Method and system for industrial internet of things

Abstract

A method and system for collecting data in an industrial environment and a method and system for using the collected data are provided. [Solution] The system includes a platform including a computing environment connected to a local data collection system having a first sensor signal and a second sensor signal from a first machine, a first sensor in the local data collection system connected to the first machine, a second sensor in the local data collection system, and a crosspoint switch having a plurality of inputs including a first input of the first sensor and a second input of the second sensor and a plurality of outputs, the plurality of outputs configured to be switchable between a state in which the first output switches between delivering the first sensor signal and delivering the second sensor signal and a state in which there is simultaneous delivery of the first sensor signal from the first output and the second sensor signal from the second output.

Description

[Technical Field]

[0001] This application is the subject of U.S. patent application Ser. No. 62 / 333,589, filed May 9, 2016, entitled "Strong Force, Industrial Internet of Things Matrix"; U.S. patent application Ser. No. 62 / 350,672, filed June 15, 2016, entitled "Automated Sequence List for Streaming Long-Term and Gap-Free Waveform Data to Storage for More Flexible Post-Processing"; and U.S. patent application Ser. No. 62 / 350,672, filed October 26, 2016, entitled "Industrial Internet of Things." 62 / 412,843, filed November 28, 2016, which claims priority to U.S. patent application Ser. No. 62 / 427,141, entitled "Method and Apparatus for Industrial Internet of Things," which claims priority to International Application No. PCT / US17 / 31721, filed May 9, 2017, and published November 16, 2017, as WO / 2017 / 196821, entitled "Method and System for Industrial Internet of Things." All of the above-referenced applications are incorporated by reference as if fully set forth herein.

[0002] The present disclosure relates to methods and systems for data collection in industrial environments, and for utilizing the collected data for monitoring, remote control, autonomous operation, and other activities in industrial environments. [Background technology]

[0003] Heavy industrial environments, such as large-scale manufacturing (e.g., aircraft, ships, trucks, automobiles, and large industrial machinery), energy production environments (e.g., oil and gas plants and renewable energy environments), energy extraction (e.g., mining and drilling), and construction environments (e.g., large building construction), contain highly complex machinery, equipment, systems, and workflows that require operators to consider numerous parameters, metrics, and other factors to optimize the design, development, deployment, and operation of these environments, as well as the operation of various technologies to improve overall results. Historically, data in heavy industrial environments was collected by humans using dedicated data collectors, sometimes by recording batches of specific sensor data to media such as tape and hard drives for later analysis. These batches of data were then returned to a central location for analysis, typically by performing signal processing or other analysis on the data collected by the various sensors. The analysis was then used as the basis for diagnosing problems in the environment and / or proposing ways to improve operations. This work has historically been performed on timescales of weeks or months and focused on limited data sets. Summary of the Invention [Problem to be solved by the invention]

[0004] The advent of the Internet of Things (IoT) has made it possible to continuously connect to a wider range of devices. Most of these devices are consumer devices, such as lights, thermostats, etc. More complex industrial environments are even more challenging because the range of available data can be limited, and the complexity of processing data from multiple sensors makes it even more difficult to create "smart" solutions that are effective in the industrial sector. There is a need for improved methods and systems for data collection in industrial environments, as well as improved methods and systems for using the collected data to provide improved monitoring, control, and intelligent diagnosis of problems and optimization of operations in various heavy industrial environments. [Means for solving the problem]

[0005] Provided herein are methods and systems for data collection in industrial environments, as well as improved methods and systems for using collected data to provide improved monitoring, control, and intelligent diagnosis of problems and optimization of operations in various heavy industrial environments. These methods and systems include methods, systems, components, devices, workflows, services, processes, and the like, deployed in a variety of configurations and locations, such as (a) the "edge" of the IoT, such as the local environment of a heavy industrial machine; (b) data transport networks that move data between the local environment of a heavy industrial machine and other environments (e.g., other machines or remote controllers) (e.g., a company that owns or operates the machine or the facility where the machine is operated); and (c) locations where facilities are located to control the machine or its environment (e.g., a cloud computing environment or on-premise computing environment of a company that owns or controls the heavy industrial environment or the machines, devices, or systems deployed in that environment). These methods and systems include a range of methods and systems for providing improved data collection, as well as methods and systems for deploying improved intelligence at the edge, the network, and the cloud or facility of controllers in industrial environments.

[0006] Disclosed herein are methods and systems for continuous ultrasonic monitoring, including providing continuous ultrasonic monitoring of rotating elements and bearings of energy production equipment.

[0007] Disclosed herein is a method and system for cloud-based machine pattern recognition based on the fusion of remote analog industrial sensors.

[0008] Disclosed herein are methods and systems for cloud-based machine pattern analysis of condition information from multiple analog industrial sensors to provide predictive condition information for industrial systems.

[0009] Disclosed herein are methods and systems for on-device sensor fusion and data storage for industrial IoT devices, including on-device sensor fusion and data storage for industrial IoT devices, where data from multiple sensors is multiplexed at the device for storing a fused data stream.

[0010] Disclosed herein are methods and systems for a self-organizing data marketplace for Industrial IoT data, including a method and system for a self-organizing data marketplace for Industrial IoT data, where available data elements are organized by consumers in a consumer marketplace based on training a self-organizing facility using feedback from a training set and market success measures.

[0011] Disclosed herein are methods and systems for self-organizing data pools, including self-organizing data pools based on utilization and / or yield metrics, including utilization and / or yield metrics tracked for multiple data pools.

[0012] Disclosed herein are methods and systems for training artificial intelligence (AI) models based on industry-specific feedback, including methods and systems for training AI models based on industry-specific feedback reflecting utilization, yield, or impact criteria, where the AI ​​models run on sensor data from an industrial environment.

[0013] Disclosed herein are methods and systems for a self-organizing swarm of industrial data collectors, including a method and system for a self-organizing swarm of industrial data collectors that self-organize to optimize data collection based on the capabilities and conditions of the swarm members.

[0014] Disclosed herein are methods and systems for an Industrial IoT distributed ledger, including a distributed ledger that supports tracking of transactions performed in an automated data marketplace for Industrial IoT data.

[0015] Disclosed herein are methods and systems for self-organizing collectors, including self-organizing multi-sensor data collectors that can optimize data collection, power, and / or yield based on conditions in an environment.

[0016] Disclosed herein are methods and systems for network-sensitive collectors, including network-state-aware, self-organizing, multi-sensor data collectors that can optimize based on bandwidth, quality of service (QoS), pricing, and / or other network conditions.

[0017] Disclosed herein are methods and systems for remotely orchestrating a universal data collector that can power up and down sensor interfaces based on identified needs and / or conditions in an industrial data collection environment.

[0018] Disclosed herein are methods and systems for self-organizing storage for multi-sensor data collectors, including self-organizing storage of multi-sensor data collectors for industrial sensor data.

[0019] Disclosed herein are methods and systems for self-organizing network coding for multi-sensor data networks, including self-organizing network coding for data networks that transfer data from multiple sensors in industrial data collection environments.

[0020] Disclosed herein are methods and systems for tactile or multi-sensory user interfaces, including wearable haptic or multi-sensory user interfaces including vibration, thermal, electrical, and / or acoustic outputs, for industrial sensor data collectors.

[0021] Disclosed herein are methods and systems for a presentation layer for augmented reality / virtual reality (AR / VR) industrial glasses in which heatmap elements are presented based on patterns and / or parameters of collected data.

[0022] Disclosed herein are methods and systems for condition-dependent, self-organized adjustment of AR / VR interfaces based on feedback metrics and / or training in industrial environments.

[0023] In an embodiment, a system for data collection, processing, and utilization of signals from at least a first element of a first machine in an industrial environment includes a platform including a computing environment connected to a local data collection system having at least a first sensor signal and a second sensor signal obtained from the at least first machine in the industrial environment. The system includes a first sensor in the local data collection system configured to be connected to the first machine and a second sensor in the local data collection system. The system further includes a crosspoint switch in the local data collection system having multiple inputs, including a first input connected to the first sensor and a second input connected to the second sensor, and multiple outputs. The multiple outputs include a first output and a second output configured to be switchable between a state in which the first output is configured to switch between delivering the first sensor signal and delivering the second sensor signal and a state in which there is simultaneous delivery of the first sensor signal from the first output and the second sensor signal from the second output. Each of the multiple inputs is configured to be individually assigned to one of the multiple outputs, and unassigned outputs are configured to be switched off to generate a high impedance state.

[0024] In an embodiment, the first sensor signal and the second sensor signal are continuous vibration data related to an industrial environment. In an embodiment, a second sensor in the local data collection system is configured to be connected to a first machine. In an embodiment, the second sensor in the local data collection system is configured to be connected to a second machine in the industrial environment. In an embodiment, the computing environment of the platform is configured to compare the relative phase of the first sensor signal and the second sensor signal. In an embodiment, the first sensor is a single-axis sensor and the second sensor is a three-axis sensor. In an embodiment, at least one of the multiple inputs of the crosspoint switch includes Internet Protocol front-end signal conditioning for an improved signal-to-noise ratio. In an embodiment, the crosspoint switch includes a third input configured to provide a continuous monitoring alarm having a default trigger condition when the third input is not assigned to any of the multiple outputs.

[0025] In an embodiment, the local data collection system includes a plurality of multiplexing units and a plurality of data acquisition units that receive a plurality of data streams from a plurality of machines in an industrial environment. In an embodiment, the local data collection system includes a distributed complex programmable logic device (CPLD) chip, each dedicated to a data bus for logical control of the plurality of multiplexing units and the plurality of data acquisition units that receive a plurality of data streams from a plurality of machines in an industrial environment. In an embodiment, the local data collection system is configured with high current input capability using solid-state relays. In an embodiment, the local data collection system is configured to power down at least one of the analog sensor channels and the component boards.

[0026] In an embodiment, the local data collection system includes an external voltage reference for an A / D zero reference that is independent of the voltages of the first and second sensors. In an embodiment, the local data collection system includes a phase-locked loop bandpass tracking filter configured to obtain slow RPM and phase information. In an embodiment, the local data collection system is configured to digitally derive phase using at least one trigger channel and an onboard timer for at least one of the multiple inputs. In an embodiment, the local data collection system includes a peak detector configured for autoscaling using a separate analog-to-digital converter for peak detection. In an embodiment, the local data collection system is configured to route at least one of the raw and buffered trigger channels to at least one of the multiple inputs. In an embodiment, the local data collection system includes at least one delta-sigma analog-to-digital converter configured to increase the input oversampling rate to reduce the sampling rate output and minimize anti-aliasing filter requirements. In an embodiment, the distributed CPLD chips each dedicated to a data bus for logical control of the multiplexing units and the multiple data acquisition units include a high frequency crystal clock reference configured to be divided by at least one of the distributed CPLD chips to reduce the sampling rate of the at least one delta-sigma analog-to-digital converter without digital resampling.

[0027] In an embodiment, the local data collection system is configured to acquire long data blocks at a single relatively high sampling rate, as opposed to multiple data sets acquired at different sampling rates. In an embodiment, the single relatively high sampling rate corresponds to a maximum frequency of approximately 40 kilohertz. In an embodiment, the long data block is greater than one minute in duration. In an embodiment, the local data collection system is configured to include multiple data collection units, each having an onboard card set, the onboard card set configured to store calibration information and maintenance history for the data collection unit in which the onboard card set is located. In an embodiment, the local data collection system is configured to plan a data collection path based on a hierarchical template.

[0028] In an embodiment, the local data collection system is configured to manage data collection bands. In an embodiment, the data collection bands define specific frequency bands and at least one of the group consisting of spectral peaks, true peak levels, crest factors derived from time waveforms, and overall waveforms derived from vibration envelopes. In an embodiment, the local data collection system includes a neural net expert system that uses intelligent management of the data collection bands. In an embodiment, the local data collection system is configured to create data collection paths based on hierarchical templates, each of which includes a data collection band associated with a machine with which the data collection path is associated. In an embodiment, at least one of the hierarchical templates is associated with multiple interconnected elements of a first machine. In an embodiment, at least one of the hierarchical templates is associated with similar elements of at least the first machine and a second machine. In an embodiment, at least one of the hierarchical templates is associated with a first machine proximate to a location of at least the second machine.

[0029] In an embodiment, the local data collection system includes a graphical user interface (GUI) system configured to manage data collection bandwidth. In an embodiment, the GUI system includes an expert system diagnostic tool. In an embodiment, the platform includes cloud-based machine pattern analysis of condition information from a plurality of sensors that provides predicted condition information for the industrial environment. In an embodiment, the platform is configured to provide self-organization of the data pool based on at least one of a utilization metric and a yield metric. In an embodiment, the platform includes a self-organized swarm of industrial data collectors. In an embodiment, the local data collection system includes a wearable tactile user interface for the industrial sensor data collector with at least one of vibration, thermal, electrical, and sound outputs.

[0030] In an embodiment, the multiple inputs of the crosspoint switch include a third input connected to a second sensor and a fourth input connected to the second sensor, and the first sensor signal is from a single-axis sensor at a fixed location associated with the first machine. In an embodiment, the second sensor is a three-axis sensor. In an embodiment, the local data collection system is configured to simultaneously record gap-free digital waveform data from at least the first input, the second input, the third input, and the fourth input. In an embodiment, the platform is configured to determine a change in relative phase based on the simultaneously recorded gap-free digital waveform data. In an embodiment, the second sensor is configured to be movable to multiple positions associated with the first machine while acquiring the simultaneously recorded gap-free digital waveform data. In an embodiment, the multiple outputs of the crosspoint switch include a third output and a fourth output, and the second, third, and fourth outputs are jointly assigned to a set of three-axis sensors each located at a different position associated with the machine. In an embodiment, the platform is configured to determine a deflection shape of the motion based on the change in relative phase and the simultaneously recorded gap-free digital waveform data.

[0031] In an embodiment, the invariant location position is a position associated with a rotating shaft of the first machine. In an embodiment, the three-axis sensors in the series of three-axis sensors are each located at a different position on the first machine and each associated with a different bearing of the machine. In an embodiment, the three-axis sensors in the series of three-axis sensors are each located at a similar position associated with a similar bearing and each associated with a different machine. In an embodiment, the local data collection system is configured to acquire simultaneously recorded gap-free digital waveform data from the first machine while both the first machine and the second machine are operating. In an embodiment, the local data collection system is configured to characterize contributions from the first machine and the second machine in the simultaneously recorded gap-free digital waveform data from the first machine. In an embodiment, the simultaneously recorded gap-free digital waveform data has a duration of more than one minute.

[0032] In an embodiment, a method for monitoring a machine having at least one shaft supported by a set of bearings includes monitoring a first data channel assigned to a single-axis sensor at a constant location associated with the machine. The method includes monitoring second, third, and fourth data channels assigned to respective axes of a three-axis sensor. The method includes simultaneously recording gap-free digital waveform data from all data channels while the machine is operating. The method includes determining a change in relative phase based on the digital waveform data.

[0033] In an embodiment, during acquisition of the digital waveform data, the three-axis sensors are positioned at multiple locations associated with the machine. In an embodiment, the second, third, and fourth channels are assigned together to a series of three-axis sensors each positioned at a different location associated with the machine. In an embodiment, data is received simultaneously from all sensors. In an embodiment, the method includes determining a deflection shape of the motion based on the relative phase information and changes in the waveform data. In an embodiment, the location of the invariant location is a location associated with a shaft of the machine. In an embodiment, the three-axis sensors in the series of three-axis sensors are each positioned at a different location and each associated with a different bearing of the machine. In an embodiment, the location of the invariant location is a location relative to the shaft of the machine. The three-axis sensors in the series of three-axis sensors are each positioned at a different location and each associated with a different bearing supporting the shaft in the machine.

[0034] In an embodiment, the method includes monitoring a first data channel assigned to a single-axis sensor at a fixed location associated with the second machine. The method includes monitoring second, third, and fourth data channels assigned to axes of a three-axis sensor located at a location associated with the second machine. The method also includes simultaneously recording gap-free digital waveform data from all data channels of the second machine while both machines are operating. In an embodiment, the method includes characterizing contributions from each machine in the simultaneous gap-free digital waveform data from the second machine.

[0035] In an embodiment, a method for data collection, processing, and utilization of signals using a monitoring platform at least at a first element of a first machine in an industrial environment includes automatically acquiring, using a computing environment, at least a first sensor signal and a second sensor signal using a local data collection system monitoring at least the first machine. The method includes connecting a first input of a crosspoint switch in the local data collection system to a first sensor and connecting a second input of the crosspoint switch to a second sensor. The method includes switching the crosspoint switch between a state in which a first output toggles between delivering the first sensor signal and delivering the second sensor signal and a state in which there is simultaneous delivery of the first sensor signal from the first output and the second sensor signal from the second output. The method also includes switching off unassigned outputs of the crosspoint switch to a high impedance state.

[0036] In an embodiment, the first sensor signal and the second sensor signal are continuous vibration data from an industrial environment. In an embodiment, a second sensor in the local data collection system is connected to a first machine. In an embodiment, the second sensor in the local data collection system is connected to a second machine in the industrial environment. In an embodiment, the method further includes automatically comparing the relative phase of the first sensor signal and the second sensor signal using a computing environment. In an embodiment, the first sensor is a single-axis sensor and the second sensor is a three-axis sensor. In an embodiment, at least a first input of the crosspoint switch includes Internet Protocol front-end signal conditioning for improved signal-to-noise ratio.

[0037] In an embodiment, the method further includes continuously monitoring at least a third input of the crosspoint switch using an alarm having a predefined trigger condition when the third input is not assigned to any of the multiple outputs of the crosspoint switch. In an embodiment, the local data collection system includes a plurality of multiplexing units and a plurality of data acquisition units that receive a plurality of data streams from a plurality of machines in an industrial environment. In an embodiment, the local data collection system includes a distributed complex programmable hardware device (CPLD) chip, each dedicated to a data bus for logical control of the multiple multiplexing units and the plurality of data acquisition units that receive the multiple data streams from the plurality of machines in the industrial environment. In an embodiment, the local data collection system has high current input capability using solid state relays.

[0038] In an embodiment, the method further includes the local data collection system powering down at least one of the analog sensor channel and the component board. In an embodiment, the local data collection system includes an external voltage reference for the A / D zero reference that is independent of the voltage of the first sensor and the second sensor. In an embodiment, the local data collection system includes a phase-locked loop bandpass tracking filter that obtains slow RPM and phase information. In an embodiment, the method includes digitally deriving the phase using at least one trigger channel on the crosspoint switch and an onboard timer for at least one of the multiple inputs.

[0039] In an embodiment, the method includes autoscaling using a peak detector with a separate analog-to-digital converter for peak detection. In an embodiment, the method includes routing at least one raw and / or buffered trigger channel on the crosspoint switch to at least one of a plurality of inputs. In an embodiment, the method includes increasing the input oversampling rate using at least one delta-sigma analog-to-digital converter to reduce the sampling rate output and minimize anti-aliasing filter requirements. In an embodiment, distributed CPLD chips are each dedicated to a data bus for logical control of the multiplexing units and the plurality of data acquisition units, and each includes a high-frequency crystal clock reference configured to be divided by at least one of the distributed CPLD chips to reduce the sampling rate without digital resampling. In an embodiment, the method includes acquiring a long data block with a single relatively high sampling rate data, as opposed to multiple data sets acquired at different sampling rates. In an embodiment, the single relatively high sampling rate corresponds to a maximum frequency of approximately 40 kilohertz. In an embodiment, the long data block is greater than one minute in duration. In an embodiment, the local data collection system includes a plurality of data collection units, each having an on-board card set that stores calibration information and maintenance history for the data collection unit in which the on-board card set is located.

[0040] In an embodiment, the method includes planning a data collection path based on a hierarchical template associated with at least a first element of a first machine in an industrial environment. In an embodiment, the local data collection system manages data collection bands, the data collection bands defining specific frequency bands and at least one of the group consisting of spectral peaks, true peak levels, crest factors derived from time waveforms, and overall waveforms derived from vibration envelopes. In an embodiment, the local data collection system includes a neural net expert system that uses intelligent management of the data collection bands. In an embodiment, the local data collection system creates data collection paths based on hierarchical templates, each of which includes a data collection band associated with a machine with which the data collection path is associated. In an embodiment, at least one of the hierarchical templates is associated with multiple interconnected elements of the first machine. In an embodiment, at least one of the hierarchical templates is associated with similar elements of at least the first machine and a second machine. In an embodiment, at least one of the hierarchical templates is associated with the first machine proximate to the location of at least the second machine.

[0041] In an embodiment, the method includes controlling a GUI system of a local data collection system to manage data collection bandwidth. The GUI system includes an expert system diagnostic tool. In an embodiment, the platform's computing environment includes cloud-based machine pattern analysis of condition information from a plurality of sensors that provides predicted condition information for the industrial environment. In an embodiment, the platform's computing environment provides self-organization of a data pool based on at least one of a utilization metric and a yield metric. In an embodiment, the platform's computing environment includes a self-organized swarm of industrial data collectors. In an embodiment, each of a plurality of inputs of a crosspoint switch is individually assignable to any of a plurality of outputs of the crosspoint switch. [Brief explanation of the drawings]

[0042] [Figure 1] 1 through 5 are schematic diagrams each illustrating an overall Industrial Internet of Things (IoT) data collection, monitoring, and control system according to the present disclosure. [Figure 2] FIG. 2 and FIGS. 1 to 5 are schematic diagrams showing the overall industrial Internet of Things (IoT) data collection, monitoring, and control system according to the present disclosure, respectively. [Figure 3] FIG. 3 and FIG. 1 to FIG. 5 are schematic diagrams showing the overall industrial Internet of Things (IoT) data collection, monitoring, and control system according to the present disclosure, respectively. [Figure 4] FIG. 4 and FIG. 1 to FIG. 5 are schematic diagrams showing the overall industrial Internet of Things (IoT) data collection, monitoring, and control system according to the present disclosure, respectively. [Figure 5] 5 is a schematic diagram illustrating an overall industrial Internet of Things (IoT) data collection, monitoring, and control system according to the present disclosure, respectively.

[0043] [Figure 6] FIG. 6 is a schematic diagram illustrating a platform including a local data collection system deployed in an industrial environment to collect data regarding or from elements of the environment, such as machines, components, systems, subsystems, ambient conditions, states, workflows, processes, and other elements, in accordance with the present disclosure.

[0044] [Figure 7] FIG. 7 is a schematic diagram illustrating elements of an industrial data collection system for collecting analog sensor data in an industrial environment in accordance with the present disclosure.

[0045] [Figure 8] FIG. 8 is a schematic diagram of a rotating or oscillating machine having a data acquisition module configured to collect waveform data in accordance with the present disclosure.

[0046] [Figure 9] FIG. 9 is a schematic diagram illustrating an example three-axis sensor mounted on a motor bearing of an example rotating machine according to the present disclosure.

[0047] [Figure 10] 10 and 11 are schematic diagrams illustrating an example three-axis sensor and a single-axis sensor mounted on a motor bearing of an example rotating machine according to the present disclosure. [Figure 11] 10 and 11 are schematic diagrams illustrating an example three-axis sensor and a single-axis sensor mounted on a motor bearing of an example rotating machine according to the present disclosure.

[0048] [Figure 12] FIG. 12 is a schematic diagram showing multiple machines under investigation using an array of multiple sensors in accordance with the present disclosure.

[0049] [Figure 13] FIG. 13 is a schematic diagram illustrating a hybrid relational metadata and binary storage approach in accordance with the present disclosure.

[0050] [Figure 14] FIG. 14 is a schematic diagram illustrating components and interactions of a data collection architecture with application of cognitive and machine learning systems for data collection and processing according to the present disclosure.

[0051] [Figure 15] FIG. 15 is a schematic diagram illustrating components and interactions of a data collection architecture with applications of a cognitive data marketplace platform according to the present disclosure.

[0052] [Figure 16] FIG. 16 is a schematic diagram illustrating components and interactions of a data collection architecture involving application of a self-organized swarm of data collectors according to the present disclosure.

[0053] [Figure 17] FIG. 17 is a schematic diagram illustrating components and interactions of a data collection architecture with a haptic user interface application according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0054] Detailed embodiments of the present disclosure are disclosed herein. However, the disclosed embodiments are merely exemplary of the present disclosure, which may be embodied in various forms. Therefore, the specific structural and functional details disclosed herein should not be construed as limiting, but should be construed solely as a basis for the claims, and as a representative basis for teaching those skilled in the art how to utilize the present disclosure in a variety of ways in virtually any suitable detailed structure.

[0055] The terms "a" or "an," as used herein, are defined as one or more than one. The term "another," as used herein, is defined as at least a second or more. The terms "including" and / or "having," as used herein, are defined as comprising (i.e., open transitions).

[0056] While only certain embodiments of the present disclosure have been shown and described, it will be apparent to those skilled in the art that many changes and modifications can be made therein without departing from the spirit and scope of the disclosure, as set forth in the following claims. All patent applications and patents (both foreign and domestic) and all other publications referenced herein are incorporated herein in their entirety to the extent permitted by law.

[0057] Figures 1 through 5 show a portion of an overall view of an Industrial Internet of Things (IoT) data collection, monitoring, and control system 10. Figure 2 shows the upper left portion of the schematic diagram of the Industrial Internet of Things (IoT) system 10 of Figures 1 through 5. Figure 2 includes a mobile ad hoc network (MANET) 20 that forms a secure and temporary network connection 22 (which may be connected or isolated) with a cloud 30 or other remote networking system, allowing network functions to be performed across the MANET 20 within the environment without the need for an external network, and information to be transmitted to and from a central location. This enables the benefits of networking and control technology to be used in industrial environments and provides security, such as preventing cyberattacks. The MANET 20 uses cognitive radio technology 40, including routers 42, MAC 44, and physical layer technology 46, which form the equivalent of IP protocols. Network-sensitive or network-aware transport of data across the network to and from data collection devices or heavy industrial machinery is also shown.

[0058] FIG. 3 shows the upper right portion of the schematic diagram of the Industrial Internet of Things system 10 of FIGS. 1 through 5. It includes intelligent data collection technology 50 deployed locally at the edge of the IoT deployment where heavy industrial machinery is located. This includes various sensors 52, IoT devices 54, data storage capabilities 56 (including intelligent self-organizing storage), sensor fusion (including self-organizing sensor fusion), etc. FIG. 3 shows interfaces for data collection, including multi-sensory interfaces, tablets, smartphones 58, etc. FIG. 3 also shows a data pool 60 that collects data published by machines or sensors that detect machine conditions, such as for later consumption by local or remote intelligence. A distributed ledger system 62 distributes storage across local storage in various elements of the environment or more broadly across the entire system.

[0059] Figure 1 shows the central portion of the Industrial IoT system schematics of Figures 1 through 5. This includes the use of network coding (including self-organizing network coding) to configure network coding models based on feedback measurements, network conditions, etc., to efficiently transmit large amounts of data over the network to and from data collection systems and the cloud. A wide range of capabilities, including intelligence, analytics, remote control, remote operation, and remote optimization, are deployed in the cloud or on the premises of the enterprise owner or operator (see Figure 1). This includes various storage configurations, including distributed ledger storage, to support transaction data and other elements of the system, etc.

[0060] Figures 1, 4, and 5 illustrate the lower right corner of the schematic diagram of the Industrial IoT system of Figures 1 through 5. This includes a program data market 70, which may be a self-organizing marketplace for making available data collected in an industrial environment, such as data collectors, data pools, distributed ledgers, and other elements disclosed herein and illustrated in Figures 1 through 5. Figures 1, 4, and 5 also illustrate on-device sensor fusion 80 for storing device data, for example, from multiple analog sensors 82, which are analyzed locally or in the cloud, including by machine learning 84, for example, by training a machine based on an initial human-created model that is augmented by providing feedback (e.g., based on measures of success) when operating the methods and systems disclosed herein. Additional details regarding the various components and subcomponents of Figures 1 through 5 are provided throughout this disclosure.

[0061] In an embodiment, a method and system are provided for a system for collecting, processing, and utilizing data in an industrial environment, referred to herein as platform 100. Referring to FIG. 6 , platform 100 includes a local data collection system 102 that is deployed in an environment 104, such as an industrial environment, to collect data from or about elements of the environment, such as machines, components, systems, subsystems, ambient conditions, states, workflows, processes, and other elements. Platform 100 connects to or includes a portion of the industrial IoT data collection, monitoring, and control system 10, illustrated in FIGS. 1 through 5 . Platform 100 also includes a network data transfer system 108, such as one deployed within a computing environment or enterprise campus, or consisting of distributed components that interact with each other to process data collected by local data collection system 102, such as via a network 110 with a host processing system 112. The host processing system 112, sometimes referred to as the host system 112 for convenience, includes various systems, components, methods, processes, equipment, etc. for enabling automated or automation-assisted processing of data, such as for monitoring one or more environments 104 or networks 110 or for remotely controlling one or more elements within the local environment 104 or network 110.Platform 100 includes one or more local autonomous systems 114, e.g., enabling autonomous operation, e.g., reflecting artificial or machine-based intelligence, e.g., enabling automated action based on application of a set of rules or models on input data from local data collection system 102 or from one or more input sources 116, including information feeds and inputs from a wide range of sources, including from the local environment 104, network 110, host system 112, or one or more external systems, databases, etc. Platform 100 includes one or more intelligent systems 118, which are located, integrated with, or serve as inputs to one or more components of platform 100. Details of these and other components of platform 100 are provided throughout this disclosure.

[0062] Intelligent systems include cognitive systems 120 that enable some degree of cognitive behavior as a result of the coordination of processing elements, such as mesh systems, peer-to-peer systems, ring systems, serial systems, and other architectures, where one or more node elements cooperate with other node elements to provide collectively coordinated operations to support processing, communication, data collection, etc. The MANET 20 shown in FIG. 2 uses cognitive radio technology, including equivalents to IP protocols, such as routers 42, MAC 44, and physical layer technology 46. In one example, the cognitive system technology stack includes that disclosed in U.S. Patent No. 8,060,017 to Schlicht et al. (issued November 15, 2011), which is incorporated herein by reference as if fully set forth herein. The intelligent system includes a machine learning system 122 that learns, for example, from one or more datasets. The one or more datasets may include information collected using the local data collection system 102 or other information from input sources 116. The information identifies states, objects, events, patterns, situations, etc., which are then used for processing by the processing system 112 as input to components of the platform 100 and parts of the industrial IoT data collection, monitoring, and control system 100, etc. The learning may be human-supervised or fully automated, and the supervision and automation may use one or more input sources 116 to provide datasets, for example, with information about the items to be learned. Machine learning is the generation of one or more models, rules, semantic understanding, workflows, or other structured or semi-structured understanding of the world, for example, for automated optimization of control of a system or process based on feedback or feedforward to the system or process. One such machine learning technique for semantic and contextual understanding, workflow, or other structured or semi-structured understanding is disclosed in U.S. Patent No. 8,200,775 to Moore, issued June 12, 2012. Machine learning is used to improve the above, for example, by adjusting one or more weights, structures, rules, etc. (e.g., changing functions within the model), based on feedback or iteration, where the feedback, for example, relates to the model's success in a given situation, and the iteration, for example, is a recursive process. Machine learning may occur without an underlying model when a full understanding of the system's underlying structure and operation is unknown, when insufficient data is available, or when other situations are preferred for various reasons. That is, input sources may be weighted and structured within the machine learning facility without a priori understanding of the structure, and results (e.g., based on measures of success in achieving various desired goals) may be continuously fed to the machine learning system to learn how to achieve those goals.For example, the system may learn to recognize failures, recognize patterns, develop models or functions, develop rules, optimize performance, minimize failure rates, optimize profits, optimize resource utilization, optimize flow (e.g., traffic), or many other parameters associated with successful outcomes, such as across a wide range of environments. Machine learning uses genetic programming techniques, such as promoting or demoting one or more input sources, structures, data types, objects, weights, nodes, links, or other factors based on feedback, so that successful elements emerge over a series of generations. For example, available sensor inputs for the data collection system 102 may be arranged in alternative configurations and arrangements using genetic programming techniques over a series of data collection events to determine which arrangements will produce successful results based on various conditions, such as conditions of the platform 100 components, conditions of the network 110, conditions of the data collection system 102, and conditions of the environment 104. In embodiments, local machine learning turns on or off one or more sensors in multi-sensor data collector 102, sequenced over time, while tracking success contributes to performance indicators (e.g., efficiency, effectiveness, return on investment, yield, etc.) and contributes to optimizing one or more parameters, patterns (e.g., related to threats, failure modes, success modes, etc.). For example, the system learns what set of sensors should be turned on or off under given conditions to achieve the highest utilization of data collector 102. In embodiments, similar techniques are used to handle optimization of data transfer within platform 100, such as by using genetic programming or other machine learning techniques to learn to configure network components within network 110, such as configuring network transmission paths, network coding types, network architectures, network security components, etc.

[0063] In an embodiment, the local data collection system 102 may include a high-performance, multi-sensor data collector with many novel features for collecting and processing analog and other sensor data. In an embodiment, the local data collection system 102 may be deployed in the industrial facility illustrated in FIG. 3. The local data collection system 102 may be deployed monitoring other machines, such as machine 2300 in FIGS. 9 and 10, machines 2400, 2600, 2800, 2950, ​​and 3000 illustrated in FIG. 12, and machines 3202 and 3204 illustrated in FIG. 13. The data collection system 102 may include an on-board intelligent system for learning to optimize the configuration and operation of the data collector, such as configuring the arrangement and combination of sensors based on context and conditions. In one example, the data collection system 102 includes a crosspoint switch 130. The automated, intelligent configuration of the local data collection system 102 may be based on various types of information, such as from various input sources and based on, for example, available power, power requirements of sensors, value of collected data (e.g., based on feedback information from other elements of the platform 100), relative value of information (based on availability of other sources of the same or similar information), power availability (e.g., to power sensors), network conditions, ambient conditions, operating conditions, operating context, operating events, and the like.

[0064] FIG. 7 illustrates elements and subcomponents of a data collection and analysis system 1100 for sensor data (e.g., analog sensor data) collected in an industrial environment. As illustrated in FIG. 7, embodiments of the disclosed methods and systems may include hardware with several different modules, beginning with a multiplexer (“Mux”) 1104. In an embodiment, the Mux 1104 is comprised of a main board and an option board 1108. The main board is where sensors connect to the system. These connections are located on the top for ease of installation. Numerous settings are located on the underside of this board and on a Mux option board, which connects to the main board via two headers on either end of the board. In an embodiment, the Mux option board has a male header that mates with a female header on the main Mux board. This reduces real estate and allows for stacking.

[0065] In an embodiment, the main Mux connects via cables to mother (e.g., with four simultaneous channels) and daughter (e.g., with four additional channels for eight total channels) analog boards 1110, where some of the signal conditioning, such as hardware integration, occurs. The signal then travels from the analog board 1110 to an anti-aliasing board, where some of the potential aliasing is removed. The remainder of the aliasing occurs on the delta-sigma board 1112, which is connected via cables to the delta-sigma board 1112. The delta-sigma board 1112 provides more aliasing protection, along with other conditioning and digitization of the signal. The data then travels to the Jennic board 1114 for more digitization, as well as communication with a computer via USB or Ethernet. In an embodiment, the Jennic board 1114 may be replaced with a PIC board 1118 for more advanced and efficient data collection and communication. Once the data is transferred to the computer software 1102, the computer software 1102 manipulates the data to present trends, spectra, waveforms, statistics, and analysis.

[0066] In embodiments, the system is meant to capture any type of data, from volts to 4-20 mA signals. In embodiments, open formats for data storage and communication may be used. In some instances, portions of the system may be proprietary, with some surveys and data specifically associated with analysis and reporting. In embodiments, smart band analysis is a way to break down data into easily analyzed parts that can be combined with other smart bands to create new, simpler, yet sophisticated analyses. In embodiments, this inherent information is taken to represent conditions, and graphics are used because pictorial depictions are more useful to users. In embodiments, complex programs and user interfaces are simplified, allowing any user to manipulate data like an expert.

[0067] In an embodiment, the system essentially operates in a large loop. It starts in software with a generic user interface. Most, if not all, online systems require the OEM to create or develop a system GUI 1124. In an embodiment, rapid route creation utilizes hierarchical templates. In an embodiment, a graphical user interface ("GUI") is created, allowing any generic user to input their own information into a simple template. Once the template is created, the user can copy and paste what they need. Additionally, users can develop their own templates for future ease of use and knowledge institutionalization. When the user has entered all of their information and connected all of their sensors, the user can begin the system acquiring data. In some applications, rotating machinery accumulates electrical charge that can be harmful to electrical equipment. In an embodiment, to reduce the effect of this charge on the device, inherent electrostatic protection for trigger and vibration inputs is pre-placed on the Mux and DAQ hardware to dissipate this charge as the signal passes from the sensor to the hardware. In an embodiment, the Mux and Analog boards may also provide precedence circuitry and wider traces in high current input capabilities, if desired, using solid state relays and design topologies that allow the system to handle high amperage inputs.

[0068] In an embodiment, a key part of the front end of the Mux is the front end signal conditioning on the Mux for improved signal-to-noise ratio, providing proactive signal conditioning. Most multiplexers are afterthoughts, and OEM manufacturers typically don't care or consider the quality of the signal coming from them. As a result, signal quality can degrade by 30 dB or more. Since any system is only as strong as its weakest link, even with a 24-bit DAQ with a 110 dB SNR, signal quality is already lost through the Mux. When the signal-to-noise ratio degrades to 80 dB at the Mux, it's not much better than a 16-bit system from 20 years ago.

[0069] In addition to providing a better signal, multiplexers can also play an important role in enhancing the system. While truly continuous systems monitor all sensors constantly, these systems are prohibitively expensive. Multiplexer systems typically monitor only a set number of channels at a time and can switch between banks of sensors from a larger set. As a result, sensors that are not being acquired are not being monitored, and users may not know if a level rises. In embodiments, the multiplexer's continuous monitoring alarm function provides a continuous monitoring alarm multiplexer by placing circuitry on the multiplexer that can measure levels relative to known alarms even when the data acquisition ("DAQ") is not monitoring the channel. This essentially makes the system continuous without the ability to instantly capture the data of interest, as in a true continuous system. In embodiments, combining these capabilities with alarms, adaptive scheduling techniques for continuous monitoring and continuous monitoring system software for adapting and adjusting data collection sequences based on statistics, analytics, data alarms, and dynamic analysis, allows the system to quickly collect dynamic spectral data on alarm sensors immediately after an alarm sounds.

[0070] Another limitation of multiplexers is that they often have a limited number of channels. In embodiments, by using a distributed complex programmable logic device ("CPLD") chip with dedicated buses for logical control of multiple Muxes and data acquisition sections, the CPLD can control multiple Muxes and DAQs with no limit on the number of channels the system can handle. In embodiments, the Muxes and DAQs are stacked on top of each other to provide additional input and output channels to the system.

[0071] In addition to having a limited number of channels, multiplexers can typically only collect sensors in the same bank. For detailed analysis, this is very limiting, since there is great value in being able to simultaneously review data from sensors on the same machine. In embodiments, the use of analog crosspoint switches to collect variable groups of vibration input channels addresses this issue by using crosspoint switches commonly used in the telephone industry, and also providing matrix circuitry so that the system can access any set of eight channels from the total number of input sensors.

[0072] In an embodiment, the system allows for remote balancing of low speed machinery, such as in a paper mill, as well as on-site providing additional analysis from the data to the phase locked loop bandpass tracking filter method for low rotational speeds ("RPM") and balancing phases.

[0073] In embodiments, the ability to control multiple multiplexers using a distributed CPLD chip with multiple Muxes and a dedicated bus for logic control of the data acquisition section is enhanced by a hierarchical multiplexer that allows multiple DAQs to collect data from multiple multiplexers. In embodiments, this allows for faster data acquisition and more channels of simultaneous data acquisition for improved analysis. In embodiments, the Muxes, with minimal configuration, make it portable and can use the data acquisition parking feature, thereby turning the SV3X DAQ into a protected system.

[0074] In embodiments, once the signal leaves the multiplexer and hierarchical Mux, it travels to an analog board where further expansion occurs. In embodiments, other power saving measures, including powering down analog channels when not in use and powering down component boards, allow the system to power down channels on the motherboard and daughter analog boards to conserve power. In embodiments, this can provide the same power savings to a protection system, especially when running on battery or solar power. In embodiments, a peak detector for autoscaling, routed to a separate A / D to maximize signal-to-noise ratio and provide the best data, provides the system with the highest peak in each data set so it can rapidly scale data up to that peak. In embodiments, improved integration using both analog and digital methods provides an innovative hybrid integration that improves or maintains the highest possible signal-to-noise ratio.

[0075] In an embodiment, a section of the analog board allows trigger channels to be routed to other analog channels, either raw or buffered. This allows users to route triggers to any channel for analysis and troubleshooting. In one embodiment, once the signal leaves the analog board, it travels to a delta-sigma board, where a precise voltage reference for the A / D zero reference provides more accurate direct current (DC) sensor data. The high speed of the delta-sigma allows for the use of higher input oversampling of the delta-sigma A / D for low sampling rate outputs, thereby minimizing anti-aliasing (AA) filter requirements for oversampling data at higher inputs, minimizing anti-aliasing requirements. In an embodiment, a CPLD is used as a clock divider for the delta-sigma A / D, thereby achieving lower sampling rates without the need for digital resampling, allowing the delta-sigma A / D to achieve lower sampling rates without digital resampling.

[0076] In an embodiment, the data then travels from the Delta Sigma board to the Jennic board, which digitally derives the phase for the input and trigger channels using onboard timers to digitally derive the phase from the input signals and triggers. In an embodiment, the Jennic board also has the ability to store calibration data and system maintenance correction history data on an onboard card set. In an embodiment, the Jennic board allows for the acquisition of long blocks of data at a high sampling rate, as opposed to multiple data sets obtained at different sampling rates, so that blocks of data can be streamed and acquired for future advanced analysis.

[0077] In an embodiment, the signal passes through a Jennic board before being sent to a computer. Once on the computer, the software has numerous enhancements that improve system analysis capabilities. In an embodiment, rapid route creation utilizes hierarchy templates, providing rapid route creation for all equipment using simple templates that also speed software deployment. In an embodiment, the software will be used to add intelligence to the system. Starting with an expert GUI graphical approach to defining smart bands and diagnostics for the expert system, the expert system will provide a graphical expert system with a simple user interface so that anyone can develop complex analyses. In an embodiment, this user interface revolves around the smart band, which is a simplified approach to complex yet flexible analyses for the average user. In an embodiment, the smart band is paired with a self-learning neural network for a more advanced analytical approach. In an embodiment, the system also uses the machine's hierarchy for additional analytical insights. One important part of predictive maintenance is the ability to learn from known information during repairs or inspections. In an embodiment, a graphical approach for back calculations can improve smart bands and correlations based on known defects or problems.

[0078] In embodiments, bearing analysis methods are provided in addition to detailed analysis by smart bands. Recent years have seen strong activity in the industry to conserve power and have led to the influx of variable frequency drives. In embodiments, analysis utilizing torsional vibration detection and transient signal analysis provides advanced torsional vibration analysis for a more comprehensive method of diagnosing machines where torsional forces are involved (e.g., machines with rotating components). In embodiments, the system itself can deploy numerous intelligent features for better data and more comprehensive analysis. In embodiments, this intelligence begins with smart routing software, which can adapt sensors collecting simultaneously to gain additional correlation intelligence. In embodiments, a smart operational data store ("ODS") allows the system to select to collect operational deflection shape analysis to further examine machine condition. In embodiments, in addition to modifying the routing, adaptive scheduling techniques for continuous monitoring allow the system to modify the scheduled data collected for full spectrum analysis across numerous (e.g., eight) correlation channels. System intelligence provides data that allows for expanded statistical capabilities for continuous monitoring, as well as ambient and local temperature and local vibration for combined analysis of changes in vibration levels to identify machine problems.

[0079] Embodiments of the methods and systems disclosed herein may include a self-contained DAQ box. In embodiments, data acquisition devices may be controlled by a personal computer ("PC") to implement desired data acquisition commands. In embodiments, the system has self-contained capabilities and can acquire, process, analyze, and monitor data independently of external PC control. Embodiments of the methods and systems disclosed herein may include Secure Digital ("SD") card storage. In embodiments, significant additional storage capacity is provided by utilizing SD cards in cameras, smartphones, and the like. This is important for monitoring applications that can persistently store important data. Also, if a power outage occurs, the most recent data may be stored despite not being offloaded to another system. Embodiments of the methods and systems disclosed herein may include a data acquisition ("DAQ") system. The current trend is for DAQ systems to communicate with the outside world as much as possible, usually via some form of network, including wireless. While in the past it was common to control DAQ systems with a microprocessor or microcontroller / microprocessor paired with a PC using a dedicated bus, today networking requirements are so great that this new design prototype arises outside of this environment. In embodiments, multiple microprocessors / microcontrollers or dedicated processors are used to perform various aspects of this increased DAQ functionality, with one or more processor units focused primarily on the communication aspects with the outside world. This negates the need to constantly interrupt the main process, which includes controlling signal conditioning circuitry, triggering, acquiring raw data using an A / D, directing the A / D output to appropriate onboard memory, and processing that data. In embodiments, a dedicated microcontroller / microprocessor is designated for all communication with the outside world. These include USB, Ethernet, and wireless, with the ability to provide an IP address to host a web page.All communication with the outside world is performed through the use of simple text-based menus. A regular array of commands (in fact, over 100) is provided, such as InitializeCard, AcquireData, StopAcquisition, and RetrieveCalibrationInfo. Furthermore, in embodiments, other powerful signal processing operations, including resampling, weighting, filtering, and spectral processing, can be performed by dedicated processors such as field programmable gate arrays ("FPGAs"), digital signal processors ("DSPs"), microprocessors, microcontrollers, or combinations thereof. In embodiments, this subsystem communicates with the communications processing section through a dedicated hardware bus, which may be facilitated by dual-port memory, semaphore logic, and the like. This embodiment not only provides significant improvements in efficiency, but can also significantly improve processing capabilities, including data streaming and other high-end analysis techniques.

[0080] Embodiments of the methods and systems disclosed herein can include sensor overload identification. Systems need to be monitored to identify when a sensor is overloaded. While a monitoring system can identify when the system is overloaded, in embodiments, the system can look at the sensor voltage to determine if the overload is coming from the sensor, which can be useful for the user to obtain a different sensor that is more suitable for the situation, or the user can attempt to recollect the data. Often, there are situations with high frequency inputs that saturate the standard 100 mV / g sensors most commonly used in the industry, and having the ability to detect overload improves data quality for better analysis.

[0081] Embodiments of the methods and systems disclosed herein can include RF ID (radio frequency identification) on the accelerometer and RF ID on the inclinometer or other sensor so that the sensor can tell the system / software which machine / bearing and mounting orientation it is on, and the software can automatically set up and store the data without user instruction. In embodiments, a user can place the system on any machine, and the system will automatically set itself up and be ready to collect data in seconds.

[0082] Embodiments of the methods and systems disclosed herein can include ultrasonic online monitoring by placing ultrasonic sensors inside transformers, motor control centers (MCCs), breakers, etc., where the system continuously searches through the acoustic spectrum for patterns that identify arcing, corona, and other electrical issues that indicate a fault or problem. In embodiments, an analytical engine is used for the ultrasonic online monitoring and to identify other faults by combining the data with other parameters such as vibration, temperature, pressure, heat flux, magnetic field, electric field, current, voltage, capacitance, inductance, other parameters, or combinations thereof (e.g., simple ratios).

[0083] Embodiments of the methods and systems disclosed herein can include the use of analog crosspoint switches to collect variable groups of vibration input channels. For vibration analysis, it is useful to simultaneously acquire multiple channels from vibration transducers mounted in multiple orientations on different parts of a machine. By acquiring simultaneous readings, for example, the relative phase of the inputs can be compared for the purpose of diagnosing various mechanical faults. Other types of cross-channel analysis, such as cross-correlation, transfer function, and operational deflection shape ("ODS"), are also possible. Current systems using traditional fixed-bank multiplexers can only compare a limited number of channels (based on the number of channels per bank), assigned to specific groups at installation. The only way to provide some flexibility is to duplicate channels or build a lot of redundancy into the system, both of which add significant expense (which can be exponential in cost vs. flexibility). The simplest Mux design selects one of many inputs and sends it to a single output line. Banked designs consist of groups of these simple building blocks, each processing a fixed group of inputs and sending them to their respective outputs. Typically, inputs are non-overlapping, so an input from one Mux group cannot be routed to another Mux group. Unlike traditional Mux chips, which typically switch a fixed group or bank (e.g., a group of 2, 4, 8, etc.) of channels to a single output, a crosspoint Mux allows the user to assign any input to any output. Previously, crosspoint multiplexers were used for specialized applications such as RGB digital video applications, and were noisy for analog applications such as vibration analysis. However, more recent advances in technology have made it feasible. Another advantage of the crosspoint Mux is the ability to disable outputs by placing them in a high-impedance state.This is ideal for output buses so that multiple Mux cards can be stacked and their output buses combined without the need for a bus switch.

[0084] Embodiments of the methods and systems disclosed herein can include front-end signal conditioning on the Mux for improved signal-to-noise ratio. Embodiments can perform signal input (e.g., range / gain control, integration, filtering, etc.) before Mux switching to achieve maximum signal-to-noise ratio.

[0085] Embodiments of the methods and systems disclosed herein may include a Mux Continuous Monitor Alarm feature, in which a Continuous Monitor Mux Bypass allows channels not currently being sampled by the Mux system to be continuously monitored for significant alarm conditions through a number of trigger conditions using filtered peak-hold circuits, or similar functionality, and conveniently passed to the monitoring system using hardware interrupts or other means.

[0086] Embodiments of the disclosed method and system can include the use of distributed CPLD chips with dedicated buses for logical control of multiple Mux and data acquisition sections. Connecting to multiple types of predictive maintenance and vibration transducers requires a large amount of switching. This includes AC / DC coupling, 4-20 interfaces, IEPE (Integrated Electronic Piezoelectric Transducers), channel power-down (to conserve op-amp power), single-ended or differential grounding options, and more. Control is also required for digital pots for range and gain control, switches for hardware integration, AA filtering, and triggering. This logic can be performed by a series of CPLD chips strategically placed for the task they control. A single, large CPLD would require long circuit paths with very high density in that single, large CPLD. In embodiments, distributed CPLDs address these concerns while also providing significant flexibility. A bus is created where each CPLD has its own unique device address with a fixed assignment. Jumpers are provided to set multiple addresses for multiple boards, e.g., multiple Mux boards. In another example, three bits allow for up to eight jumper-configurable boards. In an embodiment, a bus protocol is defined such that each CPLD on the bus can be addressed individually or as a group.

[0087] Embodiments of the methods and systems disclosed herein can include high-amperage input capabilities using solid-state relays and design topologies. Vibration data collectors are typically not designed to handle large input voltages due to cost and the fact that they are often not necessary. The need for these data collectors to acquire various types of PM data continues to grow as technology improves and costs plummet. In embodiments, a method uses established OptoMOS technology, which allows for high-voltage signal front switching, rather than using traditional reed relay techniques. Many historical concerns about nonlinear zero crossings or other nonlinear solid-state behavior are eliminated with respect to the passing of weakly buffered analog signals. Furthermore, in embodiments, the printed circuit board (PCB) routing topology places all of the individual channel input circuitry as close as possible to the input connector.

[0088] Embodiments of the methods and systems disclosed herein can include other power saving measures, including powering down analog channels when not in use and powering down component boards. In embodiments, the powering down of analog signal processing op-amps for unselected channels, as well as the ability to power down component boards and other hardware through low-level firmware for the DAQ system, makes high-level control over power saving features relatively simple. Explicit control of hardware is always possible, but is not required by default.

[0089] Embodiments of the methods and systems disclosed herein can include inherent electrostatic protection for trigger inputs and vibration inputs. In many critical industrial environments where large electrostatic forces can create, such as low-speed balancing using large belts, proper transducer and trigger input protection is necessary. In embodiments, a low-cost yet efficient method for such protection is described that does not require external supplemental devices.

[0090] Embodiments of the methods and systems disclosed herein can include an accurate voltage reference for the A / D zero reference. Some A / D chips provide an internal zero voltage reference to be used as the midscale value of external signal conditioning circuitry, ensuring that both the A / D and the external op amp use the same reference. While this is reasonable in principle, it presents practical complications. Often, these references are essentially based on the power supply voltage using a resistive voltage divider. For many current systems, especially those powered from a PC via USB or similar bus, this provides an unreliable reference because the power supply voltage can vary significantly with the load. This is especially true for delta-sigma A / D chips, which require increased signal processing. Although offsets can drift with load, problems arise if measurements need to be digitally calibrated. It is common to digitally modify the voltage offset, expressed as counts from the A / D, to compensate for DC drift. However, in this case, once the appropriate calibration offset is determined for one set of load conditions, it does not apply to other conditions. The absolute DC offset, expressed in counts, would no longer apply. This results in a need to calibrate for all loading conditions which is complex, unreliable and ultimately unmanageable. In an embodiment, an external voltage reference that is simply independent of the power supply voltage is used to use as a zero offset.

[0091] Embodiments of the methods and systems disclosed herein can include a phase-locked loop bandpass tracking filter method for obtaining slow-speed RPM and phase for balancing purposes. For balancing purposes, it is sometimes necessary to balance at very slow speeds. Typical tracking filters can be constructed based on phase-locked loop or PLL designs. However, stability and speed range are paramount concerns. In embodiments, multiple digitally controlled switches are used to select appropriate RC and damping constants. The switching can be performed entirely automatically after measuring the frequency of the incoming tachometer signal. Embodiments of the methods and systems disclosed herein can include digital derivation of phase for input and trigger channels using on-board timers. In embodiments, the digital phase derivation uses digital timers to ascertain the precise delay from the trigger event to the exact start of data acquisition. This delay or offset can be further refined using interpolation to obtain a more precise offset, which is applied to the analytically determined phase of the acquired data such that the phase is essentially absolute, with precise mechanical meaning useful for one-shot balancing, alignment analysis, and the like.

[0092] Embodiments of the methods and systems disclosed herein can include a peak detector for autoscaling routed to a separate A / D. Many microprocessors in use today have built-in A / D converters. For vibration analysis purposes, the number of bits, number of channels, or sampling frequency relative to a significantly slower microprocessor is often insufficient. Despite these limitations, using them for autoscaling purposes is convenient. In embodiments, a separate A / D with reduced performance and low cost can be used. For each input channel, after the signal is buffered (usually with appropriate coupling: AC or DC) and before signal conditioning, it is fed directly to a microprocessor or low-cost A / D. Unlike the conditioned signal, where range, gain, and filter switches are thrown, the switches are not changed. This can enable simultaneous sampling of autoscaling data, while the input data is signal conditioned and fed to a more robust external A / D and routed to onboard memory using a direct memory access (DMA) method, which can be accessed without the need for a CPU. This greatly simplifies the autoscaling process by eliminating the need to flip switches and consider settling times, thereby significantly slowing down the autoscaling process. Furthermore, data can be collected simultaneously, thereby ensuring the best signal-to-noise ratio. The reduced number of bits and other features are usually more than sufficient for autoscaling purposes.

[0093] Embodiments of the methods and systems disclosed herein can include routing trigger channels, either raw or buffered, to other analog channels. Many systems have trigger channels for purposes such as determining the relative phase between various input data sets or acquiring important data without needlessly repeating unwanted inputs. In embodiments, a digitally controlled relay switches either a raw or buffered trigger signal onto one of the input channels. Inspecting the quality of trigger pulses is very useful, as problems often arise for a variety of reasons, including improper placement of the trigger sensor, wiring issues, or poor setup issues such as a dirty piece of reflective tape when using optical sensors. The ability to view either the raw or buffered signal provides an excellent diagnostic or debugging tool. Utilizing the recorded data signal for various signal processing techniques, such as variable-speed filtering algorithms, can also provide improved phase analysis capabilities.

[0094] Embodiments of the methods and systems disclosed herein may include using higher input oversampling for the delta-sigma A / D for lower sampling rate outputs to minimize AA filter requirements. In embodiments, for lower sampling rate output data, a higher input oversampling rate for the delta-sigma A / D is used to minimize AA filter requirements. For higher sampling rates, a lower oversampling rate can be used. For example, for an Fmax range of 200 Hz and 500 Hz, a third-order AA filter set with a minimum sampling requirement of 256 Hz (Fmax of 100 Hz) is appropriate. Then, for an Fmax range of 1 kHz or higher (where the second-order filter starts at 2.56 times (2.56x) the highest sampling rate of 128 kHz), another high-cutoff AA filter can be used. Embodiments of the methods and systems disclosed herein may include using a CPLD as a clock divider for the delta-sigma A / D to achieve lower sampling rates without the need for digital resampling. In an embodiment, a high frequency crystal reference can be divided down to a lower frequency by using a CPLD as a programmable clock divider. The accuracy of the divided down frequency is more accurate than the original source for those longer time periods. This also minimizes or eliminates the need for resampling by a delta-sigma A / D.

[0095] Embodiments of the methods and systems disclosed herein may include signal processing firmware / hardware. In embodiments, long blocks of data are acquired at a high sampling rate, as opposed to multiple data sets acquired at different sampling rates. In general, modern route collection for vibration analysis typically involves collecting data at a specified data length and a fixed sampling rate. The sampling rate and data length may vary from route point to route point based on the specific mechanical analysis requirements at hand. For example, a motor may require a relatively low sampling rate with high resolution to distinguish harmonics of the running speed from harmonics of the line frequency. However, the practical tradeoff here is that it takes more collection time to achieve this improved resolution. In contrast, some high-speed compressors or gear sets require a much higher sampling rate to measure the amplitude of relatively high-frequency data, but precise resolution may not be necessary. However, ideally, it would be better to collect data at a very high sampling rate and with a very long sample length. When digital acquisition devices began to become popular in the early 1980s, A / D sampling, digital storage, and computing power were not nearly what they are today, so compromises were made between the time required for data acquisition and the desired resolution and accuracy. These limitations sometimes prevented field analysts from abandoning analog tape-based recording systems, which did not suffer from many of the same digitization drawbacks. Some hybrid systems were also employed, in which recorded analog data was digitized for playback at multiple sampling rates and lengths as desired, but these systems were not obviously automated. As noted above, a more common approach is to balance data acquisition time with analysis capabilities by digitally acquiring blocks of data at multiple sampling rates and lengths and digitally storing these blocks separately. In an embodiment, long data lengths are stored at the highest practical sampling rate (e.g., 102.4 kHz).The data can be collected and stored at a frequency of 10 kHz (corresponding to an Fmax of 40 kHz). This long data block can be acquired in the same amount of time as a shorter one at a lower sampling rate used in conventional methods, so no significant delay is added to the sampling at the measurement points, which are always related to root collection. In embodiments, the analog tape-style recording of data is digitally simulated with such accuracy that for purposes of embodiments of the present disclosure, it can be considered substantially continuous or "analog" for many applications, unless the context indicates otherwise.

[0096] Embodiments of the methods and systems disclosed herein can include storage of calibration data and maintenance history for onboard card sets. Many data acquisition devices that rely on interfacing with a PC to function store their calibration coefficients on the PC. This is especially true for complex data acquisition devices, where calibration tables can become very large due to the many signal paths. In embodiments, the calibration coefficients are stored in flash memory, which for all practical purposes permanently stores this data and other important information about the device. This information may include nameplate information, such as individual component serial numbers, firmware or software version numbers, maintenance history, and calibration tables. In embodiments, the DAQ box remains calibrated and retains all of this important information, regardless of which computer the box is ultimately connected to. A PC or external device can poll this information at any time for embedding or information exchange purposes.

[0097] Embodiments of the methods and systems disclosed herein may include rapid route generation capabilities utilizing hierarchical templates. In the fields of vibration monitoring and parameter monitoring in general, the existence of data monitoring points needs to be organized into a database or functional equivalent. These points have various attributes associated with them, including transducer attributes, data collection settings, machine parameters, and operational parameter categories. Transducer attributes include probe type, probe attachment type, and probe attachment orientation or axial direction. Measurement-related data collection attributes include sampling rate, data length, integrated electronic piezoelectric (IEPE) probe power and coupling requirements, hardware integration requirements, 4-20 mA (4-20) or voltage interface, applicable range and gain settings, filter requirements, etc. Machine parameter requirements for specific points include items such as operating speed, bearing type, and bearing parameter data for rolling bearings, including pitch diameter, number of balls, inner ring diameter, and outer ring diameter. For tilting pads supporting the bearings, this includes the number of pads, etc. For some measurement points on equipment such as gearboxes, required parameters include, for example, the number of teeth on each gear. For induction motors, this includes the number of rotors and the number of poles. For belt / pulley systems, this includes the number of belts and the associated belt pass frequency, which can be calculated from the pulley dimensions and center-to-center distance. For measurements near the coupling, this includes the type of coupling and the number of teeth required for the gear coupling, etc. Operating parameter data includes the operating load, expressed as megawatts, flow rate (either air or fluid), percentage, horsepower, feet per minute, etc. Operating temperature, pressure, humidity, both ambient and operational, may also be relevant. As can be seen, the setup information required for each measurement point can be quite large. It is also important to perform a proper analysis of the data. Machine-, equipment-, and bearing-specific information is essential to identifying failure frequencies and predicting the various types of specific failures that can be expected. Transducer attributes and data collection parameters are essential to properly interpreting the data, in addition to providing limitations on the type of analysis technique appropriate.Traditional methods of entering this data are manual and very tedious, typically at the lowest levels of hierarchy—for example, at the bearing level for machine parameters or the transducer level for data collection configuration information. However, it cannot be emphasized enough how important the hierarchical relationships required to organize the data are not only for data storage and movement, but also for data analysis and interpretation purposes. This discussion focuses primarily on data storage and movement. By its nature, the configuration information described above is highly redundant at the lowest levels of hierarchy. However, its strong hierarchical nature allows it to be stored very efficiently in that format. In embodiments, this hierarchical nature can be exploited when copying data in the form of templates. As an example, a hierarchical storage structure suitable for many applications is defined for companies, plants or sites, units or processes, machines, equipment, shaft elements, bearings, and transducers, from the general to the specific. Copying data related to a specific machine, piece of equipment, shaft element, or bearing is much easier than copying only at the lowest transducer level. In embodiments, the system not only stores data in this hierarchical manner, but also provides robust support for fast copying of data using these hierarchical templates. The similarity of elements at a particular hierarchical level lends itself to effective data storage in a hierarchical format. For example, many machines share common elements such as motors, gearboxes, compressors, belts, fans, etc. More specifically, many motors can be easily classified as induction, DC, fixed speed, or variable speed. Many gearboxes can be grouped into common found groups such as input / output, input pinion / intermediate pinion / output pinion, 4-posters, etc. Within a factory or company, there are many similar types of equipment that are purchased and standardized for both cost and maintenance reasons. This results in a huge amount of duplication of similar types of equipment, resulting in a great opportunity to utilize a hierarchical template approach.

[0098] Embodiments of the methods and systems disclosed herein can include a smart band. A smart band references processed signal characteristics derived from a dynamic input or group of inputs in order to analyze the data and achieve a correct diagnosis. Additionally, a smart band can include small or relatively simple diagnostics in order to achieve more robust and complex diagnostics. Historically, in the field of mechanical vibration analysis, alarm bands have been used to define important spectral frequency bands in order to analyze and / or trend important vibration patterns. Alarm bands typically consist of spectral (amplitude plotted against frequency) regions defined by boundaries between low and high bands. The amplitudes between those boundaries are summed in the same way that all amplitudes are calculated. Smart bands are more flexible in that they can reference not only specific frequency bands but also groups of spectral peaks, such as single-peak harmonics, true peak levels or crest factors derived from time waveforms, vibration envelope spectra or other specialized signal analysis techniques, or logical combinations (e.g., AND, OR, XOR) of these signal attributes. Additionally, a wide variety of other parametric data can be used as well, including system load, motor voltage and phase information, bearing temperature, flow rate, etc., as a basis for creating additional smart bands. In embodiments, the symptoms of the smart band can be used as building blocks of an expert system, whose engine utilizes these inputs to derive diagnoses. Some of these mini-diagnoses may be used as symptoms of the smart band (which can include diagnoses) for more generalized diagnoses.

[0099] Embodiments of the methods and systems disclosed herein can include a neural net expert system using a smart band. Typical vibration analysis engines are rule-based, i.e., they use a list of expert rules that, when met, trigger a particular diagnosis. In contrast, a neural approach utilizes weighted triggering of multiple input stimuli to smaller analysis engines or neurons, which then provide simplified weighted outputs to other neurons. The outputs of these neurons are also classified as smart bands, which are then provided to other neurons. This results in a more layered approach to expert diagnosis, as opposed to the one-off approach of rule-based systems. In embodiments, the expert system utilizes this neural approach using a smart band, but this does not preclude rule-based diagnoses from being reclassified as smart bands as additional stimuli utilized by the expert system. From this perspective, it can be broadly described as a hybrid approach, although at the highest level it is essentially neural.

[0100] Embodiments of the methods and systems disclosed herein can include the use of a database hierarchy in the analysis. Smart band verification methods and diagnostics can be assigned to various hierarchical database levels. For example, a smart band may call for "looseness" at the bearing level, which causes "looseness" at the device level, which causes "looseness" at the machine level. Another example would be to perform a smart band diagnostic called "Horizontal Plane Phase Flip" across the coupling and a smart band diagnostic for "Vertical Coupling Misalignment" at the machine level.

[0101] Embodiments of the methods and systems disclosed herein can include an expert system GUI. In embodiments, the system implements a graphical approach to defining smart bands and diagnoses for the expert system. Entering verification methods, rules, or more generally smart bands to create specific machine diagnoses can be lengthy and time-consuming. One way to make the process more convenient and efficient is to provide a graphical means of utilizing connectivity. The proposed graphical interface consists of four main components: a symptom parts bin, a diagnosis bin, a tools bin, and a graphical connectivity area ("GWA"). In embodiments, the verification method parts bin includes various spectrum, waveform, envelope, and any type of signal processing characteristics or characteristic groups, such as spectral peaks, spectral harmonics, waveform true peaks, waveform crest factors, spectral alarm bands, etc. Additional characteristics can be assigned to each part. For example, a spectral peak part can be assigned a frequency or order (multiple) of execution speed. Some parts are predefined or user defined, such as 1x, 2x, 3x execution speed, 1x, 2x, 3x gear mesh, 1x, 2x, 3x blade path, number of motor rotors x execution speed.

[0102] In embodiments, the diagnostic binary includes various predefined and user-defined diagnoses, such as misalignment, unbalance, looseness, and bearing fault. Similar to parts, diagnoses can also be used as parts to create more complex diagnoses. In embodiments, the tool binary includes logical operations such as AND, OR, and XOR, or other methods of combining the various parts, such as maximum and minimum search, interpolation, averaging, and other statistical operations. In embodiments, the graphical connection area includes parts from the part binary or diagnoses from the diagnostic binary, which can be combined using the tool to create a diagnosis. The various parts, tools, and diagnoses are represented by icons that are simply graphically connected in a desired manner. Embodiments of the methods and systems disclosed herein can include an expert system GUI graphical approach for defining smart bands and diagnoses for an expert system. Entering validation methods, rules, or more generally smart bands, to create a specific machine diagnosis can be lengthy and time-consuming. One way to make the process more convenient and efficient is to provide a graphical means of using connections. In an embodiment, the graphical interface consists of four main components: a verification method part binary, a diagnostic binary, a tool binary, and a graphical binding area ("GWA"). The verification method part binary consists of various spectrum, waveform, envelope, and any type of signal processing characteristics or characteristic groups, such as spectral peaks, spectral harmonics, waveform true peaks, waveform crest factors, spectral alarm bands, etc. Additional characteristics can be assigned to each part; for example, a spectral peak part can be assigned a frequency or order (multiple) of the execution speed. Some parts are pre-defined or user-defined, such as 1x, 2x, 3x the execution speed, 1x, 2x, 3x the gear mesh, 1x, 2x, 3x the blade path, or number of motor rotors x execution speed.Diagnostic binaries are composed of various predefined and user-defined diagnoses, such as misalignment, unbalance, looseness, bearing fault, etc. Similar to parts, diagnoses can also be used as parts to create more complex diagnoses. Tool binaries are composed of logical operations such as AND, OR, XOR, or other methods of combining the various parts listed above, such as maximum and minimum search, interpolation, averaging, and other statistical operations. GWAs are generally composed of parts from part binaries or diagnoses from diagnostic binaries, which can be combined using tools to create diagnoses. The various parts, tools, and diagnoses are represented by icons that can be simply graphically connected in any desired manner.

[0103] Embodiments of the methods and systems disclosed herein can include a graphical approach for backcalculation definition. In embodiments, the expert system also provides an opportunity for the system to learn. If a unique set of stimuli or smart band is known to correspond to a particular disorder or diagnosis, it is possible to backcalculate a set of coefficients that, when applied to a future set of similar stimuli, will lead to the same diagnosis. In embodiments, a best-fit approach can be used when multiple data sets exist. Unlike the smart band GUI, this embodiment self-generates the binding diagram. In embodiments, a user can adjust backpropagation settings and use a database browser to match a particular data set with a desired diagnosis. In embodiments, a desired diagnosis can be created or custom-tuned using the smart band GUI. In embodiments, the user can then press the Create button, and a dynamic binding of confirmation methods to diagnoses can appear on the screen as the algorithm works to achieve the best fit. In embodiments, upon completion, various statistics are presented detailing how well the mapping process went. In some cases, the mapping cannot be achieved, for example, if the input data is all zeros or incorrect (misassigned) data. Embodiments of the methods and systems disclosed herein can include bearing analysis techniques, which in embodiments can be used in combination with computer aided design ("CAD"), predictive deconvolution, minimum variance distortion response ("MVDR"), and spectrum sum-of-harmonics.

[0104] Embodiments of the methods and systems disclosed herein can include torsional vibration detection and analysis using transient signal analysis. In recent years, there has been a significant trend toward the widespread use of variable-speed machinery. Due primarily to the reduction in the cost of motor speed control systems and the increasing cost and awareness of energy usage, leveraging the enormous energy savings potential of load control has become more economically justified. Unfortunately, an often-overlooked design aspect of this problem is vibration. When a machine is designed to operate at only one speed, it is much easier to design the physical structure to avoid both structural and torsional mechanical resonances, which can dramatically shorten the machine's mechanical integrity. This includes structural characteristics such as the type of material used, its weight, stiffening member requirements and placement, bearing type, bearing location, base support constraints, etc. Even for machines that operate at one speed, designing the structure to minimize vibration can be a daunting task, potentially requiring computer modeling, finite element analysis, and field testing. Mixing various speeds often makes it impossible to design for all desired speeds. This problem is one that can be minimized, for example, by speed avoidance. This is why many modern motor controllers are typically programmed to skip or quickly pass through certain speed ranges or bands. Embodiments can include speed range identification in vibration monitoring systems. Non-torsional structural resonances are typically fairly easy to detect using traditional vibration analysis techniques. However, this is not the case for torsion. One particular area of ​​current focus is the increased incidence of torsional resonance problems, apparently due to increased torsional stresses with speed changes and operation of equipment at torsional resonance speeds. Unlike non-torsional structural resonances, whose effects are generally manifested by dramatic increases in casing or external vibrations, torsional resonances generally do not exhibit such effects. In the case of shaft torsional resonance, the torsional motion induced by the resonance can only be identified by looking for changes in speed and / or phase. The current standard technique for analyzing torsional vibration involves the use of specialized instrumentation.The methods and systems disclosed herein enable the analysis of torsional vibrations without such specialized instrumentation. This involves stopping the machine and using strain gauges and / or other specialized fixtures, such as speed encoder plates and / or gears. Friction wheels are another alternative, but typically require manual implementation and specialized analysts. Generally, these techniques can be very expensive and / or inconvenient. Due to reduced costs and increased convenience (e.g., remote access), continuous vibration monitoring systems are becoming increasingly popular. In embodiments, the vibration signal alone provides the ability to identify torsional speed and / or phase changes. In embodiments, transient analysis techniques can be utilized to distinguish torsional-induced vibrations from simple speed changes due to process control. In embodiments, identification factors may focus on one or more of the following aspects: the rate of speed change by the variable speed motor control is relatively slow, sustained, and purposeful; torsional speed changes tend to be short, impulsive, and not sustained; torsional speed changes tend to be oscillatory, perhaps exponentially decaying, while process speed changes are not; and monitoring phase behavior suggests that smaller speed changes related to torsion relative to shaft rotational speed indicate fast or transient bursts of speed, as opposed to slow phase changes historically related to the slope of machine speed increases and decreases (as shown in Bode or Nyquist plots).

[0105] Embodiments of the methods and systems disclosed herein can include improved integration using both analog and digital methods. When integrating a signal digitally using software, essentially the low-end frequency data of the spectrum has its amplitude multiplied by a function that swells rapidly as it approaches zero, creating what is known in the industry as a "ski-slope" effect. The ski-slope amplitude is essentially the noise floor of the device. A simple solution to this is a conventional hardware integrator, which can operate with a much higher signal-to-noise ratio than the already digitized signal. The amplification factor can also be limited to a reasonable level so that multiplication by very large numbers is essentially prohibited. However, at increasingly higher frequencies, the original amplitude, which may be significantly above the noise floor, is multiplied by a very small number (1 / f) that is significantly below the noise floor. While the hardware integrator has a fixed noise floor, the now low-amplitude high-frequency data does not have its bottom scaled down. In contrast, the same digital multiplication of a digitized high-frequency signal scales the noise floor proportionally. In embodiments, hardware integration can be used below the unity gain point (usually at a value determined by units and / or a desired signal-to-noise ratio based on gain) to produce ideal results, and software integration can be used above the unity gain value. In embodiments, this integration is performed in the frequency domain. In embodiments, the resulting hybrid data can be back-converted into a waveform with a much better signal-to-noise ratio when compared to either the hardware-integrated data or the software-integrated data. In embodiments, the advantages of hardware integration are used in conjunction with the advantages of digital-software integration to achieve the maximum signal-to-noise ratio. In embodiments, a first-order gradual hardware integrator high-pass filter, along with curve fitting, can pass relatively low frequency data to reduce or eliminate noise, allowing for very useful analytical data where steeper filters would be compromised and salvaged.

[0106] Embodiments of the methods and systems disclosed herein can include adaptive scheduling techniques for continuous monitoring. Continuous monitoring is often performed with an up-front multiplexer, where the goal is to select a few data channels from many to provide hardware signal processing, A / D, and DAQ system processing components. This is primarily driven by practical cost considerations. The tradeoff is that not all points are continuously monitored (though alternative hardware approaches would monitor less). In embodiments, multiple scheduling levels are provided. In embodiments, at the lowest level, which is continuous for the most part, all measurement points are cycled through in a round-robin fashion. For example, if it takes 30 seconds to capture and process a measurement point and there are 30 points, each point is processed once every 15 minutes. However, points that warrant alarms based on user-selected criteria can be prioritized and processed more frequently. Because there can be multiple severities per alarm, multiple levels of prioritization for monitoring can be established. In embodiments, more severe alarms are monitored more frequently. In embodiments, many additional high-level signal processing techniques can be applied less frequently. Embodiments may take advantage of the increased processing power of a PC, allowing the PC to temporarily suspend the round-robin route collection process (with its multiple collection layers) and stream the required amount of data for the selected point. Embodiments may include various advanced processing techniques, such as envelope processing, wavelet analysis, and many other signal processing techniques. In embodiments, after acquiring this data, the DAQ card set continues its route at the interrupted point. In embodiments, various PC-scheduled data acquisitions follow their own schedules, less frequently than the DAQ card routes. They can be set hourly or daily by route cycle count (e.g., once every 10 cycles), and can also be incremental, like scheduling, based on their alarm severity priority or the type of measurement (e.g., motors are monitored differently than fans).

[0107] Embodiments of the methods and systems disclosed herein can include data acquisition parking capabilities. In embodiments, a data collection box used for route collection, real-time analysis, and generally as a collection device can be detached from its PC (tablet or otherwise) and powered by an external power source or suitable batteries. In embodiments, the data collector still retains continuous monitoring capabilities, and its on-board firmware can perform dedicated monitoring functions over time or can be remotely controlled for further analysis. Embodiments of the methods and systems disclosed herein can include expanded statistical capabilities for continuous monitoring.

[0108] Embodiments of the methods and systems disclosed herein can include ambient sensing plus local sensing plus vibration for analysis. In embodiments, ambient temperature and pressure, sensed temperature and pressure, can be combined with long-term / medium-term vibration analysis to predict any range of conditions or characteristics. In advanced versions, infrared sensing, infrared thermography, ultrasound, and many other types of sensors and input types can be added in combination with vibration or with each other. Embodiments of the methods and systems disclosed herein can include smart routes. In embodiments, the software of the continuous monitoring system adapts / adjusts the data collection sequence based on statistics, analytics, data alarms, and dynamic analysis. Typically, routes are set based on the channels to which sensors are attached. In embodiments, a multiplexer can couple any input multiplexer channel to (e.g., eight) output channels using a crosspoint switch. In embodiments, when a channel goes into alarm or the system identifies key deviations, the normal route set by the software is suspended, and specific synchronized data is collected from channels that share significant statistical changes for more advanced analysis. In embodiments, smart ODSs or smart transfer functions are implemented.

[0109] Embodiments of the methods and systems disclosed herein can include a smart ODS and one or more transfer functions. In embodiments, the system's multiplexer and crosspoint switch can perform ODS, transfer functions, or other specialized tests on all vibration sensors attached to a machine / structure, accurately showing how machine points move in relation to one another. In embodiments, 40-50 kHz and longer data lengths (e.g., at least one minute) can be streamed, potentially revealing information different from that shown by a typical ODS or transfer function. In embodiments, the system can decide to use a smart route function based on data / statistics / analysis, which deviates from the standard route to perform an ODS on a machine, structure, or between multiple machines and structures, and the ODS can show correlations as conditions / data dictate. In embodiments, for the transfer function, an impact hammer is used on one channel and compared to other vibration sensors on the machine. In embodiments, the system can implement the transfer function using changes in conditions, such as load, speed, temperature, or other changes in the machine or system. In embodiments, various transfer functions can be compared to each other over time. In embodiments, various transfer functions can be strung together like a movie that can show how a machine fault changes, such as a bearing that can show how it moves through four stages of bearing failure. Embodiments of the methods and systems disclosed herein can include a hierarchical multiplexer. In embodiments, the hierarchical multiplexer allows for modular output of more channels, such as 16, 24, or even more multiples of an 8-channel card set, thereby allowing for the collection of more data-synchronous channels for more complex analysis and faster data collection. Disclosed herein are methods and systems for continuous ultrasonic monitoring, including providing continuous ultrasonic monitoring of rotating elements and bearings in energy production equipment.

[0110] Referring to FIG. 8 , the present disclosure generally involves digitally collecting or streaming waveform data 2010 from a machine 2020 whose operating speeds may vary from relatively slow rotational or oscillating speeds to much higher speeds under various circumstances. The waveform data 2010 on at least one machine may include data from a single-axis sensor 2030 installed at a fixed reference location 2040 and data from a three-axis sensor 2050 installed at varying locations (or at multiple locations) including location 2052. In an embodiment, the waveform data 2010 may be vibration data acquired simultaneously from each sensor 2030, 2050 in a gap-free format over multiple minutes using a maximum resolution frequency large enough to capture periodic and transient impact events. As an example of this, the waveform data 2010 may include vibration data that can be used to generate operating deflection profiles. It may also be used to diagnose vibrations from which machine repair solutions can be defined, if desired.

[0111] In an embodiment, the machine 2020 may further include a housing 2100 containing a drive motor 2110 capable of driving a shaft 2120. The shaft 2120 may be supported for rotation or oscillation by a set of bearings 2130, such as a first bearing 2140 and a second bearing 2150. A data collection module 2160 may be connected to (or resident on) the machine 2020. In one example, the data collection module 2160 may be accessible via a cloud network facility 2170 to collect waveform data 2010 from the machine 2020 and distribute the waveform data 2010 to a remote location. The working end 2180 of the drive shaft 2120 of the machine 2020 may drive a windmill, fan, pump, drill, gear system, drive system, or other working element, and the techniques described herein may be applied to a wide range of machines, devices, equipment, or the like, including rotating or oscillating elements. In another example, a generator can be used in place of the motor 2110, and the working end of the drive shaft 2120 can deliver rotational energy to the generator, generating power rather than consuming it.

[0112] In an embodiment, the waveform data 2010 can be acquired using a predetermined route format based on the layout of the machine 2020. The waveform data 2010 can include data from a single-axis sensor 2030 and a three-axis sensor 2050. The single-axis sensor 2030 can function as a reference probe using its one data channel and can be fixed to a fixed location 2040 on the machine under investigation. The three-axis sensor 2050 can function as a three-axis probe (e.g., three orthogonal axes) using its three data channels and can be moved from one test point to the next according to a predetermined diagnostic route format. In one example, both sensors 2030, 2050 can be manually attached to the machine 2020 and, in certain application examples, can be connected to separate portable computers. The reference probe can remain in one position while a user can move the three-axis vibration probe along a predetermined route, such as from bearing to bearing on the machine. In this example, the user is instructed to place the sensors in predetermined locations to complete an investigation (or portion thereof) of the machine.

[0113] 9, there is shown a portion of a schematic machine 2200 having a three-axis sensor 2210 mounted at a location 2220 associated with a motor bearing of the machine 2200, the machine having an output shaft 2230 and an output member 2240, in accordance with the present disclosure. Referring to FIGS. 9 and 10, there is shown a schematic machine 2300 having a three-axis sensor 2310 and a single-axis vibration sensor 2320 mounted at a fixed location on the machine 2300 during a vibration study, the single-axis vibration sensor 2320 serving as a reference sensor, in accordance with the present disclosure. The three-axis sensor 2310 and the single-axis vibration sensor 2320 can be connected to a data acquisition system 2330.

[0114] In further examples, sensors and data acquisition modules and devices may be integrated into or resident on the rotating machine. In these examples, a machine may include multiple single-axis sensors and multiple three-axis sensors at predetermined locations. These sensors may be original equipment, provided by the original equipment manufacturer, or retrofitted at a later time. A data collection module 2160 or the like may select and use one single-axis sensor from which to acquire data exclusively while collecting waveform data 2010, moving to each of the three-axis sensors. The data collection module 2160 may be resident on the machine 2020 and / or connected via a cloud network facility 2170.

[0115] Referring to FIG. 8 , various embodiments include collecting waveform data 2010 by locally digitally recording or streaming via cloud network facility 2170. The waveform data 2010 may be collected in an uninterrupted, gap-free manner, similar in some respects to analog recording of waveform data. Depending on the rotational or oscillating speed of the machine being monitored, waveform data 2010 from all channels may be collected over a period of one to two minutes. In embodiments, the sampling rate of the data may be relatively high relative to the operating frequency of the machine 2020.

[0116] In some embodiments, a second reference sensor may be used and a fifth data channel may be collected. Thus, a single-axis sensor may occupy the first channel, and triaxial vibration may occupy the second, third, and fourth data channels. This second reference sensor, like the first sensor, may be a single-axis sensor, such as an accelerometer. In some embodiments, the second reference sensor, like the first reference sensor, may remain in the same location on the machine for an overall vibration study of the machine. The location of the first reference sensor (i.e., the single-axis sensor) may be different from the location of the second reference sensor (i.e., another single-axis sensor). In a particular example, a second reference sensor may be used when a machine has two shafts with different operating speeds, with the two reference sensors located on two different shafts. According to this example, an additional single-axis reference sensor may be used in an additional, but different, permanent location on the rotating machine.

[0117] In embodiments, waveform data can be electronically transmitted in a gap-free, open format at a significantly high sampling rate for a relatively long period of time. In one example, the period is 60 to 120 seconds. In another example, the sampling rate is 100 kHz and the maximum resolution frequency (Fmax) is 40 kHz. In light of this disclosure, it will be appreciated that waveform data can be presented to more closely approximate some of the wealth of data available from analog recording of the waveform data in the previous example.

[0118] In embodiments, sampling, band selection, and filtering techniques can be used to undersample or oversample one or more portions of a long data stream (e.g., 1-2 minutes) to achieve different effective sampling rates. To this end, interpolation and decimation can be used to further achieve different effective sampling rates. For example, oversampling can be applied to a frequency band near the rotational or oscillating operating speed of the sample machine or its harmonics, where vibration effects may tend to be more pronounced at those frequencies across the machine's operating range. In embodiments, the digital sample data set can be decimated to generate a lower sampling rate. It will be understood in light of this disclosure that decimation in this context may be the opposite of interpolation. In embodiments, decimating a data set can include first applying a low-pass filter to the digital sample data set and then undersampling the data set.

[0119] In one example, a waveform sampled at 100 Hz can be undersampled by undersampling every 10th point of the digital waveform to produce an effective sampling rate of 10 Hz, but the remaining 9 points in that portion of the waveform are effectively discarded and not included in modeling the sampled waveform. Furthermore, this type of bare undersampling can produce ghost frequencies due to the rate of undersampling (i.e., 10 Hz) relative to the 100 Hz sampled waveform.

[0120] Most analog-to-digital conversion hardware uses sample-and-hold circuits that can charge a capacitor over a given time period so that the average value of the waveform over that time period can be determined. It will be appreciated in light of this disclosure that the value of the waveform over that time period is not linear, but more similar to a cardinal sinusoidal ("sinc") function, and therefore, more emphasis can be placed on the waveform data at the center of the sampling interval, with the cardinal sine wave signal exponentially decaying from that center.

[0121] Continuing with the above example, a waveform sampled at 100 Hz can be hardware sampled at 10 Hz, and therefore each sample point is averaged over 100 milliseconds (e.g., a signal sampled at 100 Hz can have each point averaged over 10 milliseconds). As opposed to effectively discarding 9 out of 10 data points of a sample waveform as described above, the present disclosure can include weighting of adjacent data. The adjacent data can include a reference to a previously discarded sample point and the remaining point that is retained. In one example, a low-pass filter can linearly average adjacent sample data, i.e., sum every 10 points and divide the sum by 10. In a further example, adjacent data can be weighted with a sinc function. The process of weighting the original waveform with a sinc function can be referred to as an impulse function, or in the time domain, as convolution.

[0122] The present disclosure is applicable not only to digitizing waveform signals based on detected voltages, but also to digitizing waveform signals based on image processing signals, including current waveforms, vibration waveforms, and rasterization of video signals. In one example, the resizing of a window on a computer screen can be decimated in at least two directions. It will be appreciated that in these further examples, undersampling by itself may prove insufficient. To that end, interpolation can be used like decimation, but oversampling or upsampling by itself may also prove insufficient, as may be used instead of undersampling by itself alone.

[0123] It will be understood in light of this disclosure that interpolation in this context may refer to first applying a low-pass filter to digitally sampled waveform data and then upsampling the waveform data. It will be understood in light of this disclosure that real-world examples may frequently require the use of non-integer factors for decimation, interpolation, or both. To that end, this disclosure includes sequential interpolation and decimation to achieve non-integer factor rates for interpolation and decimation. In one example, sequential interpolation and decimation may be defined as applying a low-pass filter to a sample waveform, interpolating the waveform after the low-pass filter, and decimating the waveform after the interpolation. In embodiments, vibration data may be looped to intentionally emulate a traditional tape recording loop, and digital filtering techniques, along with effective splicing, may be used to facilitate longer analysis. It will be understood in light of this disclosure that the above techniques do not preclude waveform, spectral, and other types of analysis to be processed and displayed in a user GUI as they are collected. It will be appreciated in light of this disclosure that newer systems are capable of performing this functionality in parallel with high performance acquisition of raw waveform data.

[0124] Regarding the issue of acquisition time, it will be appreciated that older systems that use a compromised approach of improving data resolution by acquiring at various sampling rates and data lengths do not actually save time as expected. Therefore, latency issues can arise each time the data acquisition hardware is stopped and started, especially if the hardware is auto-scaling. The same is true for data retrieval of path information (i.e., test locations), which is often in database format and can be very slow. Storing raw data in bursts to disk (solid-state or not) can be uncomfortably slow.

[0125] In contrast, many embodiments, as disclosed herein, involve digitally streaming waveform data 2010 while simultaneously benefiting from the need to load path parameter information while configuring data acquisition hardware only once. Because waveform data 2010 is streamed to only one file, there is no need to open and close files or switch between read and write operations on storage media. It can be seen that collecting and storing waveform data 2010 as described herein produces relatively meaningful data in a much shorter time than traditional batch data acquisition approaches. An example of this includes an electric motor, where waveform data can be collected at a data length of 4K points (i.e., 4,096), which is high enough resolution to distinguish electrical sideband frequencies, among other things. For fans or blowers, a reduced resolution of 1K (i.e., 1,024) can be used. In some cases, 1K is the minimum waveform data length requirement. The sampling rate is 1,280 Hz, which corresponds to an Fmax of 500 Hz. It will be appreciated in light of this disclosure that oversampling by the industry standard factor of 2.56 satisfies the 2x oversampling required by the Nyquist criterion, with additional margin to accommodate anti-aliasing filter roll-off. The time to acquire this waveform data is 800 milliseconds for 1,024 points at 1,280 Hz.

[0126] To improve accuracy, waveform data can be averaged. For example, eight averages with 50% overlap can be used. This extends the time from 800 ms to 3.6 seconds, which is equal to 800 ms x 8 averages x 0.5 (overlapping) + 0.5 x 800 ms (non-overlapping leading and trailing). After waveform data collection at Fmax = 500 Hz, a higher sampling rate can be used. As an example, ten times (10x) the previous sampling rate can be used, with Fmax = 10 kHz. In this example, waveform data can be collected at this higher rate using eight averages with 50% overlap, resulting in a collection time of 360 ms, or 0.36 seconds. It will be appreciated in light of this disclosure that it is necessary to enable automatic scaling of the hardware as well as read the hardware collection parameters for the higher sampling rate from the route list and / or reset other necessary hardware collection parameters. To do this, a waiting period of a few seconds can be added to accommodate the change in sampling rate. In another example, the wait time can be inserted to accommodate hardware autoscaling and changes in hardware acquisition parameters that are required when using the lower sampling rates disclosed herein. In addition to adapting to the change in sampling rate, additional time is required to read route point information (i.e., where to monitor and where to monitor next) from the database, display the route information, and process the waveform data. Furthermore, displaying the waveform data and / or associated spectra can also consume significant time. With this in mind, 15 to 20 seconds may elapse between acquiring waveform data at each measurement point.

[0127] In further examples, additional sampling rates can be added, which may further increase the total time for the vibration survey due to the time required to switch from one sampling rate to another and to acquire additional data at a different sampling rate. In one example, a lower sampling rate, such as a 128 Hz sampling rate with Fmax = 50 Hz, is used. In this example, the vibration survey requires an additional 36 seconds for the first set of averaged data at this sampling rate, in addition to the others mentioned above, resulting in a more dramatic increase in the total time spent at each measurement point. Further embodiments include using gap-free digital streaming of waveform data similar to that disclosed herein for use with wind turbines and other machines that may have relatively slow rotating or oscillating systems. In many examples, the collected waveform data may include long sample data at relatively long sampling rates. In one example, the sampling rate may be 100 kHz, and the sampling period may be 2 minutes for all channels being recorded. In many examples, one channel may be for a single-axis reference sensor and three additional data channels may be for a three-axis, three-channel sensor. It will be understood in light of the present disclosure that long data lengths may be provided to facilitate detection of very low frequency phenomena. Long data lengths may also be provided to accommodate speed variability inherent in wind turbine operation. Additionally, long data lengths may also be provided to achieve very high spectral resolution, provide the opportunity to use multiple averages as discussed herein, and to create suitable tape loops for specific spectral analysis. Many diverse and sophisticated analysis techniques become available because such techniques can use continuous lengths of waveform data of useful lengths in accordance with the present disclosure.

[0128] In light of the present disclosure, it will also be appreciated that collecting waveform data from multiple channels simultaneously can facilitate the implementation of transfer functions between the multiple channels. Furthermore, collecting waveform data from multiple channels simultaneously facilitates the establishment of phase relationships across the machine, so that a greater degree of correlation can be exploited by relying on the fact that waveforms from each channel are collected simultaneously. In another example, by enabling the simultaneous acquisition of waveform data from multiple sensors that would otherwise be acquired in a serial manner moving from sensor to sensor in a vibration survey, more channels in the data collection can be used to reduce the time required to complete the entire vibration survey.

[0129] The present disclosure includes using at least one uniaxial reference probe in one of the channels to enable relative phase comparison between channels. The reference probe may be an accelerometer or other type of transducer that is fixed in a fixed position and does not move during a vibration study of a machine. Multiple reference probes may be placed in suitable fixed locations (i.e., fixed positions) while acquiring vibration data during a vibration study. In a specific example, up to seven reference probes may be placed, depending on the capacity of the data collection module 2160, etc. Using transfer functions or similar techniques, the relative phase of all channels can be compared to each other at all selected frequencies. By keeping one or more reference probes fixed in their fixed positions while moving or monitoring other triaxial vibration sensors, the entire machine can be mapped in terms of amplitude and relative phase. This can be properly illustrated even when there are more measurement points than channels of data collection. This information can be used to create a motion deflection shape that can show the dynamic behavior of the machine in 3D, providing a valuable diagnostic tool. In an embodiment, one or more reference probes can provide relative phase rather than absolute phase. It will be appreciated in light of this disclosure that although relative phase may not be as meaningful as absolute phase for some purposes, relative phase information can still prove very useful.

[0130] In embodiments, the sampling rate used during the vibration study can be digitally synchronized to a predetermined operating frequency, which can be related to a suitable parameter of the machine, such as rotational or oscillation speed. Doing so allows for more information to be extracted using synchronous averaging techniques. It will be appreciated in light of this disclosure that this can be done without using key phasors or reference pulses from the rotating shaft, which are typically unavailable in route-collected data. Thus, asynchronous signals can be removed from complex signals without the need to deploy synchronous averaging using key phasors. This proves very useful when analyzing a specific pinion in a gearbox, or more generally when applied to components within complex mechanical mechanisms. While key phasors and reference pulses are often unavailable in route-collected data, the techniques disclosed herein can overcome this lack. In embodiments, there may be multiple shafts operating at various speeds within the machine being analyzed. In certain instances, there may be a single-axis reference probe for each axis. In other instances, only one single-axis reference probe at a fixed position on one shaft may be used to relate the phase of one shaft to another. In embodiments, variable speed machines can be more easily analyzed with relatively long data durations compared to single speed machines. Vibration studies can be performed at several machine speeds within the same continuous vibration data set using the same techniques disclosed herein. These techniques can also enable the study of changing relationships between vibration and speed changes that were not previously available.

[0131] In embodiments, as disclosed herein, raw waveform data can be captured in a gap-free digital format, enabling numerous analytical techniques. The gap-free digital format can facilitate numerous means for analyzing waveform data in numerous ways after identifying a specific problem. Vibration data collected in accordance with the techniques disclosed herein can provide analysis of transient, semi-periodic, and very low frequency phenomena. Waveform data acquired in accordance with the present disclosure can include relatively long streams of raw gap-free waveform data that can be conveniently replayed as needed and upon which many diverse and sophisticated analytical techniques can be performed. Many of these techniques can provide various forms of filtering to extract low amplitude modulations from the transient impulse data that can be contained in the relatively long streams of raw gap-free waveform data. It will be understood in light of the present disclosure that in previous data collection practices, these types of phenomena were typically lost in the averaging process of spectral processing algorithms because the goal of previous data acquisition modules was pure periodic signals, or in file size reduction techniques because much of the content from the original raw signal was typically discarded when it was determined that it would not be used.

[0132] In one embodiment, a method for monitoring vibrations of a machine having at least one shaft supported by a set of bearings is provided. The method includes monitoring a first data channel assigned to a single-axis sensor at a fixed position relative to the machine. The method also includes monitoring second, third, and fourth data channels assigned to a three-axis sensor. The method further includes simultaneously recording gap-free digital waveform data from all data channels while the machine is operating and measuring relative phase changes based on the digital waveform data. The method also includes positioning three-axis sensors at multiple positions relative to the machine and acquiring the digital waveforms. In one embodiment, the second, third, and fourth channels are assigned together to a set of three-axis sensors, each positioned at a different position relative to the machine. In one embodiment, data is received simultaneously from all sensors on all channels.

[0133] The method further includes determining a motion deflection shape based on the relative phase information and the change in waveform data. In an embodiment, the invariant position of the reference sensor is a position relative to the shaft of the machine. In an embodiment, each of the three-axis sensors in the series of three-axis sensors is located at a different location and associated with a different bearing within the machine. In an embodiment, the invariant position is a position relative to the shaft of the machine, and each of the three-axis sensors in the series of three-axis sensors is located at a different location and associated with a different bearing supporting the shaft within the machine. Various embodiments include a method for simultaneously and continuously monitoring vibration or similar process parameters and signals of rotating, oscillating, or similar process machines from multiple channels, known as an ensemble. In various examples, an ensemble can include one to eight channels. In a further example, an ensemble can represent a logical measurement group on a device, whether the locations being measured are temporary for the measurement, provided by the original device manufacturer, added later, or one or more combinations thereof.

[0134] In one example, the ensemble can monitor bearing vibration in one direction. In a further example, the ensemble can monitor three different directions (e.g., orthogonal directions) using a triaxial sensor. In yet another example, the ensemble can monitor four or more channels, where a first channel can monitor a uniaxial vibration sensor and second, third, and fourth channels can monitor each of the three directions of the triaxial sensor. In another example, the ensemble can be fixed to a group of adjacent bearings on the same piece of equipment or associated shaft. Various embodiments provide a method including a procedure for collecting waveform data from various ensembles deployed in a vibration test or the like in a relatively efficient manner. Furthermore, the method includes simultaneously monitoring a reference channel assigned to a constant reference position associated with the ensemble monitoring the machine. Coordination with the reference channel can be demonstrated to support more complete correlation of waveforms collected from the ensembles. The reference sensor on the reference channel can be a uniaxial vibration sensor or a phased reference sensor that can be triggered by a reference position, such as a rotating shaft. As disclosed herein, the method may further include simultaneously recording gap-free digital waveform data from all channels of each ensemble during operation of the machine being monitored at a relatively high sampling rate to include all frequencies deemed necessary for proper analysis of the machine. Data from the ensembles may be streamed gap-free to a storage medium for subsequent processing, which may be connected to a cloud network facility, a local data link, a Bluetooth connection, a cellular data connection, etc.

[0135] In embodiments, the methods disclosed herein include procedures for collecting data from various ensembles, including digital signal processing techniques that can subsequently be applied to the data from the ensembles to emphasize or favorably isolate specific frequency or waveform phenomena. This may contrast with current methods of collecting multiple data sets at different sampling rates or with different hardware filtering configurations, including integration, that offer relatively little flexibility in post-processing due to a commitment to those configurations (known as a priori hardware configurations). These same hardware configurations may also result in longer vibration studies due to the latency delays associated with configuring the hardware for each independent test. In embodiments, the methods for collecting data from various ensembles include data marker techniques that can be used to classify sections of the streamed data into homogeneous groups and assign them to specific ensembles. In one example, the classification may be defined by operating speed. This allows for the creation of multiple ensembles from what conventional systems only collect. Many embodiments include post-processing analysis techniques to compare the relative phase of all frequencies of interest not only between each channel of the collected ensemble, but also, where applicable, between all channels of all ensembles being monitored.

[0136] 12 , many embodiments include a first machine 2400 having a rotating or vibrating component 2410 supported by bearing sets 2420, each of which includes bearing packs 2422, 2424, 2426, and optionally additional bearing packs, or both. The first machine 2400 can be monitored by a first sensor ensemble 2450. The first sensor ensemble 2450 can be configured to receive signals from multiple sensors originally installed (or later added) on the first machine 2400. These sensors on the first machine 2400 can include multiple single-axis sensors 2460, such as single-axis sensors 2462, 2464, and additional single-axis sensors as needed. In many examples, the multiple single-axis sensors 2460 can be positioned at locations on the first machine 2400 that can detect one of the multiple rotating or vibrating components 2410 of the first machine 2400.

[0137] The first machine 2400 may also include multiple three-axis (orthogonal axis) sensors 2480, such as three-axis sensors 2482, 2484, and additional three-axis sensors as needed. In many examples, the three-axis sensors 2480 may be located at locations on the first machine 2400 capable of sensing one of each bearing pack of the bearing set 2420 associated with a rotating or vibrating component of the first machine 2400. The first machine 2400 may also include multiple temperature sensors 2500, such as temperature sensors 2502, 2504, and additional temperature sensors as needed. The first machine 2400 may also include tachometer sensors 2510, or additional tachometer sensors as needed, each precisely detecting the revolutions per minute (RPM) of one of its rotating components. Using the above example, the first sensor ensemble 2450 may interrogate multiple of the aforementioned sensors associated with the first machine 2400. As such, the first sensor ensemble 2450 may be configured to receive eight channels. In other examples, first sensor ensemble 2450 can be configured to have more or less than eight channels, as needed. In this example, the eight channels include two channels capable of monitoring one single-axis sensor signal and three channels capable of monitoring one three-axis sensor signal. The remaining three channels can monitor two temperature signals and a signal from a tachometer. In one example, first sensor ensemble 2450 can monitor one-axis sensors 2462, 2464, a three-axis sensor 2482, temperature sensors 2502, 2504, and a tachometer sensor 2510, all of which are relevant to the present disclosure. During a vibration investigation of first machine 2400, first sensor ensemble 2450 can first monitor three-axis sensor 2482 and then move on to the next three-axis sensor 2484.

[0138] After monitoring the three-axis sensor 2484, the first sensor ensemble 2450 may monitor additional three-axis sensors of the first machine 2400 as needed, which are part of a predetermined route list associated with a vibration investigation of the first machine 2400 in accordance with the present disclosure. During the vibration investigation, the first sensor ensemble 2450 may continualy monitor the one-axis sensors 2462, 2464, the two temperature sensors 2502, 2504, and the tachometer sensor 2510, and the first sensor ensemble 2450 may serially monitor the multiple three-axis sensors 2480 in the predetermined route plan for the vibration investigation.

[0139] 12 , many embodiments include a second machine 2600 having a rotating or vibrating component 2610 supported by bearing sets 2620, each of which includes bearing packs 2622, 2624, 2626 and, optionally, additional bearing packs, or both. The second machine 2600 can be monitored by a second sensor ensemble 2650. The second sensor ensemble 2650 can be configured to receive signals from multiple sensors originally installed (or later added) on the second machine 2600. These sensors on the second machine 2600 can include multiple single-axis sensors 2660, such as single-axis sensors 2662, 2664 and, optionally, additional single-axis sensors. In many examples, the multiple two-axis sensors 2660 can be positioned at a location on the second machine 2600 that can detect one of the multiple rotating or vibrating components 2610 of the second machine 2600.

[0140] The second machine 2600 can also include multiple three-axis (orthogonal axis) sensors 2680, such as three-axis sensors 2682, 2684, 2686, 2688, and additional three-axis sensors as needed. In many examples, the three-axis sensors 2680 can be located on the second machine 2600 at locations capable of sensing one of each bearing pack of the bearing set 2620 associated with a rotating or vibrating component of the second machine 2600. The second machine 2600 can also include multiple temperature sensors 2700, such as temperature sensors 2702, 2704, and additional temperature sensors as needed. The second machine 2600 can also include tachometer sensors 2710, or additional tachometer sensors as needed, each accurately sensing the revolutions per minute (RPM) of one of its rotating components.

[0141] Using the above example, second sensor ensemble 2650 can interrogate multiple sensors associated with second machine 2600. To that end, second sensor ensemble 2650 can be configured to receive eight channels. In other examples, second sensor ensemble 6450 can be configured to have more or less than eight channels, as needed. In this example, the eight channels include one channel that can monitor a reference signal of one single-axis sensor and six channels that can monitor two three-axis sensor signals. The remaining channels can monitor temperature signals. In one example, second sensor ensemble 2650 can monitor one-axis sensor 2662, three-axis sensor 2682, three-axis sensor 2684, and temperature sensor 2702. During vibration interrogation of machine 2600 according to the present disclosure, second sensor ensemble 2650 can first monitor three-axis sensor 2682 simultaneously with three-axis sensor 2684, and then move on to three-axis sensor 2686 and three-axis sensor 2688 simultaneously.

[0142] After monitoring the three-axis sensor 2680, the second sensor ensemble 2650 may monitor additional three-axis sensors (simultaneous pairs) of the machine 2600 as needed, which are part of a pre-defined route list associated with a vibration survey of the machine 2600 in accordance with the present disclosure. During the vibration survey, the second sensor ensemble 2650 may monitor the multiple three-axis sensors sequentially in a pre-defined route plan for the vibration survey, while continuously monitoring the one-axis sensor 2662 at its fixed location and continuously monitoring the temperature sensor 2702.

[0143] 12 , many embodiments include a third machine 2800 having a rotating or oscillating component 2810, or both, each of which is supported by a bearing set 2820 including bearing packs 2822, 2824, 2826 and, optionally, additional bearing packs. The third machine 2800 can be monitored by a third sensor ensemble 2850. The third sensor ensemble 2850 can be comprised of one single-axis sensor 2860 and two three-axis (e.g., orthogonal) sensors 2880, 2882. In many examples, the single-axis sensor 2860 can be secured by a user to a position on the machine 2800 that allows for sensing of one of the rotating or oscillating components of the machine 2800. The three-axis sensors 2880, 2882 can also be placed by a user on the machine 2800 in a position that allows for sensing of one of the bearings in the set of bearings associated with each of the rotating or oscillating components of the machine 2800. The third sensor ensemble 2850 may also include a temperature sensor 2900. The third sensor ensemble 2850 and its sensors are distinct from the first and second sensor ensembles 2450, 2650 and may be transported to other machines.

[0144] Many embodiments also include a fourth machine 2950 having a rotating and / or vibrating component 2960 each supported by a set of bearings 2970, including bearing packs 2972, 2974, 2976, and optionally additional bearing packs. The fourth machine 2950 can be monitored by a third sensor ensemble 2850 by a user moving the third sensor ensemble 2850 to the fourth machine 2950. Many embodiments also include a fifth machine 3000 having a rotating and / or vibrating component 3010. The fifth machine 3000 is not explicitly monitored by any sensor or sensor ensemble during operation, but can generate vibrations or other impact energy of sufficient magnitude to be recorded in data associated with any of the machines 2400, 2600, 2800, 2950 under vibration investigation.

[0145] Many embodiments include monitoring a first sensor ensemble 2450 on a first machine 2400 via a predetermined route as disclosed herein. Many embodiments also include monitoring a second sensor ensemble 2650 on a second machine 2600 via a predetermined route. The location of machine 2400 in proximity to machine 2600 can be included in the contextual metadata for both vibration surveys. A third sensor ensemble 2850 can move between machines 2800, 2950, ​​and other suitable machines. Machine 3000 does not implement sensors as described above, but can be monitored by third sensor ensemble 2850 if desired. Machine 3000 and its operating characteristics can be recorded in metadata in connection with vibration surveys of other machines, recording its contribution due to proximity.

[0146] Many embodiments include hybrid database adaptations that blend relational metadata with streaming raw data formats. Unlike older systems that utilized traditional database structures to associate nameplates and operational parameters (sometimes considered metadata) with discrete and relatively simple individual data measurements, it will be appreciated in light of this disclosure that more modern systems are able to collect much larger amounts of raw streaming data at higher sampling rates and higher resolutions. At the same time, it will be appreciated in light of this disclosure that the network of metadata that links and captures and / or correlates with this raw data is ever-increasing and expanding.

[0147] In one example, a single overall vibration level can be collected as part of a routed or predetermined list of measurement points. This collected data can then be associated with database measurement location information for points located on the surface of a bearing housing on a particular component of the machine adjacent to the coupling in the vertical direction. Mechanical analysis parameters relevant to a suitable analysis can be associated with the points located on the surface of the bearing housing. An example of a mechanical analysis parameter relevant to a suitable analysis can include the running speed of a shaft passing through the measurement points on the surface of the bearing housing. Further examples of mechanical analysis parameters relevant to a suitable analysis can include one or a combination of the running speed of all component shafts of the device and / or machine component, the bearing type being analyzed, such as sleeve or roller bearing, the number of gear teeth if a gearbox is present, the number of motor poles, the motor slip and line frequency, the dimensions of the roller bearing elements, the number of fan blades, etc. Further examples of mechanical analysis parameters relevant to a suitable analysis can include machine operating conditions, such as the load on the machine and whether the load is expressed as a percentage, wattage, air flow rate, head pressure, horsepower, etc. Further examples of mechanical analysis parameters include information related to adjacent machinery that may affect the data obtained during the vibration test.

[0148] It will be appreciated in light of this disclosure that the vast array of equipment and machine types can support many different classifications, each of which can be analyzed in a distinctly different manner. For example, machines such as screw compressors and hammer mills can be displayed as generating more noise and are expected to vibrate more than other machines. Machines known to vibrate more can be displayed as requiring a change in vibration level to be considered acceptable relative to quieter machines.

[0149] The present disclosure further includes hierarchical relationships found in collected vibration data that can be used to aid in proper analysis of the data. One example of hierarchical data includes the interconnections of machine components, such as bearings, measured in a vibration survey, including how the bearing connects to a particular shaft attached to a particular pinion in a particular gearbox, and the relationship between the shaft, pinion, and gearbox. The hierarchical data can further include where a particular point in the machine gear train where the bearing is being monitored is located relative to other components within that machine. The hierarchical data can also detail whether a bearing measured on a machine is in close proximity to another machine whose vibration may affect what is being measured on the machine being vibration tested.

[0150] Analysis of vibration data from bearings or other components relative to one another in hierarchical data can use table lookups, searching for correlations between multiple frequency patterns derived from the raw data, and specific frequencies from the machine's metadata. In some embodiments, this can be stored in and retrieved from a relational database. In some embodiments, the National Instruments Technical Data Management Solution (TDMS) file format can be used. The TDMS file format can be optimized for streaming various types of measurement data (i.e., binary digital samples of waveforms) and can also handle hierarchical metadata.

[0151] Many embodiments include a hybrid relational metadata-binary storage approach (HRM-BSA). The HRM-BSA can include a Structured Query Language (SQL)-based relational database engine. The Structured Query Language-based relational database engine also includes a raw data engine that can be optimized for throughput and storage density of flat, relatively unstructured data. In light of the disclosure, it will be understood that benefits can be demonstrated in collaboration between hierarchical metadata and an SQL relational database engine. In one example, marker technology and pointer indicators can be used to create correlations between the raw database engine and the SQL relational database engine. Three examples of correlations between the raw database engine and the SQL relational database engine linkage include: (1) pointers from the SQL database to the raw data; (2) pointers from auxiliary metadata tables or similar groupings of raw data to the SQL database; and (3) independent storage tables outside the domain of the SQL database or raw data technology.

[0152] 13 , the present disclosure may include pointers for Group 1 and Group 2 containing associated metadata such as associated file names, path information, table names, database key fields employed in existing SQL database technology that can be used to associate specific database segments or locations, asset properties to specific measurement raw data streams, records with associated time / date stamps, or operational parameters, panel conditions, etc. As an example of this, plant 3200 may include machine 1 3202, machine 2 3204, and many other machines. Machine 1 3202 may include gearbox 3210, motor 3212, and other elements. Machine 2 3204 may include motor 3220 and other elements. Many waveforms 3230, including waveform 3240, waveform 3242, waveform 3244, and additional waveforms as needed, may be acquired from machines 1 and 2 3202, 3204 of plant 3200. The waveform 3230 can be associated with a local marker link table 3300 and a link table raw data table 3400. Machines 1, 2 3202, 3204 and their elements can be associated with a link table with a relational database 3500. The link table raw data table 3400 and the link table with the relational database 3500 can be associated with a link table with an optional separate storage table 3600.

[0153] The present disclosure can be applied to time marks or sample lengths within raw waveform data. The present disclosure can include markers that can be applied to time marks or sample lengths within raw waveform data. The markers generally fall into two categories: preset or dynamic. Preset markers can be correlated to preset or existing operating conditions, such as load, head pressure, cubic feet of airflow per minute, ambient temperature, revolutions per minute (RPM), etc. These preset markers can be fed directly into a data acquisition system. In some cases, the preset markers can be collected on a data channel in parallel with the waveform data (e.g., vibration, current, voltage, etc.). Alternatively, the values ​​of the preset markers can be entered manually.

[0154] For dynamic markers, such as trend data, it may be important to compare similar data, such as vibration amplitude and pattern, with a repeatable set of operating parameters. An example of this disclosure involves one of the parallel channel inputs being a key phasor trigger pulse from the operating shaft, which can provide RPM information at the moment of collection. This dynamic marker example allows for marking the appropriate speed or speed range in a section of the collected waveform data.

[0155] The present disclosure may also include dynamic markers that can be correlated to data resulting from post-processing and analysis performed on sampled waveforms. In further embodiments, dynamic markers may also be correlated to other performance-derived metrics, such as collected parameters including revolutions per minute (RPM), as well as alarm conditions, such as maximum RPM. In a specific example, many modern machines that are candidates for vibration investigation using the portable data collection systems described herein do not include tachometer information. This may be true because, even when RPM measurement is of primary importance to vibration investigation and analysis, adding a tachometer is not always practical or cost-justified. For fixed-speed machines, it is understood that obtaining an accurate RPM measurement is less important if the machine's approximate speed can be determined in advance. However, variable-speed drives are becoming increasingly common. In light of the present disclosure, it will also be appreciated that various signal processing techniques allow RPM to be derived from raw data without the need for a dedicated tachometer signal.

[0156] In many embodiments, RPM information can be used to mark segments of raw waveform data in their collection history. Further embodiments include techniques for collecting equipment data following a predetermined path of vibration testing. Dynamic markers can enable analysis and trending software to utilize multiple segments of the collection interval (e.g., two minutes) indicated by the marker as multiple historical collection ensembles, whereas in conventional systems, the route collection system stores historical data for only one RPM setting, which is done only once. This can be similarly extended to other operating parameters, such as load setting, ambient temperature, etc., as previously mentioned. However, dynamic markers placed in a type of index file that points to the raw data stream can group portions of the stream in a similar manner that can be more easily compared to portions of previously collected raw data streams.

[0157] Many embodiments include a hybrid relational metadata-binary storage approach that can utilize the best of existing technology for both relational and raw data streams. In some embodiments, the hybrid relational metadata-binary storage approach can combine them with various marker linkages. Marker linkages can enable rapid searching through relational metadata and more efficient analysis of raw data using conventional SQL techniques with existing technology. This can enable many features, linkages, compatibility, and extensions not offered by conventional database technologies.

[0158] Marker linkage can also enable rapid and efficient storage of raw data using conventional binary storage and data compression techniques. This can be shown to enable leveraging many of the features, linkages, compatibility, and extensions offered by conventional raw data technologies such as TMDS (National Instruments) and UFF (Universal File Format, e.g., UFF58). Marker linkage can further enable the use of marker technology links, which can accumulate rich data sets from ensembles in the same collection time as conventional systems. Richer data sets from ensembles can be stored as data snapshots related to predetermined collection criteria, and the proposed system can derive multiple snapshots from collected data streams using marker technology. Doing so can enable relatively rich analysis of collected data to be achieved. One such advantage is the ability to include more trend points of vibrations in RPM, load, operating temperature, flow rate, etc. for a particular frequency or order of operating speed, which can be collected in a similar time to that consumed by conventional systems.

[0159] In an embodiment, the platform 100 includes a local data collection system 102 deployed within the environment 104 to monitor machines, machine elements, and the environment of machines, including large machines deployed at local or distributed work sites under common control. Large machines can include industrial machines installed in various locations, such as earthmoving machines, large industrial on-road vehicles, large industrial off-road vehicles, turbines, turbomachinery, generators, pumps, pulley systems, manifolds, and valve systems. In an embodiment, heavy industrial machines can also include earthmoving machines, soil compaction equipment, transporting equipment, lifting equipment, conveying equipment, integrated production equipment, equipment used in concrete construction, and towing equipment. By way of example, earthworking machines can include excavators, backhoes, loaders, bulldozers, skid steer loaders, trenchers, motor graders, motor scrapers, crawler loaders, and wheel loading shovels. In an example, construction vehicles can include dumpers, tankers, tippers, and trailers. In an example, material handling equipment can include cranes, conveyors, forklifts, and hoists. By way of example, construction machinery can include underground handling equipment, road rollers, concrete mixers, hot mix plants, road compactors, stone crushers, pavers, slurry sealers, spreaders and plastering machines, and large pumps. Further examples of heavy industrial equipment can include different systems such as traction implements, structures, drivelines, controls, and information implementations. Heavy industrial equipment can include many different powertrains and combinations thereof to provide power for locomotion and also power auxiliary machinery and onboard functions. In each of these examples, the platform 100 can deploy a local data collection system 102 in the environment 104 in which these machines, motors, pumps, etc. operate and to which each machine, motor, pump, etc. is connected.

[0160] In an embodiment, the platform 100 may include a local data acquisition system 102 deployed within the environment 104 to monitor signals from operating or assembled machinery, such as turbine and generator sets, such as a Siemens® SGT6-5000F® gas turbine, an SST-900® steam turbine, an SGen6-100A® generator, and an SGen6-1000A® generator. In an embodiment, the local data acquisition system 102 is deployed to monitor a steam turbine driven by a flow of hot steam passing through the turbine, although other power sources, such as gas-fired burners, nuclear, molten salt loops, etc., may also be used. In these systems, the local data acquisition system 102 may monitor the turbine and water or other fluid in a closed-loop cycle in which the water condenses and is then heated until it evaporates again. The local data acquisition system 102 may monitor the steam turbine separately from a fuel source deployed to heat the water to steam. In an example, the operating temperature of the steam turbine may be between 500°C and 650°C. In many embodiments, an array of steam turbines may be arranged and configured for high, intermediate, and low pressure to optimally convert the respective steam pressures into rotary motion.

[0161] The local data acquisition system 102 may also be deployed on a gas turbine unit, thus monitoring not only the operating turbine, but also the supply of hot combustion gases sent to the turbine, which can exceed 1,500°C. Because these gases are much hotter than those in a steam turbine, the blades can be cooled by air that exits through small openings and forms a protective film, or boundary layer, between the exhaust gases and the blades. This temperature profile can be monitored by the local data acquisition system 102. Unlike a typical steam turbine, a gas turbine engine includes a compressor, a combustion chamber, and a turbine, all of which are journaled for rotation about a rotating shaft. The configuration and operation of each of these components can be monitored by the local data acquisition system 102.

[0162] In an embodiment, the platform 100 can deploy the environmental data collection system 102 in the environment 104 to monitor signals from a water turbine acting as a rotary engine to recover energy from moving water and use it to generate electricity. The type of water turbine or hydro selected for a project can be based on the still water height, often referred to as head, and the flow rate or volume of water at the site. In this example, a generator can be located at the top of a shaft that connects to the hydro turbine. As the turbine rotates, capturing naturally moving water within its blades, the turbine transmits rotational energy to the generator, generating electrical energy. In doing so, the platform 100 can monitor signals from the generator, turbine, local water system, and flow controllers, such as dam windows and gates. Additionally, the platform 100 can monitor local conditions on the electrical grid, including load, predicted demand, frequency response, and the like, and include such information in the monitoring and control deployed by the platform 100 in these hydro settings.

[0163] In embodiments, platform 100 can include a local data collection system 102 deployed in environment 104 to monitor signals from energy-producing environments, such as thermal, nuclear, geothermal, chemical, biomass, carbon-based fuel, and hybrid renewable energy plants. Many of these plants can use multiple types of energy recovery devices, such as wind turbines, hydroelectric turbines, and steam turbines powered by heat from nuclear, gas-fired, solar, and molten salt sources. In embodiments, elements of such systems can include power lines, heat exchangers, desulfurization scrubbers, pumps, chillers, recuperators, and coolers. In some embodiments, specific implementations of turbomachinery, turbines, scroll compressors, and the like can be configured into arrayed controls to monitor large-scale facilities that produce electricity for consumption, provide refrigeration, and generate steam for on-site manufacturing and heating. Arrayed control platforms can be offered by industrial equipment providers, such as Honeywell and their Experion® PSK platform. In embodiments, platform 100 can specifically communicate and integrate with on-site manufacturer-specific controls and enable one manufacturer's equipment to communicate with another's equipment. Additionally, platform 100 allows local data collection systems 102 to collect information across systems from many different manufacturers. In an embodiment, platform 100 may include local data collection systems deployed in environment 104 to monitor signals from marine industrial equipment, marine diesel engines, shipbuilding, oil and gas plants, refineries, petrochemical plants, ballast water treatment solutions, marine pumps and turbines, etc.

[0164] In an embodiment, the platform 100 includes a local data collection system 102 deployed within the environment 104, which can monitor signals from heavy industrial equipment and processes, including monitoring one or more sensors. By way of example, a sensor may be a device used to detect or respond to any type of input from the physical environment, such as an electrical, thermal, or optical signal. In an embodiment, the local data collection system 102 can include multiple sensors, such as, without limitation, a temperature sensor, a pressure sensor, a torque sensor, a flow sensor, a heat sensor, a smoke sensor, an arc sensor, a radiation sensor, a position sensor, an acceleration sensor, a strain sensor, a pressure cycle sensor, a pressure sensor, and an air temperature sensor. The torque sensor can include a magnetic torsion angle sensor. In one example, the torque and speed sensors of the local data collection system 102 are similar to those described in U.S. Patent No. 8,352,149, issued January 8, 2013, to Meacham, which is incorporated herein by reference as if fully set forth herein. In an embodiment, one or more sensors can be provided, such as a tactile sensor, a biosensor, a chemical sensor, an image sensor, a humidity sensor, an inertial sensor, or the like.

[0165] In an embodiment, the platform 100 includes a local data acquisition system 102 deployed in the environment 104 that can monitor signals from sensors that can provide signals for fault detection, including excessive vibration, incorrect materials, incorrect material properties, proper dimensions, proper shape, proper weight, and balance. Additional fault sensors include those for inventory control and inspection, such as sensors that indicate part packaging, that parts are within tolerances, packaging damage or stress, and potential impact or damage during shipping. Additional fault sensors include those that detect insufficient lubrication, excessive lubrication, the need to clean a sensor detection window, the need for maintenance due to low lubrication, and the need for maintenance due to blocked or reduced flow in a lubricated area.

[0166] In an embodiment, platform 100 may include a local data collection system 102 deployed in environment 104, including aircraft operation and manufacturing, that includes monitoring signals from specialized sensors, such as aircraft attitude and heading reference systems (AHRS), gyroscopes, accelerometers, magnetometers, etc. Platform 100 may include a local data collection system 102 deployed in environment 104 that monitors signals from image sensors, such as semiconductor charge-coupled devices (CCDs), active pixel sensors, image sensors in complementary metal-oxide semiconductor (CMOS) or negative-type metal-oxide semiconductor (NMOS, Live MOS) technology. In an embodiment, platform 100 may include a local data collection system 102 deployed in environment 104 to monitor signals from sensors, such as infrared (IR) sensors, ultraviolet (UV) sensors, touch sensors, proximity sensors, etc. In an embodiment, the platform 100 may include a local data collection system 102 deployed in the environment 104 that monitors signals from sensors configured for optical character recognition (OCR), barcode reading, surface acoustic wave detection, transponder detection, communication with home automation systems, medical diagnostics, health monitoring, etc.

[0167] In an embodiment, the platform 100 may include a local data acquisition system 102 deployed in the environment 104 to monitor signals from sensors such as microelectromechanical system (MEMS) sensors, such as an STMicroelectronics® LSM303AH smart MEMS sensor, which may include an ultra-low power, high performance system-in-package 3D digital linear acceleration sensor and a 3D digital magnetic sensor.

[0168] In an embodiment, the platform 100 can include a local data acquisition system 100 deployed in the environment 104 to monitor signals from additional large machinery such as turbines, windmills, industrial vehicles, robots, and the like. These large pieces of machinery include multiple components and elements that provide multiple subsystems for each machine. To that end, the platform 100 can include a local data acquisition system 102 deployed in the environment 104 to monitor signals from individual elements such as axles, bearings, belts, buckets, gears, shafts, gearboxes, cams, carriages, camshafts, clutches, brakes, drums, dynamos, feeders, flywheels, gaskets, pumps, jaws, robotic arms, seals, sockets, sleeves, valves, wheels, actuators, motors, and servomotors. Many of the machines and their elements can include servomotors. The local data acquisition system 102 can monitor motors, rotary encoders, and potentiometers of servomechanisms to provide three-dimensional details of the position, location, and progress of the industrial process.

[0169] In an embodiment, the platform 100 includes a local data collection system 102 deployed in the environment 104 to monitor signals from gear drives, powertrains, transfer cases, multi-speed axles, transmissions, direct drives, chain drives, belt drives, shaft drives, magnetic drives, and similar intermeshing mechanical drives. In an embodiment, the platform 100 includes a local data collection system 102 deployed in the environment 104 to monitor signals from fault conditions in the industrial machinery, which may include overheating, noise, gear squealing, locked gears, excessive vibration, shaking, underpressurization, overpressurization, and the like. During operation, installation, and maintenance, operational faults, maintenance indicators, and interactions with other machinery can occur that cause maintenance and operational issues. These faults can occur in the mechanisms of the industrial machinery, but also in the infrastructure supporting the machinery, such as wiring and local installation platforms. In an embodiment, large industrial machinery may face different types of fault conditions such as overheating, noise, squealing gears, excessive vibration of machine parts, fan vibration issues, problems with rotating parts of large industrial machinery, etc.

[0170] In an embodiment, the platform 100 includes a local data collection system 102 deployed within the environment 104 to monitor signals from industrial machinery, including potential faults caused by premature bearing failure due to contamination or loss of bearing lubrication. In another example, mechanical defects such as misaligned bearings can occur. Because many factors can contribute to failures such as metal fatigue, the local data collection system 102 monitors cyclic and local stresses. As an example, the platform 100 can monitor for improper operation of machine components, lack of maintenance and service on components, corrosion of critical machine components such as couplings or gearboxes, and misalignment of machine components. While failures cannot be completely prevented, they can be mitigated in many industrial sectors, reducing operational and financial losses. The platform 100 provides real-time monitoring and predictive maintenance in many industrial environments, demonstrating cost savings over regularly scheduled maintenance processes that replace parts according to fixed deadlines rather than actual loads and wear. To that end, the platform 10 may provide for the attention or implementation of several preventative measures such as adhering to the machine's operating manual and mode instructions, proper lubrication, and maintenance of machine parts, minimizing or eliminating machine overruns beyond their specified capacity, using worn but functional parts as needed, and properly training personnel for the use of the machine.

[0171] In an embodiment, platform 100 includes a local data acquisition system 102 deployed in environment 104 and is capable of monitoring multiple signals conveyed in multiple physical, electronic, and symbolic formats or signals. Platform 100 can employ signal processing, including multiple mathematical, statistical, computational, heuristic, and linguistic techniques, as well as multiple operations necessary to process multiple signals and extract useful information from signal processing operations, such as displaying, modeling, analyzing, synthesizing, detecting, acquiring, and extracting information from signals. By way of example, signal processing can be performed using multiple techniques, including, but not limited to, transforms, spectral estimation, statistical operations, probabilistic and stochastic operations, numerical theory analysis, data mining, etc. The processing of various types of signals forms the basis of many electrical or computational processes. As a result, signal processing is applied to nearly every field and application in industrial environments, such as audio and video processing, image processing, wireless communications, process control, industrial automation, financial systems, and quality improvement, such as feature extraction, noise reduction, and image enhancement. Image signal processing includes pattern recognition for manufacturing inspection, quality inspection, and automated operational inspection and maintenance. The platform 100 can use pattern recognition techniques to classify input data into classes based on key features with the goal of recognizing patterns or regularities in the data. The platform 100 can also perform pattern recognition processes using machine learning operations and can be used in applications such as computer vision, speech and text processing, radar processing, handwriting recognition, and computer-aided design (CAD) systems. The platform 100 can use supervised and unsupervised classification. Supervised learning classification algorithms can be based on creating classifiers for image or pattern recognition based on training data obtained from different object classes. Unsupervised learning classification algorithms can operate by using advanced analytical techniques such as segmentation and clustering to find hidden structures in unlabeled data.For example, some of the analytical techniques used in unsupervised learning can include K-means, Gaussian mixture models, hidden Markov models, etc. The algorithms used in supervised and unsupervised learning methods of pattern recognition enable the use of pattern recognition in a variety of high-precision applications. Platform 100 can use pattern recognition in security systems, tracking, sports-related applications, fingerprint analysis, medical and forensic applications, navigation and guidance systems, vehicle tracking, public infrastructure systems such as transportation systems, and face detection-related applications such as license plate surveillance.

[0172] In an embodiment, the platform 100 may include a local data collection system 102 deployed in an environment 104 that uses machine learning to enable the derivation of learning outcomes from a computer without the need for programming. Thus, the platform 100 may learn from a dataset and make decisions by making data-driven predictions and adapting according to the dataset. In an embodiment, the machine learning may include performing multiple machine learning tasks with a machine learning system, such as supervised learning, unsupervised learning, and reinforcement learning. Supervised learning involves presenting a machine learning system with a series of input examples and a desired output. Unsupervised learning may include the learning algorithm itself structuring the input through methods such as pattern detection and / or feature learning. Reinforcement learning may include a machine learning system running in a dynamic learning environment and providing feedback on correct and incorrect decisions. In an example, the machine learning may include multiple other tasks based on the output of the machine learning system. In another example, the task may be classified as a machine learning problem, such as classification, regression, clustering, density estimation, dimensionality reduction, anomaly detection, etc. In an example, the machine learning may include multiple mathematical and statistical techniques. By way of example, many types of machine learning algorithms can include machine learning algorithms such as decision tree-based learning, association rule learning, deep learning, artificial neural networks, genetic learning algorithms, inductive logic programming, support vector machines (SVMs), Bayesian networks, reinforcement learning, representation learning, rule-based machine learning, sparse dictionary learning, similarity metric learning, learning classifier systems (LCS), logistic regression, random forests, K-means, gradient boosting and Adaboost, and K-nearest neighbors (KNN), Apriori algorithms, etc. In embodiments, certain machine learning algorithms can be used, such as genetic algorithms defined to solve both constrained and unconstrained optimization problems based on natural selection, the process that drives biological evolution.As an example of this, genetic algorithms can be deployed to solve a variety of optimization problems that are not well suited to standard optimization algorithms, including problems where the objective function is discontinuous, not differentiable, stochastic, or highly nonlinear. In one example, genetic algorithms can be used to address mixed-integer programming problems where some components are restricted to integer values. Genetic algorithms and machine learning techniques and systems can be used in computer intelligence systems, computer vision, natural language processing (NLP), recommendation systems, reinforcement learning, graphical model building, and more. As in this example, machine learning systems can be used to perform intelligent computing-based control and respond to tasks in a wide variety of systems, such as interactive websites and portals, brain-machine interfaces, online security and fraud detection systems, diagnostic and therapeutic support systems, and medical applications such as DNA sequence classification. In other examples, machine learning systems can be used in advanced computing applications, such as online advertising, natural language processing, robotics, search engines, software engineering, speech and handwriting recognition, pattern matching, gameplay, computational anatomy, and bioinformatics systems. In some instances, machine learning can also be used in financial and marketing systems, such as user behavior analysis, online advertising, economic forecasting, and financial market analysis.

[0173] Further details are provided below in connection with the methods, systems, devices, and components illustrated in connection with Figures 1-6. In embodiments, methods and systems are disclosed herein for cloud-based machine pattern recognition based on the fusion of remote, analog industrial sensors. For example, data streams from vibration, pressure, temperature, accelerometers, magnetic, electric field, and other analog sensors can be multiplexed or fused, relayed over a network, and fed to a cloud-based machine learning facility, which can use one or more models related to the operating characteristics of industrial machinery, industrial processes, or their components or elements. The models can be created by humans with experience in industrial environments and may be associated with training data sets, such as those created by human or machine analysis of environmental or other similar environmental sensors. The learning machine can then operate on other data using a set of model rules or elements to initially classify the data into types, recognize specific patterns (such as those indicating the presence of a fault or an operating condition such as fuel efficiency or energy production), and provide various outputs. The machine learning facility can obtain feedback, such as one or more inputs or success strategies, such that the initial model can be trained or improved by adjusting weights, rules, parameters, etc. based on the feedback. For example, a model of fuel consumption by an industrial machine can include physics model parameters that characterize weight, motion, resistance, momentum, inertia, acceleration, and other factors indicative of consumption, and chemistry model parameters, such as those that predict energy produced and / or consumed via combustion, chemical reactions in charging and discharging a battery, etc. This model can be fed data from sensors located in the machine's environment, within the machine, etc., and data indicative of actual fuel consumption, so that the machine can be modified to increase the accuracy of its sensor-based fuel consumption estimates and provide outputs indicating what changes can be made to change the machine's operating parameters or other elements of the environment, such as ambient temperature, operation of nearby machinery, etc., to increase fuel consumption.For example, if resonance effects between two machines are adversely affecting one of them, the model can take this into account and automatically provide an output that changes the operation of one machine to reduce the resonance and increase the efficiency of one or both machines. Machine learning facilities can self-organize to provide highly accurate models of environmental conditions, such as predicting failures and optimizing operating parameters by continuously adjusting parameters to match the output to actual conditions. This can be used to improve fuel efficiency, reduce wear, increase power output, extend life, avoid failure conditions, and for many other purposes.

[0174] FIG. 14 illustrates the components and interactions of a data collection architecture involving the application of cognitive functions and machine learning for data collection and processing. Referring to FIG. 14 , the data collection system 102 may be deployed in an environment, such as an industrial environment, where one or more complex systems, such as electromechanical systems and machines, are manufactured, assembled, or operated. The data collection system 102 may include on-board sensors and may take input from one or more sensors (such as any type of analog or digital sensor disclosed herein) and one or more input sources (such as Wi-Fi, Bluetooth, NFC, or other local network connection, or sources available via the Internet) via, for example, one or more input interfaces or ports 4008. The sensors may be multiplexed, such as in combination with one or more multiplexers 4002. The data may be cached or buffered in cache / buffer 4022 and made available to an external system, such as a remote host processing system 112 as described elsewhere in this disclosure (which may include a broad processing architecture 4024 including any of the elements described in connection with other embodiments illustrated in this disclosure and in the figures), via one or more output interfaces and ports 4010 (which in embodiments may be separate from or the same as input interfaces and ports 4008). The data collection system 102 may be configured to receive input from the host processing system 112, for example, input from an analysis system 4018, which operates on data from the data collection system 102 and data from other input sources 116 to provide analysis results, which in turn may be provided as learning feedback input 4012 to the data collection system to update the configuration and operation of the data collection system 102.

[0175] The combination of inputs (including the selection of which sensors or input sources to turn "on" or "off") can be performed under the control of machine-based intelligence, such as a local recognition input selection system 4004, an optional remote recognition input selection system 4114, or a combination of the two. The recognition input selection systems 4004, 4014 can use the intelligence and machine learning capabilities described elsewhere in this disclosure, such as detected conditions (information from input sources 116 or sensors), state information (including state information determined by a machine state recognition system 4020 that can determine a state), such as state information related to an operating state, an environmental state, a state within a known process or workflow, a fault or diagnostic state, or many others. This can include optimizing input selection and configuration based on learning feedback from a learning feedback system 4012, which can include training data (e.g., data from the host processing system 112, or directly from other data collection systems, or from the host 112), and can include feedback metrics such as success criteria calculated within the analysis system 4018 of the host processing system 112. For example, if a data stream consisting of a particular combination of sensors and inputs results in positive results under a given set of conditions (e.g., providing improved pattern recognition, improved predictions, improved diagnostics, improved yield, improved return on investment, improved efficiency, etc.), metrics regarding such results from the analysis system 4018 can be provided via the learning feedback system 4012 to the recognition input selection system 4004, 4014 to help configure future data collection to select combinations under those conditions (such as by reducing the output of other sensors, causing other input sources to be deselected).In embodiments, the selection and deselection of sensor combinations under the control of one or more of the cognitive input selection systems 4004 occurs with automatic variation, such as genetic programming techniques, to promote effective combinations for a given set of states or conditions and demote less effective combinations over time based on learning feedback 4012 from the analysis system 4018, resulting in an optimization and adaptation of the local data collection system to each unique environment. Thus, an automatically adapting multi-sensor data collection system is provided, where cognitive input selection is used in conjunction with feedback to improve the effectiveness, efficiency, or other performance parameters of the data collection system within its specific environment. Performance parameters may relate to overall system metrics (e.g., financial yield, process optimization results, energy production or use), analytical metrics (e.g., pattern recognition success, prediction creation, data classification), and local system metrics (e.g., bandwidth utilization, storage utilization, power consumption, etc.). In embodiments, the analysis system 4018, status system 4020, and host cognitive input selection system obtain data from multiple data collection systems 102, allowing optimization (including input selection) to occur through multiple, coordinated actions. For example, the recognition input selection system 4114 can understand that if one data collection system 102 is already collecting vibration data for the X-axis, then the X-axis vibration sensors of other data collection systems are turned off and the host processing system 112 can select to acquire Y-axis data. In this manner, coordinated collection by the host recognition input selection system 4114 allows the activity of multiple collectors 102 across different sensor hosts to provide a rich data set for the host processing system 112 without wasting energy, bandwidth, storage space, etc. As noted above, optimization may be based on overall system success metrics, analysis success metrics, and local system metrics, or a combination of the above.

[0176] Disclosed herein are cloud-based methods and systems for machine pattern analysis of status information from multiple industrial sensors to provide predicted status information for industrial systems. In some embodiments, machine learning utilizes a state machine to track the status of multiple analog and / or digital sensors, send the status to a pattern analysis facility, and determine a predicted status of the industrial system based on historical data regarding sequences of status information. For example, if the temperature status of an industrial machine exceeds a certain threshold and subsequently leads to a fault condition, such as a broken bearing set, the temperature status can be tracked by pattern recognition, which can generate an output data structure including a predicted bearing fault condition whenever a high temperature input condition is recognized. A wide range of measurements and predicted statuses can be managed by state machines related to temperature, pressure, vibration, acceleration, momentum, inertia, friction, heat, heat flow, galvanic status, magnetic field status, electric field status, capacitance status, charge / discharge status, motion, position, and many others. A state can include a combination of states, where the data structure includes a series of states, each represented by a location in a byte-type data structure. For example, an industrial machine may be characterized by a genetic structure that provides measurements of pressure, temperature, vibration, and acoustic data, a data structure whose combined state is then operated on as a byte-based data structure to compactly characterize the combined state of the machine or environment, or a compactly characterize the predicted state. This byte-based data structure can be used by state machines for machine learning, such as pattern recognition, that operates on the structure to determine patterns that reflect the combined effects of multiple conditions. A wide variety of such structures can be tracked and used in machine learning, such as representing various lengths of combinations of different elements sensible in an industrial environment. In embodiments, the byte-based structures can be used in genetic programming techniques, such as by substituting data of various types or from various sources and tracking results over time. When used in a real-world situation, one or more desirable structures emerge based on the success, such as an indication of successful prediction of the predicted state, or the achievement of a successful outcome, improved efficiency, successful routing of information, or increased profits.That is, by varying the types and sources of data used in the byte-based structures used to optimize the machine over time, a genetic programming-based machine learning facility can "evolve" a data structure set consisting of a preferred combination of data types (e.g., pressure, temperature, and vibration) from a preferred mix of data sources for a given objective (e.g., temperature from sensor X, while vibration from sensor Y). Different desired outcomes can result in different data structures that are best suited to assist in the effective achievement of those outcomes over time by applying machine learning and promoting structures with the desired outcomes for the problem using genetic programming. The promoted data structures can provide compact and efficient data for various activities such as those described throughout this disclosure, including storage in a data pool (which can be optimized by storing the desired data structure that provides the best operational results for a given environment) and publication in a data marketplace (e.g., publishing as the most effective structure for a given objective).

[0177] In an embodiment, a platform is provided that includes cloud-based machine pattern analysis of state information from multiple analog industrial sensors to provide predicted state information for an industrial system. In an embodiment, a host processing system 112, such as one deployed in the cloud, can include a state system 4020 that can be used to estimate or calculate current states or predict future states associated with certain features of the data collection system 102 or the environment in which the data collection system 102 is deployed, such as machines, components, workflows, processes, events (e.g., whether an event has occurred), objects, people, conditions, functions, etc. The retained state information can be analyzed by the host processing system 112, such as in one or more analysis systems 4018, to determine contextual information, apply semantic and conditional logic, and perform many other functions enabled by the processing architecture 4024 described throughout this disclosure.

[0178] In an embodiment, a platform is provided that includes a cloud-based policy automation engine for the Internet of Things (IoT) along with the creation, deployment, and management of IoT devices. In an embodiment, the platform 100 includes (integrated with or included in) a host processing system 112 on a cloud platform, a policy automation engine 4032, etc., for automatically creating, deploying, and managing policies to Internet of Things (IoT) devices. Policies may include access policies, network usage policies, storage usage policies, bandwidth usage policies, device connection policies, security policies, rule-based policies, role-based policies, etc., which may be required to manage the use of IoT devices. For example, because IoT devices can have many different network and data communications with other devices, policies are needed that indicate which devices a given device can connect to, what data it can pass, and what data it can receive. In the near future, billions of devices with countless potential connections are expected to be deployed, making it impossible for humans to configure policies for IoT devices for each connection. Therefore, the intelligent policy automation engine 4032 may include cognitive capabilities for creating, configuring, and managing policies. The policy automation engine 4032 can consume information about possible policies, such as a policy database or library containing one or more public sources of available policies. These can be written in one or more conventional policy languages ​​or scripts. The policy automation engine 4032 can apply policies according to one or more models, such as based on the characteristics of a given device, machine, or environment. For example, a large machine, such as one used for power generation, might include a policy that only a verifiable local controller can change certain parameters of the power generation, thereby preventing remote "takeover" by hackers. This can in turn be accomplished by automatically discovering and applying security policies that prohibit the machine's control infrastructure from connecting to the Internet, such as by requiring access authentication.Policy automation engine 4032 can include cognitive functions such as applying policies, changing policy configurations, etc., based on state information from state system 4020. Policy automation engine 4032 can take feedback from system-wide results (e.g., the extent of security violations and policies), local results, and learning feedback systems such as based on analytical results, such as from one or more analytical results from analysis system 4018. By changing and selecting based on such feedback, policy automation engine 4032 can learn to automatically create, deploy, configure, and manage policies over time and across large numbers of devices, such as managing policies for configuring connections between IoT devices.

[0179] Methods and systems disclosed herein are for on-device sensor fusion and data storage for the Industrial IoT, including on-device sensor fusion and data storage for the Industrial IoT, where data from multiple sensors is fused at the device for a fused data stream. For example, pressure and temperature data can be multiplexed into a data stream that combines pressure and temperature in a time series, such as in a byte-based structure (where time, pressure, and temperature are multiple bytes in the data structure, the pressure and temperature remain linked to time without requiring separate processing of the streams by an external system), or the fused data can be stored at the device by performing operations such as addition, division, multiplication, or subtraction. Any of the sensor data types described throughout this disclosure can be fused in this manner and stored at the IoT device, such as a local data pool, storage, or data collector, a machine component, or other device.

[0180] In embodiments, the platform provides on-device sensor fusion and data storage for industrial IoT devices. In embodiments, the cognitive system is used for a self-organizing storage system 4028 for the data collection system 102. Sensor data, especially analog sensor data, can consume a large amount of storage capacity, especially if the data collector 102 has multiple sensor inputs onboard or from the local environment. Simply storing all data indefinitely is generally not a desirable option, and transmitting all data can become bandwidth-constrained or exceed bandwidth allowances (e.g., exceeding a cellular data plan). Therefore, storage strategies are necessary. These generally include capturing only a portion of the data (e.g., a snapshot), storing the data for a limited time, storing a portion of the data (e.g., in an intermediate or abstracted form), etc. With so many options to choose from, determining the correct storage strategy can be very complex. In embodiments, the self-organizing storage system 4028 can use cognitive systems based on the learning feedback 4012, and can use various metrics from other systems such as the analysis system 4018 or the host cognitive input selection system 4114, and can use overall system metrics, analytical metrics, and local performance indicators. The self-organizing storage system 4028 can automatically change storage parameters, such as storage location (local storage on the data collection system 102, storage near the data collection system 102 (such as using peer-to-peer organization), and remote storage such as network-based storage), storage amount, storage duration, type of data stored (individual sensors or input sources 116, as well as various combined or multiplexed data, e.g., data selected under the cognitive input selection system 4004, 4014), storage type (e.g., RAM, flash, or other short-term memory versus available hard drive space), storage configuration (e.g., raw format, tier, etc.), and so forth.Such parameter changes can be made with feedback, and over time, the data collection system 102 adapts its storage of data to optimize itself for its environment, such as a particular industrial environment, resulting in the storage of data in the appropriate amount needed and in the appropriate format available to the user.

[0181] In embodiments, the local cognitive input selection system 4004 can organize fusion data from various on-board sensors, external sensors (such as from the local environment), and other input sources 116 into one or more fusion data streams to the local collection system 102, for example, using a multiplexer 4002 to generate various signals representing combinations, permutations, blends, layers, abstractions, and data-metadata combinations of analog and / or digital source data to be processed by the data collection system 102. The selection of a particular fusion of sensors can be determined locally by the cognitive input selection system 4004, for example, based on learning feedback from the learning feedback system 4012, various overall system, analysis system, and local system results and metrics. In embodiments, the system learns to fuse particular combinations and permutations of sensors to best achieve accurate prediction of state, as indicated by feedback from the analysis system 4018 related to the analysis system's 4018's ability to predict future states, such as various state processing by the state system 4020. For example, the input selection system 4004 may indicate a selection of a subset of sensors from a larger set of available sensors, and the inputs from the selected sensors may be combined by arranging the inputs from each of those sensors into a defined multi-bit data structure (such as by taking the signal from each at a given sampling rate or time, arranging the result into a byte structure, and then collecting and processing those bytes over time), by multiplexing in a multiplexer 4002, by additive mixing of successive signals, etc. Any of a wide range of signal processing and data processing techniques for combining and fusing may be used, including convolution techniques, coercion techniques, transformation techniques, etc. The particular fusion in question may be adapted to a given situation by cognitive learning, such as by the cognitive input selection system 4004 learning based on feedback 4012 from results (such as conveyed by the analysis system 4018), so that the local data collection system 102 can perform context-adaptive sensor fusion.

[0182] In embodiments, the analysis system 4018 may apply a wide range of analytical techniques, including statistical and econometric techniques (such as linear regression analysis, the use of affinity matrices, heat map-based techniques, etc.), inference techniques (e.g., Bayesian inference, rule-based inference, inductive inference, etc.), iterative methods (such as feedback, recursion, feedforward, and other techniques), signal processing techniques (such as Fourier transforms and other transforms), pattern recognition techniques (such as Kalman and other Fermat techniques), search techniques, probabilistic techniques (such as random walks, random forest algorithms, etc.), simulation techniques (such as random walks, random forest algorithms, linear optimization, etc.), and others, which may include computing various statistics or metrics. In embodiments, the analysis system 4018 is at least partially located in the data collection system 102, and a local analysis system may compute one or more metrics related to any of the items described throughout this disclosure. For example, metrics of efficiency, power usage, storage usage, redundancy, entropy, and other factors can be computed on-board, allowing data collection 102 to enable the various cognitive and learning functions mentioned throughout this disclosure without relying on remote analysis systems (e.g., cloud-based).

[0183] In an embodiment, the host processing system 112, the data collection systems 102, or both, can include connection or integration with a self-organizing network system 4020, which can include a cognitive system for providing machine-based intelligence or organization of network utilization for the transfer of data in the data collection systems, such as handling analog and other sensor data or other source data, such as between one or more local data collection systems 102 and the host system 112. This can include utilization of an organized network for source data sent to the data collection systems, analytical data provided to or via the learning feedback system 4012, data to support marketing (as described in connection with other embodiments), and output data from one or more data collection systems 102 via output interfaces and ports 4010.

[0184] The disclosed methods and systems for a self-organizing data marketplace for the Industrial IoT include organizing available data elements into a marketplace for consumption by consumers based on training a self-organizing facility with feedback from a training set and metrics of marketplace success. The marketplace can be initially configured to make available data collected from one or more industrial environments and present the data in a menu or hierarchy, e.g., by type, source, environment, machine, one or more patterns, etc. The marketplace can vary the data collected, the organization of the data, the presentation of the data (including pushing the data to external sites, providing links, configuring multiple APIs to access the data, etc.), the pricing of the data, etc., under machine learning, which can vary any of the different parameters described above. The machine learning system can self-organize to manage all of these parameters, for example, by varying parameters over time (including the elements of the data types presented, the data sources used to obtain each type of data, the data structures presented (such as byte-based structures, fused or multiplexed structures (e.g., representing multiple sensor types), and statistical structures (e.g., representing various mathematical products of sensor information)), pricing of the data, if and how the data is presented (e.g., via API, link, push message), how the data is stored, how the data is retrieved, etc. As parameters are changed, feedback is obtained in terms of number of views, yield per access (e.g., price paid), total yield, revenue per unit, total profit, and many other success metrics, and the self-organizing machine learning facility promotes configurations that improve the success metrics and demotes configurations that do not, so that over time the marketplace is configured to offer combinations of desirable data types (e.g., those that provide robust predictions of the predicted state of a given type of specific industrial environment) from desirable sources (e.g., those that are reliable, accurate, and low-priced) with increasingly efficient pricing (e.g., pricing that tends to provide a high total profit from the marketplace).The marketplace may include spiders, web crawlers, etc., that search input data sources to search for those that expose potentially relevant data, such as data pools, connected IoT devices, etc. These may be trained by humans and may be improved by machine learning in a similar manner as described elsewhere in this disclosure.

[0185] In an embodiment, a platform having a self-organizing data marketplace for industrial IoT data is provided. Referring to FIG. 15 , in one embodiment, a platform having a cognitive data marketplace 4102, sometimes referred to as a self-organizing data marketplace, is provided for data collected by one or more data collection systems 102 or data from other sensors or input sources 116 located in various data collection environments, such as industrial environments. In addition to data collection systems 102, this can include data collected, processed, or exchanged by IoT devices such as cameras, monitors, embedded sensors, mobile devices, diagnostic equipment and systems, instrumentation systems, telematics systems, etc., for monitoring various parameters and functions of machines, devices, components, parts, operations, functions, conditions, states, events, workflows, and other elements of such environments (collectively encompassed by the term “state”). The data can also include metadata related to any of the foregoing that enables further processing for data extraction, transformation, loading, and processing, such as describing the data, indicating its origin, indicating factors related to identity, access, roles, and permissions, providing a summary or abstract of the data, or one or more other items of data. Such data (such term includes metadata unless the context indicates otherwise) is valuable to third parties, either as individual elements (such as data about the state of an environment that can be used as a condition in a process) or as aggregates (e.g., data collected across many systems and devices, optionally in different environments, that can be used to develop behavioral models, train learning systems, etc.). As billions of IoT devices are deployed with countless connections, the amount of available data will explode. To enable data access and utilization, the cognitive data marketplace 4102 enables various components, features, services, and processes to allow users to contribute, discover, consume, and transact packages of data, such as batches of data, streams of data (including event streams), data from various data pools 4120, etc.In embodiments, the cognitive data marketplace 4102 can be included in one or more connected or integrated components of the host processing architecture 4024 of the integrated host processing system 112, such as a cloud-based system, as well as various sensors, input sources 115, data collection systems 102, etc. The cognitive data marketplace 4102 can include a marketplace interface 4108, which includes one or more supplier interfaces through which data suppliers make data available and one or more consumer interfaces through which data is discovered and retrieved. The consumer interfaces can include an interface to a data marketplace search system 4118, which includes functionality that allows users to indicate what type of data they want to retrieve, such as by entering keywords into a natural language search interface that characterizes the data or metadata. The search interface can use various search and filtering techniques, including keyword matching, collaborative filtering (such as using known consumer preferences or characteristics to match similar consumers and their past results of other consumers), and ranking techniques (such as ranking based on the success of past results according to various metrics described in connection with other embodiments in this disclosure). In some embodiments, the supply interface may enable a data owner or supplier to supply one or more packages of data to and through the cognitive data marketplace 4102, such as packaging a batch of data, a stream of data, etc. Suppliers may pre-package data by supplying data from a single input source 116, a single sensor, etc., or by providing combinations, permutations, etc. (such as multiplexed analog data, mixed bytes of data from multiple sources, results of extracts, loads and transforms, results of convolutions, etc.), as well as metadata associated with any of the foregoing. Packaging may include pricing based on pre-batch, streaming (such as subscriptions to event feeds or other feeds or streams), pre-item, revenue share, etc.For data related to pricing, the data exchange system 4114 can track orders, deliveries, and usage, including order fulfillment. The exchange system 4114 can include rich exchange functionality, including digital rights management, such as by managing access control to purchasing data and managing encryption keys to govern usage (such as allowing data to be used for a limited time, in a limited domain, by a limited set of users or roles, or for a limited purpose). The exchange system 4114 can manage payments, such as by processing credit cards, wire transfers, debits, and other forms of consideration.

[0186] In embodiments, the cognitive data packaging system 4012 of the marketplace 4102 may package data using machine-based intelligence, such as by automatically configuring data packages such as batches, streams, pools, etc. In embodiments, the packaging may be according to one or more rules, models, or parameters, such as by packaging or aggregating data that may supplement or complement existing models. For example, operational data from a group of similar machines (such as one or more of the industrial machines described throughout this disclosure) may be aggregated based on metadata indicative of the type of data or by recognizing features or characteristics of the data stream that are indicative of the data's characteristics. In embodiments, the packaging may be performed using machine learning and cognitive capabilities, such as learning to combine, sort, blend, hierarchy, etc., information from input sources 116, sensors, data pools 4120, and information from data collection system 102 that may meet user requirements or be a measure of success. The learning may be based on learning feedback 4012, such as based on criteria determined by the analysis system 4018, such as system performance criteria, data collection criteria, analysis criteria, etc. In embodiments, success measures may be correlated to market success measures such as views of the package, engagement with the package, purchase or licensing of the package, payment for the package, etc. Such criteria may be calculated by the analytics system 4018, including associating specific feedback criteria with search terms and other inputs, such that the cognitive packaging system 4110 can find and create packages designed to provide high value to consumers and increase returns to data providers. In embodiments, the cognitive data packaging system 4110 may use different combinations, permutations, blends, etc., vary the weightings applied to given input sources, sensors, data pools, etc., and use learning feedback 4012 to automatically vary packaging, such as promoting preferred packages and de-emphasizing less preferred packages.This can be done using genetic programming or similar techniques to compare the results of different packages. The feedback can include state information from the state system 4020 (e.g., regarding various operating conditions, etc.), as well as information about market conditions and conditions, such as information about pricing and potential other data sources. Thus, an adaptive cognitive data package system 4110 is provided that automatically adapts to conditions to provide a preferred data package for the market 4102.

[0187] In embodiments, a cognitive data pricing system 4112 may be provided to set pricing for data packages. In embodiments, the data pricing system 4112 may use a set of rules, models, etc., such as to set pricing based on supply conditions, demand conditions, setting pricing of various available sources, etc. For example, pricing for a package may be configured to be set based on the sum of the prices of its components (e.g., input sources, sensor data, etc.), or may be set based on a rules-based discount of the sum of the component prices. Rules and conditional logic may be applied, such as rules that consider cost factors (e.g., bandwidth and network usage, peak demand factors, scarcity factors, etc.), rules that consider usage parameters (e.g., package purpose, domain, user, role, duration, etc.), and many others. In embodiments, the cognitive data pricing system 4112 may include fully cognitive intelligent features, such as using genetic programming to automatically change pricing and track resulting feedback. The tracking feedback may include various financial yield metrics, usage metrics, etc., provided by calculating analytics system 4018 metrics on data from the data trading system 4114.

[0188] Disclosed herein are methods and systems for self-organizing data pools, which may include self-organizing data pools based on utilization and / or yield metrics, where the utilization and / or yield metrics include utilization and / or yield metrics tracked for multiple data pools. The data pools may initially include pools of unstructured or loosely structured data, including data from industrial environments, such as sensor data from or related to industrial machines or components. For example, the data pools may capture data streams from various machines or components in the environment, such as turbines, compressors, batteries, reactors, engines, motors, vehicles, pumps, rotors, axles, bearings, valves, and the like, along with data streams including analog and / or digital sensor data (of a wide variety), published data on operating conditions, diagnostic and fault data, machine or component identification data, asset tracking data, and many other types of data. Each stream may have an identifier within the pool, indicating its source and, optionally, its type. The data pool can be accessed by external systems through one or more interfaces or APIs (e.g., RESTful APIs) or by data integration elements such as gateways, brokers, bridges, connectors, and similar functions that provide access to available data streams. The data pool can be managed by a self-organizing machine learning facility that can configure the data pool, such as by determining which sources are used for the pool, which streams are available, and managing connections into and out of the API or data pool. The self-organization can obtain feedback based on success criteria. The success criteria may include utilization and yield criteria. The utilization and yield criteria may calculate the cost of acquiring and / or storing the data, and the benefit of the pool, measured by profit or other means (which may include, for example, user indications of usefulness).For example, a self-organizing data pool may recognize that chemical and radiation data from an energy-producing environment is regularly accessed and extracted, while vibration and temperature data is not being used, in which case the data pool will automatically reorganize by either ceasing to store vibration and / or temperature data or acquiring better sources of such data. This automated reorganization can also be applied to data structures to encourage different data types, different data sources, different data structures, etc. through gradual iteration and feedback.

[0189] In embodiments, a platform is provided with self-organization of data pools based on utilization and / or yield metrics. In some embodiments, data pool 4020 may be a self-organizing data pool 4020, such as organized by cognitive capabilities described throughout this disclosure. Data pool 4020 can self-organize in response to learning feedback 4012, such as based on baseline and outcome feedback, including that calculated by analysis system 4018. Organization includes determining which data or packages of data to store in the pool (presenting a particular combination, permutation, aggregation, etc.), the structure of such data (e.g., flat, hierarchical, linked, or other structure), duration of storage, characteristics of the storage media (e.g., hard disk, flash memory, SSD, network-based storage, etc.), arrangement of storage bits, and other parameters. The contents and nature of the storage can change, allowing the data pool to learn and adapt based on conditions of host system 112, one or more data collection systems 102, storage environment parameters (e.g., capacity, cost, and performance factors), data collection environment parameters, market parameters, and many others. In an embodiment, the pool 4020 can learn and adapt, such as by modifying the above and other parameters, in response to yield metrics such as return on investment, power usage optimization, revenue optimization, etc.

[0190] The methods and systems disclosed herein for training AI models based on industry-specific feedback include training an AI model based on industry-specific feedback reflecting utilization, yield, or impact criteria, where the AI ​​model operates on sensor data from an industrial environment. As described above, these models can include operational models of the industrial environment, machines, and workflows, models of predicted health, failure prediction and maintenance optimization models, self-organizing storage models (on-device, in data pools, and / or in the cloud), optimization models for data transfer (such as optimization of network coding, network state-aware routing, etc.), optimization models for data markets, and many other models.

[0191] In embodiments, a platform for training AI models based on industry-specific feedback is provided. In embodiments, various embodiments of the cognitive systems disclosed herein can obtain input and feedback from industry-specific and domain-specific sources 116, such as those related to the optimization of specific machines, components, processes, etc. Thus, learning and adaptation of other functions, such as storage organization, network usage, sensor and input data combination, data pooling, data packaging, data pricing, and other purposes of the marketplace 4102 or other host processing system 112, can be configured by learning about domain-specific feedback criteria of a given environment, e.g., an industrial environment, or application, e.g., an application involving IoT devices. This may include efficiency optimization (such as for electrical, electromechanical, magnetic, physical, thermodynamic, chemical and other processes and systems), output optimization (such as for the production of energy, materials, products, services and other outputs), failure prediction, avoidance and mitigation (such as for the aforementioned systems and processes), performance criteria optimization (such as return on investment, yield, profit margin, profit margin, revenue), cost reduction (such as labor costs, bandwidth costs, data costs, material input costs, license costs), profit optimization (such as for safety, satisfaction, health, etc.), workflow optimization (such as optimizing time and resource allocation to processes), and others.

[0192] Disclosed herein are methods and systems for a self-organized swarm of industrial data collectors, including a self-organized swarm of industrial data collectors that organizes among itself to optimize data collection based on the capabilities and status of the swarm members. Each member of the swarm can be configured with intelligence and the ability to collaborate with other members. For example, swarm members can track information about the data other members are processing and can intelligently assign data collection activities, data storage, data processing, and data publishing to the swarm, taking into account environmental conditions, the capabilities of the swarm members, operational parameters, rules (e.g., from a rules engine that controls the swarm's operation), and the current status of the members. For example, among four collectors, one with a relatively low current level (e.g., low battery) can receive a small amount of power from a reader or querying device (e.g., an RFID reader) when it needs to publish data and thus be temporarily assigned the role of publishing data. A second collector with good power levels and robust processing capabilities can be assigned more complex functions such as processing data, fusing data, and organizing the rest of the swarm (including self-organization under machine learning to optimize the swarm over time by adjusting operating parameters, rules, etc. based on feedback). A third collector in the swarm with robust storage capabilities can be tasked with collecting and storing categories of data, such as vibration sensor data, which consumes significant bandwidth. A fourth collector in the swarm, for example, one with lower storage capacity, can be assigned the role of collecting data that would normally be discarded, such as current diagnostic status, where only fault data needs to be maintained and communicated. Members of a swarm can be connected in a peer-to-peer relationship, using one member as a "master" or "hub," or by connecting in a series or ring, with each member communicating data, including instructions, to the next and recognizing the appropriate functions and instruction characteristics for the previous and / or next member.The swarm can be used for its storage allocation (e.g., using each memory as an aggregate data store). In these examples, the aggregate data store may support, for example, a distributed ledger storing transaction data. The transaction data may be data, such as transactions occurring in an industrial environment, for transactions involving data collected by the swarm. In embodiments, the transaction data may also include, for example, data used to manage the swarm, the environment, or a machine or its components. The swarm can self-organize based on either machine learning functions deployed in one or more members of the swarm or instructions from an external machine learning facility, which can optimize storage, data collection, data processing, data presentation, data transfer, and other functions based on management of parameters associated with each. The machine learning facility can start with an initial configuration and vary swarm parameters related to any of the foregoing (including varying swarm membership), and iterate based on feedback to the machine learning facility regarding, for example, success criteria, utilization criteria, efficiency criteria, success criteria in forecasting or predicting conditions, productivity criteria, yield criteria, profit criteria, etc. Over time, the swarm can be optimized to a desired configuration to achieve desired success criteria for an operator, operator, or host industrial environment or its machines, components, or its processes.

[0193] In an embodiment, a platform having a self-organized swarm of industrial data collectors is provided. In an embodiment, the host processing system 112, having its processing architecture 4024 (and optionally integrated with or including the cognitive data marketplace 4102), can integrate, connect, or utilize information from a self-organizing swarm 4202 of data collectors 102. In some embodiments, the self-organizing swarm 4202 can organize two or more data collection systems 102, such as through the deployment of cognitive capabilities of one or more data collection systems 102, and can provide coordination of the swarms 4202. The swarms 4202 can be organized based on a hierarchical organization (e.g., where a master data collector organizes and supports the activities of one or more data collection systems 102), a collaborative organization (e.g., where decisions for organizing the swarm 4202 are distributed among the data collectors 102, or where various models of decision-making are used, e.g., a voting system, a points system, a least-cost routing system, a prioritization system, etc.). In embodiments, one or more data collectors 102 may have mobility capabilities, such as when the data collectors are located on or in a mobile robot, drone, mobile underwater vehicle, etc., such that orchestration may include locating and positioning the data collectors 102. Data collection systems 102 may communicate with each other and with the host processing system 112, including sharing a single storage space that includes storage of or accessible to one or more collectively assigned collectors (which, in embodiments, may be treated as a unified storage space even when physically distributed using virtualization capabilities, etc.). Orchestration may be automated based on one or more rules, models, conditions, processes, etc., such as embodied or executed by conditional logic, and orchestration may be governed by policies, such as processed by a policy engine.Rules can be based on industry-, application-, and domain-specific objects, classes, events, workflows, processes, and systems, such as configuring the swarm 4202 to collect selected types of data at specified locations and times, as described above. For example, the swarm 4202 assigns data collectors 102 to continuously collect diagnostic, sensor, instrumentation, and / or telematic data from each of a set of machines running an industrial process (such as a robotic manufacturing process), including the time and location of inputs and outputs to each of those machines. In embodiments, the self-organization is cognitive, in which the swarm varies one or more collection parameters, adapts parameter selection, weightings applied to parameters, etc. over time. In some examples, this may be responsive to learning and feedback. The learning and feedback may be from a learning feedback system 4012, which may be based on various feedback criteria that may be determined, for example, by applying an analysis system 4018 (in embodiments, in the swarm 4202, the host processing system 112, or a combination thereof) to data handled by the swarm 4202 or other elements (including market factors, etc.) of the various embodiments disclosed herein. Thus, the swarm 4202 may exhibit adaptive behavior, such as adapting to the current state 4020 or the predicted state of its environment at a given time (taking into account market behavior), the behavior of various objects (such as IoT devices, machines, components, and systems), processes (including events, states, workflows, etc.), and other factors. Parameters may change in processes of variation (such as neural nets, self-organizing maps), selection, promotion, etc. (such as genetic programming or other AI-based techniques).Parameters that can be managed, changed, selected, and adapted through cognition and machine learning include storage parameters (such as the location, type, duration, quantity, and structure of the swarm 4202), network parameters (such as how the swarm 4202 is organized, e.g., mesh, peer-to-peer, ring, serial, hierarchical, and other network configurations, as well as bandwidth utilization, data routing, network protocol selection, network coding type, and other network parameters), security parameters (such as the configuration of various security applications and services), location and positioning parameters (such as routing of movement to locations of mobile data collectors 102, positioning and feedback of collectors 102 to points of data acquisition, to each other, and to locations where network availability is desired, etc., can be based on any of the types of feedback described herein, thereby adapting to current and predicted conditions over time to achieve a wide range of desired objectives.

[0194] Disclosed herein are methods and systems for an industrial IoT distributed ledger, including a distributed ledger that supports tracking of transactions performed in an automated data marketplace for industrial IoT data. The distributed ledger can distribute storage across devices using a secure protocol, such as one used for virtual currencies, such as the Blockchain protocol used to support the Bitcoin currency. The ledger or similar transaction record can have a structure in which successive members of a chain store data of previous transactions and competition can be established to determine which of the stored alternatives is the “best” (e.g., most complete) and can be stored across data collectors, industrial machines or components, data pools, data marketplaces, cloud computing elements, servers, and / or on the IT infrastructure of an industrial environment or enterprise, such as the owner, operator, or host of the system disclosed herein. Ledgers and transactions can be optimized using machine learning to provide storage efficiency, security, redundancy, etc.

[0195] In embodiments, the cognitive data marketplace 4102 may use a secure architecture for tracking and settling transactions, such as a distributed ledger 4004, where data package transactions are tracked in a chained, distributed data structure, such as a blockchain, enabling forensic analysis and verification, where individual devices store portions of the ledger representing data package transactions. The distributed ledger 4004 may be distributed to IoT devices, data pools 4020, data collection systems 102, etc., allowing transaction information to be authenticated without trust on a single central information repository. The transaction system 4114 may be configured to store data in the distributed ledger 4004, retrieve data from it (and from constituent devices), and settle transactions. This provides the distributed ledger 4004 for processing transactions in data, such as packages of IoT data. In embodiments, a self-organizing storage system 4028 may be used to optimize the storage of the distributed ledger data, as well as to organize packages of data, such as IoT data, that may appear on the marketplace 4120.

[0196] Disclosed herein are methods and systems for self-organizing collectors, including self-organizing, multi-sensor data collectors that can optimize data collection, power, and / or yield based on the conditions of their environment. The collectors can organize data collection by turning on and off particular sensors based on, for example, historical usage patterns or success criteria, as governed by a machine learning facility that configures and tracks the success criteria. For example, a multi-sensor collector can learn to turn off particular sensors when power levels are low or during periods of low data utilization from such sensors, or vice versa. Self-organization can also automatically organize how data is collected (sensors, external sources), how it is stored (granularity, degree of compression, length, etc.), how it is presented (fused or multiplexed structures, byte-like structures, or intermediate statistical structures such as after addition, subtraction, division, multiplication, squaring, normalization, scaling, or other operations), etc. This can be improved over time from an initial configuration by training the self-organizing facility on datasets from the actual operating environment, such as based on feedback criteria, including many types of feedback described throughout this disclosure.

[0197] Disclosed herein are methods and systems for network-aware collectors, including network-condition-aware, self-organizing, multi-sensor data collectors that can optimize based on bandwidth, quality of service (QoS), pricing, and / or other network conditions. Network sensitivity can include awareness of the price of data transfer (such as allowing the system to pull or push data during off-peak periods or within the available parameters of a paid data plan), the quality of the network (such as avoiding periods where errors are likely to occur), the quality of environmental conditions (such as delaying transmission until signal quality is good, such as when the collector leaves a shielded environment, avoiding wasted power usage when searching for a signal when shielded, such as by large metal structures common in industrial environments), etc.

[0198] Disclosed herein are methods and systems for remotely orchestrating a universal data collector that can power up and down sensor interfaces based on needs and / or conditions of interest in an industrial data collection environment. For example, the interface can recognize which sensors are available, and the interface and / or processor can turn on and obtain input from such sensors, including hardware interfaces that allow these sensors to be plugged into the collector, wireless data interfaces (where the collector can ping the sensors and optionally provide some power via interrogation signals), and software interfaces, such as for processing specific types of data. Thus, collectors capable of processing various types of data can be configured to suit specific applications in a given environment. In embodiments, configuration can be automatic or under machine learning, which can improve the configuration by optimizing parameters based on feedback criteria over time.

[0199] Disclosed herein are methods and systems for self-organizing storage for multi-sensor data, including self-organizing storage in a multi-sensor data collector for industrial sensor data. The self-organizing store can allocate storage based on the application of machine learning, thereby improving storage configuration over time based on feedback criteria. Storage can be optimized by configuring what data types to use (e.g., byte-like structures, structures representing fused data from multiple sensors, structures representing statistics or metrics calculated by applying mathematical functions to data, etc.), configuring compression, configuring data storage times, configuring write strategies (e.g., by striping data across multiple storage devices using a protocol in which one device stores instructions for other devices in the chain), and configuring storage tiers (e.g., by providing intermediate statistics calculated to facilitate faster access to frequently accessed data items). Thus, highly intelligent storage systems can be configured and optimized over time based on feedback.

[0200] Disclosed herein are methods and systems for self-organizing network coding for multi-sensor data networks, including self-organizing network coding for data networks that transfer data from multiple sensors in industrial data collection environments. Network coding, including random linear network coding, can enable highly efficient and reliable transfer of large amounts of data between various types of networks. Selecting different network coding configurations based on machine learning can optimize network coding and other network transfer characteristics based on network conditions, environmental conditions, and other factors, such as characteristics of the data being transferred, environmental conditions, operating conditions, etc., including by training a network coding selection model over time based on feedback of success criteria, such as any of the criteria described herein.

[0201] In an embodiment, a platform with self-organizing network coding for a multi-sensor data network is provided. The cognitive system can vary one or more parameters for network operation, such as network type selection (e.g., selecting among available local, cellular, satellite, WiFi, Bluetooth, NFC, Zigbee, and other networks), network selection (e.g., selecting a particular network, such as one known to have desired security features), network coding selection (e.g., selecting random linear network coding, fixed coding, or other network coding types for efficient transmission), network timing selection (e.g., configuring distribution based on network pricing conditions, traffic, etc.), network feature selection (e.g., selecting cognitive features, security features, etc.), network conditions (e.g., network quality based on current environmental or operating conditions), network characteristic selection (e.g., enabling available authentication, authorization, and similar systems), network protocol selection (e.g., HTTP, IP, TCP / IP, cellular, satellite, serial, packet, streaming, and many other protocols), and others. Considering bandwidth constraints, price fluctuations, susceptibility to environmental factors, security concerns, etc., selecting the optimal network configuration can be highly complex and context-dependent. The self-organizing networking system 4030 can vary the combinations and permutations of these parameters while obtaining input from the learning feedback system 4012 (e.g., using information from the analysis system 4018 about various outcome criteria) or while obtaining other inputs. In many examples, the outcomes may include system-wide criteria, analytical success criteria, local performance indicators, etc. In an embodiment, the input from the learning feedback system 4012 may include information from various sensors and input sources 116, information from the state system 4020 about conditions (e.g., events, environmental conditions, operational conditions, and many more), or other information.By varying and selecting alternative configurations of network parameters in different conditions, the self-organizing network system can find a configuration that best suits the environment monitored and managed by the host system 112 and that best suits the emerging network conditions, thereby providing a self-organizing network condition adaptive data collection system.

[0202] Referring to FIG. 14 , the data collection system 102 can have one or more output interfaces and / or ports 4010. These can include network ports and connections, application programming interfaces, and the like. Disclosed herein are methods and systems for haptic or multisensory user interfaces, including wearable haptic or multisensory user interfaces with vibration, thermal, electrical, and / or acoustic outputs for industrial sensor data collectors. For example, the interface can be configured to provide input or feedback to a user, such as based on data from sensors in its environment, based on data structures configured to support this. For example, when a fault condition is detected based on vibration data (such as the result of bearing wear, shaft misalignment, or a machine-to-machine resonance condition), this can be indicated in the haptic interface by a vibration of the interface, such as by vibrating a wrist-worn device. Similarly, thermal data indicating overheating can be indicated by heating or cooling a wearable device while a worker is working on a machine and cannot necessarily see the user interface. Similarly, electrical or magnetic data can be presented, such as by beeping, to indicate the presence of, for example, a broken electrical connection or wire. That is, a multisensory interface can help a user intuitively, for example, a person wearing a wearable device can quickly find out what is happening in their environment through a wearable interface that has various modes of interaction that do not require the user to see a graphical UI; in many industrial environments, the user needs to keep an eye on the environment and seeing a graphical UI would be difficult or impossible.

[0203] In an embodiment, a platform having a wearable tactile user interface with vibration, thermal, electrical, and / or acoustic output for an industrial sensor data collector is provided. In an embodiment, a tactile user interface 4302 is provided as an output of the data collection system 102, e.g., to one or more components of the data collection system 102 or another system, such as a wearable device, a mobile phone, etc., for processing and providing the vibration, thermal, electrical, and / or acoustic output. The data collection system 102 can be provided in a form factor suitable for delivering tactile input to a user, e.g., by vibration, heating, cooling, or ringing, such as disposed on headgear, an armband, a wristband, a watch, a belt, clothing, a uniform, etc. In such a case, the data collection system 102 can be integrated into gear, uniforms, equipment, etc. worn by a user, such as an individual responsible for operating or monitoring an industrial environment. In embodiments, signals from various sensors or input sources (or selective combinations, permutations, mixtures, etc., as managed by one or more cognitive input selection systems 4004, 4014) can trigger haptic feedback. For example, if nearby industrial machinery is overheating, the haptic interface can alert the user by heating or by sending a signal to another device (such as a cell phone) to also heat it. If the system is experiencing abnormal vibrations, the haptic interface can vibrate. Thus, through various forms of haptic input, the data collection system 102 can alert the user to the need to address one or more devices, machines, or other factors in the industrial environment, etc., without the user having to take their hands off their work to read a message or attend to visual alerts. The selection of the haptic interface and which output to provide can be considered in the cognitive input selection systems 4004, 4014.For example, the analysis system 4018 can monitor and analyze user behavior (e.g., responses to inputs), and feedback can be provided via the learning feedback system 4012, so that signals can be provided at the right time and in the right way based on the correct collection or packaging of sensors and inputs to optimize the efficiency of the haptic system 4202. This can include rule-based or model-based feedback (e.g., providing outputs that respond in a logical way to the source data conveyed). A cognitive haptic system can be provided in which the selection, output, timing, intensity level, duration, and other parameters of the haptic feedback input or trigger (or the weightings applied thereto) can be varied in the variation, promotion, and selection process (e.g., using genetic programming) based on feedback based on real-world responses to feedback in actual situations or based on the results of simulations and testing of user behavior. This provides an adaptive haptic interface for the data collection system 102 that learns to meet demands and optimizes the effects on user behavior, such as overall system performance, data collection performance, and analysis performance.

[0204] Disclosed herein are methods and systems for a presentation layer for AR / VR industrial glasses in which heatmap elements are presented based on patterns and / or parameters of collected data. Disclosed herein are methods and systems for state-dependent, self-organizing adjustment of AR / VR interfaces based on feedback metrics and / or training in industrial environments. In embodiments, any data, measurements, etc. described throughout this disclosure can be presented by visual elements, overlays, etc. for presentation in an AR / VR interface such as industrial glasses, in an AR / VR interface on a smartphone or tablet, in an AR / VR interface on a data collector (which may be built into a smartphone or tablet), on a display installed on a machine or component, and / or on a display installed in an industrial environment.

[0205] In an embodiment, a platform is provided having a heat map displaying collected data for AR / VR. In an embodiment, a platform is provided having a heat map 4204 displaying data collected from a data collection system 102 to provide input to an AR / VR interface 4208. In an embodiment, a heat map interface 4304 is provided as an output for the data collection system 102, such as for handling and providing information for visualization of various sensor and other data (e.g., map data, analog sensor data, and other data), such as for one or more components of the data collection system 102 or another system, such as a mobile device, tablet, dashboard, computer, AR / VR device, etc. The data collection system 102 can be provided in a form factor suitable for conveying visual input to a user, such as by presenting a map with level indicators of analog and digital sensor data, for example, to indicate levels of rotation, vibration, heating or cooling, pressure, and many other conditions. In such cases, the data collection system 102 can be integrated with equipment used by individuals responsible for operating or monitoring industrial environments. In embodiments, signals from various sensors and input sources (or selective combinations, permutations, blends, etc., as managed by one or more cognitive input selection systems 4004, 4014) can provide input data for the heat map. Coordinates can include real-world location coordinates (such as geographic locations or locations on a map of the environment) as well as other coordinates, such as time-based coordinates, frequency-based coordinates, or other coordinates, that allow for the representation of analog sensor signals, digital signals, input source information, and various combinations in a map-based visualization, such that colors represent various input levels along associated dimensions. For example, if a nearby industrial machine is overheating, the heat map interface may alert the user by showing the machine in bright red.If a system is vibrating abnormally, the heatmap interface may display a different color in the visual element for that machine, or may bring an icon or display element to the interface indicating that the machine is vibrating and call attention to that element. Clicking, touching, or other methods of interacting with the map allow the user to drill down and see the underlying sensors and input data used as input to the heatmap display. Thus, through various forms of display, the data collection system 102 can notify users of the need to pay attention to one or more devices, machines, or other elements in an industrial environment, etc., without them having to read a text-based message or input. The heatmap interface and the selection of which output to provide may be taken into account in the cognitive input selection system 4004, 4014. For example, user behavior (e.g., responses to inputs and displays) may be monitored and analyzed by analysis system 4018, and feedback may be provided via learning feedback system 4012, so that signals are provided based on the right collection or package of sensors and inputs at the right time and in the right way to optimize the effectiveness of the heatmap UI 4304. This may include rule-based or model-based feedback (e.g., providing outputs that correspond to some logical manner in response to transmitted source data). In embodiments, a cognitive heatmap system may be provided in which input or trigger selection, output selection, color, visual elements, timing, intensity level, duration, and other parameters (or weights applied thereto) for the heatmap display can be varied in the process of change, promotion, or selection (e.g., using genetic programming) based on real-world responses to feedback in actual situations, or based on the results of simulations and testing of user behavior.Thus, a heatmap interface is provided that is adapted to the data collection system 102, or the data collected thereby, or the data processed by the host processing system 112, and that can learn and adapt to feedback to meet requirements and optimize the impact on user behavior and responses, such as overall system results, data collection results, analysis results, and the like.

[0206] In embodiments, a platform is provided having an auto-adjusting function for AR / VR visualization of data collected by a data collector. Also, in embodiments, a platform is provided having an auto-adjusting AR / VR visualization system 4308 for visualizing data collected by a data collection system 102, where the data collection system 102 has an AR / VR interface 4208 or provides input to the AR / VR interface 4308, such as a mobile phone located in a virtual reality or AR headset or a set of AR glasses, etc. In embodiments, the AR / VR system 4308 is provided as an output interface to the data collection system 102, such as to process and provide information for visualization of various sensor data and other data (e.g., map data, analog sensor data, and other data) to one or more components of the data collection system 102 or to other systems, such as a mobile device, tablet, dashboard, computer, AR / VR equipment, or the like. The data collection system 102 may be provided in a form factor suitable for delivering AR or VR visual, auditory, or other sensory input to a user, such as by presenting one or more displays (e.g., 3D realistic visual images, objects, maps, camera overlays, or other overlay elements), maps, and indicators of analog and digital sensor data levels, such as rotation, vibration, heating or cooling, pressure, and many other conditions, for the input source 116, or the like. In such cases, the data collection system 102 may be integrated with equipment used by individuals operating or monitoring industrial environments.

[0207] In embodiments, signals from various sensors and input sources (or selective combinations, permutations, blends, etc., as managed by one or more cognitive input selection systems 4004, 4014) can provide input data for setting, configuring, modifying, or otherwise determining the AR / VR elements. The visual elements can include a wide variety of icons, map elements, menu elements, sliders, toggles, colors, shapes, sizes, etc., to represent analog sensor signals, digital signals, input source information, and various combinations. In many examples, the color, shape, and size of the visual overlay elements can represent changes in input levels along a magnitude associated with a sensor or combination of sensors. In a further example, if nearby industrial machinery is overheating, the AR element can alert the user by showing a flashing red icon representing that type of machinery on a portion of the display of a pair of AR glasses. Also, if the system detects abnormal vibrations, a virtual reality interface showing a visual image of a machine's components (e.g., a camera view of the machine overlaid with elements of a 3D visual image) can indicate the vibrating component with an accent color, movement, or the like to make it stand out in the virtual reality environment being used to help the user monitor and maintain the machine. Clicking, touching, looking, or other methods of interacting with the virtual elements of the AR / VR interface can allow the user to drill down and see the underlying sensors and input data used as input to the display. Thus, through various forms of display, the data collection system 102 can notify users of the need to pay attention to one or more devices, machines, or other elements in an industrial environment, etc., without requiring them to read text-based messages or inputs or to divert their attention from the applicable environment (whether a real environment with AR capabilities or a virtual environment for simulation, training, etc.).

[0208] Selection and configuration of the AR / VR output interface 4208 and which outputs and displays to provide can be handled by the cognitive input selection system 4004, 4014. For example, user behavior (e.g., responses to inputs and displays) can be monitored and analyzed by the analysis system 4018, and feedback can be provided via the learning feedback system 4012, so that the AR / VR display signal is provided based on the correct collection and package of sensors and inputs at the right time and in the right way to optimize the effectiveness of the AR / VR UI 4308. This can include rule-based or model-based feedback (e.g., providing an output that corresponds to some logical progression in relation to transmitted source data). In embodiments, a cognitively tailored AR / VR interface control system 4308 may be provided that can vary input and trigger selections for AR / VR display elements, output selections (such as color, visual elements, timing, intensity levels, duration, and other parameters), and other parameters of the AR / VR environment in a process of change, promotion, or selection (e.g., using genetic programming) based on real-world responses in actual situations or based on the results of simulating and testing user behavior. Thus, an AR / VR interface is provided that is adapted to the data collection system 102, or the data collected thereby, or the data processed by the host processing system 112, and that can learn and adapt to feedback to meet requirements and optimize the impact on user behavior and responses, such as overall system results, data collection results, analysis results, and the like.

[0209] As noted above, methods and systems for continuous ultrasonic monitoring, including providing continuous ultrasonic monitoring of rotating elements and bearings of energy generating equipment, are disclosed herein. Embodiments include using the continuous ultrasonic monitoring of an industrial environment as a source for a cloud-deployed pattern recognizer. Embodiments include using the continuous ultrasonic monitoring to provide updated state information to a state machine used as input to the cloud-based pattern recognizer. Embodiments include providing available continuous ultrasonic monitoring information to a user based on policies declared in a policy engine. Embodiments include storing ultrasonic continuous monitoring data with other data in a fusion data structure on an industrial sensor device. Embodiments include making a stream of continuous ultrasonic monitoring data from the industrial environment available as a service from a data marketplace. Embodiments include feeding the stream of continuous ultrasonic data into a self-organizing data pool. Embodiments include training a machine learning model that monitors the continuous ultrasonic monitoring data stream, the model based on a training set created from human analysis of such data stream and improved based on data collected about performance in the industrial environment. Embodiments include a population of data collectors including at least one data collector for continuous ultrasonic monitoring of an industrial environment and at least one other type of data collector. Embodiments include using a distributed ledger to store time series data from the continuous ultrasonic monitoring across multiple devices. Embodiments include collecting the continuous ultrasonic data stream in self-organizing data collectors. Embodiments include collecting the continuous ultrasonic data stream in network-aware data collectors.

[0210] Embodiments include collecting a continuous ultrasound data stream at a remote organizational data collector. Embodiments include collecting a continuous ultrasound data stream at a data collector having self-organizing storage. Embodiments include transporting a stream of ultrasound data collected from an industrial environment using self-organizing network coding. Embodiments include communicating an indicator of a parameter of the continuously collected ultrasound data stream via a sensory interface of a wearable device. Embodiments include communicating an indicator of a parameter of the continuously collected ultrasound data stream via a heat map visual interface of a wearable device. Embodiments include communicating an indicator of a parameter of the continuously collected ultrasound data stream via an interface operating with self-organizing coordination of an interface layer.

[0211] As described above, methods and systems for cloud-based machine pattern recognition based on the fusion of remote analog industrial sensors are disclosed herein. Embodiments include capturing inputs from multiple analog sensors located in an industrial environment, multiplexing the sensors into a multiplexed data stream, sending the data stream to a cloud-deployed machine learning facility, and training a model in the machine learning facility to recognize defined patterns associated with the industrial environment. Embodiments include using the cloud-based pattern recognizer with input states from a state machine that characterizes the state of the industrial environment. Embodiments include deploying policies through a policy engine to govern which data is used by which users and for what purposes in the cloud-based machine learning. Embodiments include providing inputs from multiple devices with on-device storage of fused multiple sensor streams to the cloud-based pattern recognizer. Embodiments include generating output from the cloud-based machine pattern recognizer that analyzes fused data from the remote analog industrial sensors, which is available as a data service in a data marketplace. Embodiments include using a cloud-based platform to identify patterns in data across multiple data pools, including publicly available data from industrial sensors. An embodiment includes training a model to identify a preferred set of sensors for diagnosing industrial environment conditions, where the training set is created by a human user and the model is improved based on feedback from data collected about conditions in the industrial environment.

[0212] Embodiments include a swarm of data collectors governed by policies that are automatically propagated through the swarm. Embodiments include utilizing a distributed ledger to store sensor fusion information across multiple devices. Embodiments include providing input from a set of self-organizing data collectors to a cloud-based pattern recognizer that uses data from multiple sensors for the industrial environment. Embodiments include providing input from a set of network-aware data collectors to a cloud-based pattern recognizer that uses data from multiple sensors from the industrial environment. Embodiments include providing input from a set of remotely organized data collectors to a cloud-based pattern recognizer that determines user data from multiple sensors from the industrial environment. Embodiments include providing input from a set of data collectors with self-organizing storage to a cloud-based pattern recognizer that uses data from multiple sensors from the industrial environment. Embodiments include a system for collecting data in an industrial environment with self-organizing network coding for data transfer of fused data from multiple sensors in the industrial environment. Embodiments include communicating information formed by fusing inputs from multiple sensors in an industrial data collection system in a multi-sensory interface. Embodiments include communicating information formed by fusing inputs from multiple sensors in an industrial data collection system in a heat map interface. Embodiments include communicating information formed by fusing inputs from multiple sensors in an industrial data collection system to an interface operating with self-organizing coordination of an interface layer.

[0213] As mentioned above, methods and systems are disclosed herein for cloud-based machine pattern analysis of state information from a plurality of analog industrial sensors to provide predicted state information for an industrial system. Embodiments include providing cloud-based pattern analysis of state information from a plurality of analog industrial sensors to provide predicted state information for an industrial system. Embodiments include using a policy engine to determine which state information is available for cloud-based machine analysis. Embodiments include providing inputs from a plurality of devices having on-device storage of fused sensor streams to a cloud-based pattern recognizer to determine a predicted state of the industrial environment. Embodiments include generating predicted state information from a cloud-based machine pattern recognizer that analyzes fused data from remote analog industrial sensors available as a data service in a data marketplace. Embodiments include using the cloud-based pattern recognizer to determine a predicted state of the industrial environment based on data collected from a data pool including information streams from machines in the industrial environment. Embodiments include training a model to identify favorable state information to diagnose conditions in an industrial environment, where a training set is created by a human user and the model is improved based on feedback from data collected about conditions in the industrial environment. Embodiments include a collection of data collectors providing a state machine that maintains current state information for the industrial environment. Embodiments include using a distributed ledger to store historical state information for fused sensor states of self-organizing data collectors that provide the state machine that maintains current state information for the industrial environment. Embodiments include a network-aware data collector that provides the state machine that maintains current state information for the industrial environment. Embodiments include a remotely organized data collector that provides the state machine that maintains current state information for the industrial environment. Embodiments include a data collector with self-organizing storage that provides the state machine that maintains current state information for the industrial environment.Embodiments include a system for data collection in an industrial environment with self-organizing network coding for data transfer, which maintains predicted state information for the environment. Embodiments include communicating predicted state information determined by machine learning in an industrial data collection system with a multisensory interface. Embodiments include communicating predicted state information determined by machine learning in an industrial data collection system with a heat map interface. Embodiments include communicating predicted state information determined by machine learning in an industrial data collection system with an interface operating with self-organizing coordination of an interface layer.

[0214] As described above, disclosed herein are methods and systems for a cloud-based policy automation engine for IoT, including the creation, deployment, and management of IoT devices, including a cloud-based policy automation engine for IoT that enables the creation, deployment, and management of policies applied to IoT devices. Embodiments include deploying policies regarding data usage to an on-device storage system that stores fused data from multiple industrial sensors. Embodiments include deploying policies regarding what data can be provided to whom in a self-organizing marketplace for IoT sensor data. Embodiments include deploying policies across a set of self-organizing pools of data, including data streamed from industrial sensing devices, to manage the use of data from the pools. Embodiments include training a model to determine which policies to deploy in an industrial data collection system. Embodiments include deploying policies that manage how a self-organizing population should be organized for a particular industrial environment. Embodiments include storing policies on a device that manage the use of the device's storage capacity for a distributed ledger. Embodiments include deploying policies that manage how self-organizing data collectors are organized for a particular industrial environment. Embodiments include deploying policies that govern how network-aware data collectors should use network bandwidth for a particular industrial environment. Embodiments include deploying policies that govern how remotely organized data collectors should collect and make available data related to a particular industrial environment. Embodiments include deploying policies that govern how data collectors should self-organize storage for a particular industrial environment. Embodiments include a system for collecting data in an industrial environment, comprising a policy engine for deploying policies within the system and self-organizing network coding for data transfer.Embodiments include a system for collecting data in an industrial environment comprising a policy engine for deploying policies within the system, the policies governing how data is presented in a multi-sensory interface.Embodiments include a system for collecting data in an industrial environment comprising a policy engine for deploying policies within the system, the policies governing how data is presented in a heat map visual interface.Embodiments include a system for collecting data in an industrial environment comprising a policy engine for deploying policies within the system, the policies governing how data is presented in an interface that operates with self-organizing coordination of an interface layer.

[0215] As described above, methods and systems are disclosed herein for on-device sensor fusion and data storage for industrial IoT devices, including on-device sensor fusion and data storage for industrial IoT devices, where data from multiple sensors is multiplexed on the device for storing a fused data stream. Embodiments include a self-organizing marketplace presenting fused sensor data extracted from on-device storage of IoT devices. Embodiments include streaming fused sensor information from multiple industrial sensors and on-device data storage facilities to a data pool. Embodiments include training a model to determine which data to store on devices in a data collection environment. Embodiments include a self-organizing collective of industrial data collectors that organize among themselves to optimize data collection, at least some of the data collectors having on-device storage of fused data from multiple sensors. Embodiments include storing distributed ledger information with fused sensor information on industrial IoT devices. Embodiments include on-device sensor fusion and data storage for self-organizing industrial data collectors. Embodiments include on-device sensor fusion and data storage for network-aware industrial data collectors. Embodiments include on-device sensor fusion and data storage for remotely organized industrial data collectors. Embodiments include on-device sensor fusion and self-organizing data storage for industrial data collectors. Embodiments include a system for data collection in an industrial environment with on-device sensor fusion and self-organizing network coding for data transfer. Embodiments include a system for data collection with on-device sensor fusion of industrial sensor data, where data structures are stored to support alternative multi-sensory modes of presentation. Embodiments include a system for data collection with on-device sensor fusion of industrial sensor data, where data structures are stored to support a visual heatmap mode of presentation.Embodiments include a system for data collection with on-device sensor fusion of industrial sensor data, where data structures are stored to support interfaces that operate with self-organizing coordination of the interface layer.

[0216] As noted above, methods and systems are disclosed herein for a self-organizing data marketplace for industrial IoT data, including a self-organizing data marketplace for industrial IoT data, where available data elements are organized for consumption by consumers based on training a self-organizing facility with a training set and feedback from marketplace success measures. Embodiments include organizing a set of data pools within the self-organizing data marketplace based on utilization rates of the data pools. Embodiments include training a model to determine pricing for data in the data marketplace. Embodiments include providing data streams from a self-organizing collection of industrial data collectors to the data marketplace. Embodiments include using a distributed ledger to store transaction data for the self-organizing marketplace for industrial IoT data. Embodiments include providing data streams from self-organizing industrial data collectors to the data marketplace. Embodiments include providing data streams from a set of network-aware industrial data collectors to the data marketplace. Embodiments include providing data streams from a set of remotely organized industrial data collectors to the data marketplace. Embodiments include providing data streams from a set of industrial data collectors with self-organizing storage to a data marketplace. Embodiments include using self-organizing network coding to transfer data to a marketplace for sensor data collected in an industrial environment. Embodiments include providing a library of data structures suitable for presenting data in alternative multi-sensory interface modes in the data marketplace. Embodiments include providing a library to the data marketplace of data structures suitable for presenting data in a heat map visual image. Embodiments include providing a library to the data marketplace of data structures suitable for presenting data in an interface that operates with self-organizing coordination of an interface layer.

[0217] As noted above, methods and systems are disclosed herein for self-organizing data pools, including self-organizing data pools based on utilization and / or yield, including utilization and / or yield tracked for a plurality of data pools. Embodiments include training a model to present the most valuable data in a data marketplace, the training based on industry-specific success metrics. Embodiments include populating a set of self-organizing data pools with data from a self-organizing collection of data collectors. Embodiments include using a distributed ledger to store transaction information for data deployed in the data pools, the distributed ledger distributed across the data pools. Embodiments include self-organizing data pools based on utilization and / or yield tracked for a plurality of data pools, the pools including data from self-organizing data collectors. Embodiments include populating a set of self-organizing data pools with data from a set of network-aware data collectors. Embodiments include populating a set of self-organizing data pools with data from a set of remotely organized data collectors. Embodiments include populating a set of self-organizing data pools with data from a set of data collectors having self-organizing storage.Embodiments include a system for collecting data in an industrial environment having a self-organizing pool for data storage and self-organizing network coding for data transfer.Embodiments include a system for collecting data in an industrial environment having a self-organizing pool for data storage, the data storage including source data structures for supporting data presentation in a multisensory interface.Embodiments include a system for collecting data in an industrial environment having a self-organizing pool for data storage, the data storage including source data structures for supporting data presentation in a heat map interface.An embodiment includes a system for collecting data in an industrial environment having a self-organizing pool for data storage, the data storage including source data structures to support data presentation in an interface operating with self-organizing coordination of an interface layer.

[0218] As discussed above, disclosed herein are methods and systems for training an AI model based on industry-specific feedback, including training an AI model based on industry-specific feedback reflecting utilization, yield, or impact of the AI ​​model operating on sensor data from an industrial environment. Embodiments include training a population of data collectors based on the industry-specific feedback. Embodiments include training an AI model to identify and use available storage locations in the industrial environment to store distributed ledger information. Embodiments include training a population of self-organizing data collectors based on the industry-specific feedback. Embodiments include training network-aware data collectors based on network and industrial conditions in the industrial environment. Embodiments include training a remote organizer for remotely organized data collectors based on industry-specific feedback metrics. Embodiments include training the self-organizing data collectors to configure storage based on the industry-specific feedback. Embodiments include a system for collecting data in an industrial environment with cloud-based training of a network coding model to organize network coding for data transfer. Embodiments include a system for collecting data in an industrial environment having a cloud-based training facility that manages the presentation of data in a multi-sensory interface.Embodiments include a system for collecting data in an industrial environment having a cloud-based training facility that manages the presentation of data in a heat map interface.Embodiments include a system for collecting data in an industrial environment having a cloud-based training facility that manages the presentation of data in an interface that operates with self-organizing coordination of an interface layer.

[0219] As noted above, methods and systems are disclosed herein for an autonomous swarm of industrial data collectors, including a self-organizing collection of industrial data collectors that organize themselves to improve data collection based on the performance and status of the members of the collection. Embodiments include deploying a distributed ledger data structure across the data collection. Embodiments include a self-organizing collection of self-organizing data collectors for collecting data in an industrial environment. Embodiments include a self-organizing collection of network-aware data collectors for collecting data in an industrial environment. Embodiments include a self-organizing collection of network-aware data collectors for collecting data in an industrial environment, the collection also configured for remote organization. Embodiments include a self-organizing collection of data collectors with self-organizing storage for collecting data in an industrial environment. Embodiments include a system for collecting data in an industrial environment with a self-organizing collection of data collectors and self-organizing network coding for data transfer. Embodiments include a system for collecting data in an industrial environment having a self-organizing collection of data collectors that relay information for use in a multi-sensory interface.Embodiments include a system for collecting data in an industrial environment having a self-organizing collection of data collectors that relay information for use in a heat map interface.Embodiments include a system for collecting data in an industrial environment having a self-organizing collection of data collectors that relay information for use in an interface that operates with self-organizing coordination of an interface layer.

[0220] As noted above, methods and systems for distributed ledgers for the Industrial IoT are disclosed herein, including a distributed ledger that supports tracking of transactions performed in an automated data marketplace for Industrial IoT data. Embodiments include a self-organizing data collector configured to distribute collected information to the distributed ledger. Embodiments include a network-aware data collector configured to distribute collected information to the distributed ledger based on network conditions. Embodiments include a remotely organized data collector configured to distribute collected information to the distributed ledger based on intelligent remote management of distribution. Embodiments include a data collector having self-organizing local storage configured to distribute collected information to the distributed ledger. Embodiments include a system for collecting data in an industrial environment using a distributed ledger for data storage and self-organizing network coding for data transfer. Embodiments include a system for collecting data in an industrial environment using a distributed ledger for data storage with a data structure that supports a haptic interface for data presentation. Embodiments include a system for collecting data in an industrial environment using a distributed ledger for data storage with a data structure that supports a heat map interface for data presentation. Embodiments include a system for collecting data in an industrial environment that uses a distributed ledger for data storage of data structures that support interfaces that operate with self-organizing coordination of an interface layer.

[0221] As noted above, methods and systems for self-organizing collectors are disclosed herein, including self-organizing multi-sensor data collectors capable of optimizing data collection, power, and / or yield based on conditions within their environment. Embodiments include self-organizing data collectors that organize at least in part based on network conditions. Embodiments include self-organizing data collectors that are also responsive to remote organization. Embodiments include self-organizing data collectors with self-organizing storage for collected data in an industrial data collection environment. Embodiments include systems for collecting data in industrial environments with self-organizing network coding for self-organizing data collection and data transfer. Embodiments include systems for collecting data in industrial environments with self-organizing data collectors that provide data structures that support a tactile or multi-sensory wearable interface for data presentation. Embodiments include systems for collecting data in industrial environments with self-organizing data collectors that provide data structures that support a heatmap interface for data presentation. An embodiment includes a system for collecting data in an industrial environment having a self-organizing data collector that provides data structures supporting interfaces that operate in a self-organizing coordination of an interface layer.

[0222] As noted above, methods and systems for network-aware collectors are disclosed herein, including a network-aware, self-organizing, multi-sensor data collector that can optimize based on bandwidth, quality of service (QoS), pricing, and / or other network conditions. Embodiments include a remotely organized, network-aware, universal data collector that can power on and off sensor interfaces based on needs and / or conditions identified in an industrial data collection environment, including network conditions. Embodiments include a network-aware data collector with self-organizing storage for collected data in an industrial data collection environment. Embodiments include a network-aware data collector with self-organizing network coding for data transfer in an industrial data collection environment. Embodiments include a system for collecting data in an industrial environment with a network-aware data collector that relays data structures supporting a haptic wearable interface for data presentation. Embodiments include a system for collecting data in an industrial environment with a network-aware data collector that relays data structures supporting a heatmap interface for data presentation. An embodiment includes a system for collecting data in an industrial environment with a network-aware data collector relaying data structures supporting interfaces operating in a self-organizing coordination of an interface layer.

[0223] As described above, methods and systems are disclosed herein for a remotely organized universal data collector that can power on and off sensor interfaces based on needs and / or conditions identified in an industrial data collection environment. Embodiments include a remotely organized universal data collector with self-organizing storage for data collected in the industrial data collection environment. Embodiments include a system for collecting data in an industrial environment with remote control of data collection and self-organizing network coding for data transfer. Embodiments include a remotely organized data collector for storing sensor data and communicating instructions for data usage in a haptic or multisensory wearable interface. Embodiments include a remotely organized data collector for storing sensor data and communicating instructions for data usage in a heat map visual interface. Embodiments include a remotely organized data collector for storing sensor data and communicating instructions for data usage to an interface operating in self-organizing coordination of an interface layer.

[0224] As noted above, methods and systems for self-organizing storage for multi-sensor data collectors are disclosed herein, including self-organizing storage for multi-sensor data collectors for industrial sensor data. Embodiments include systems for collecting data in industrial environments with self-organizing data storage and self-organizing network coding for data transfer. Embodiments include a data collector with self-organizing storage for storing sensor data and instructions for transfer of data for use in a haptic wearable interface. Embodiments include a data collector with self-organizing storage for storing sensor data and instructions for transfer of data for use in a heat map presentation interface. Embodiments include a data collector with self-organizing storage for storing sensor data and instructions for transfer of data for use in an interface operating with self-organizing coordination of an interface layer.

[0225] As noted above, methods and systems for self-organizing network coding for multi-sensor data networks are disclosed herein, including self-organizing network coding for data networks that transfer data from multiple sensors in an industrial data collection environment. Embodiments include a system for collecting data in an industrial environment having self-organizing network coding for data transfer and a data structure that supports a haptic wearable interface for data presentation. Embodiments include a system for collecting data in an industrial environment with self-organizing network coding for data transfer and a data structure that supports a heat map interface for data presentation. Embodiments include a system for collecting data in an industrial environment having self-organizing network coding for data transfer and a self-organizing coordination function of an interface layer for data presentation.

[0226] As described above, methods and systems for haptic or multisensory user interfaces are disclosed herein, including wearable haptic or multisensory user interfaces for industrial sensor data collectors with vibration, heat, electrical, and / or audio outputs. Embodiments include a wearable haptic user interface for communicating industrial status information from a data collector using vibration, heat, electrical, and / or audio outputs. The wearable haptic user interface also has a visual presentation layer for presenting heat maps illustrating parameters of the data. Embodiments include state-sensitive, self-organizing adjustment of AR / VR and multisensory interfaces based on feedback metrics and / or training in an industrial environment.

[0227] As noted above, disclosed herein are methods and systems for a presentation layer for AR / VR industrial glasses in which heatmap elements are presented based on patterns and / or parameters of collected data. Embodiments include state-sensitive, self-organizing adjustment of the heatmap AR / VR interface based on feedback metrics and / or training in an industrial environment. As noted above, disclosed herein are methods and systems for state-sensitive, self-organizing adjustment of the AR / VR interface based on feedback metrics and / or training in an industrial environment.

[0228] The following exemplary sections describe specific embodiments of the present disclosure. The data collection system referred to in the following disclosure can be a local data collection system 102, a host processing system 112 (e.g., using a cloud platform), or a combination of a local system and a host system. In an embodiment, a data collection system is provided that includes the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs. In an embodiment, a data collection and processing system is provided that includes the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has IP front-end signal conditioning on the multiplexer for improved signal-to-noise ratio. In an embodiment, a data collection and processing system is provided that includes the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has continuous monitoring alarm functionality for the multiplexer. In an embodiment, a data collection and processing system is provided that includes the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has multiple multiplexers and a distributed CPLD chip with a dedicated bus for logic control of the data acquisition section. In embodiments, a data acquisition and processing system is provided that has the use of analog crosspoint switches to acquire data with a variable group of analog sensor inputs and has high current input capability using solid state relays and design topology. In embodiments, a data acquisition and processing system is provided that has the use of analog crosspoint switches to acquire data with a variable group of analog sensor inputs and has power disconnect capability for at least one of the analog sensor channels and component boards. In embodiments, a data acquisition and processing system is provided that has the use of analog crosspoint switches to acquire data with a variable group of analog sensor inputs and has unique electrostatic protection for trigger and vibration inputs.In an embodiment, a data acquisition and processing system is provided that has the use of analog crosspoint switches to acquire data with a variable group of analog sensor inputs, and has a precision voltage reference for the A / D zero reference.

[0229] In an embodiment, a data acquisition and processing system is provided that uses an analog crosspoint switch to acquire data with a variable group of analog sensor inputs and has a phase-locked loop bandpass tracking filter to obtain slow RPM and phase information. In an embodiment, a data acquisition and processing system is provided that uses an analog crosspoint switch to acquire data with a variable group of analog sensor inputs and has digital differentiation of phase relative to the input and trigger channels using an onboard timer. In an embodiment, a data acquisition and processing system is provided that uses an analog crosspoint switch to acquire data with a variable group of analog sensor inputs and has a peak detector for autoscaling located in a separate analog-to-digital converter to detect peaks. In an embodiment, a data acquisition and processing system is provided that uses an analog crosspoint switch to acquire data with a variable group of analog sensor inputs and has routing of a buffered trigger channel to unprocessed or other analog channels. In an embodiment, a data acquisition and processing system is provided that uses an analog crosspoint switch to acquire data with a variable group of analog sensor inputs and has higher input oversampling to the delta-sigma A / D for a lower sampling rate output to minimize AA filter requirements. In an embodiment, a data acquisition and processing system is provided that includes the use of an analog crosspoint switch to acquire data with a variable group of analog sensor inputs, and the use of a CPLD as a clock divider for a delta-sigma analog-to-digital converter to achieve low sampling rates without the need for digital resampling.

[0230] In embodiments, a data collection and processing system is provided that uses analog crosspoint switches to collect data with variable groups of analog sensor inputs and has long blocks of data at a high sampling rate as opposed to multiple data sets acquired at different sampling rates. In embodiments, a data collection and processing system is provided that uses analog crosspoint switches to collect data with variable groups of analog sensor inputs and has storage of calibration data including the maintenance history of an on-board card set. In embodiments, a data collection and processing system is provided that uses analog crosspoint switches to collect data with variable groups of analog sensor inputs and has fast route generation using hierarchical templates. In embodiments, a data collection and processing system is provided that uses analog crosspoint switches to collect data with variable groups of analog sensor inputs and has intelligent management of data collection bands. In embodiments, a data collection and processing system is provided that uses analog crosspoint switches to collect data with variable groups of analog sensor inputs and has a neural net expert system that uses intelligent management of data collection bands.

[0231] In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has the use of a database hierarchy in sensor data analysis. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has an expert system GUI graphical approach for defining intelligent data collection bands and diagnostics for the expert system. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has a graphical approach for backcalculation definition. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has a proposed bearing analysis method. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has torsional vibration detection / analysis using transient signal analysis. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has improved integration with both analog and digital methods.

[0232] In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has an adaptive scheduling technique for continuous monitoring of analog data in a local environment. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has a data acquisition parking function. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has a self-contained data collection box. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has SD card storage. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has expanded on-board statistical capabilities for continuous monitoring. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data with variable groups of analog sensor inputs and has the use of ambient noise, local noise, and vibration noise for prediction. In an embodiment, a data collection and processing system is provided that has the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has a smart route function that changes the route based on the input data or alarms to simultaneously enable dynamic data for analysis or correlation.In an embodiment, a data collection and processing system is provided that has the use of an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has a smart ODS and transfer function.In an embodiment, a data acquisition and processing system is provided that has an analog crosspoint switch for collecting data with a variable group of analog sensor inputs and that has a hierarchical multiplexer. In an embodiment, a data acquisition and processing system is provided that has an analog crosspoint switch for collecting data with a variable group of analog sensor inputs and that has a sensor overload identification function. In an embodiment, a data acquisition and processing system is provided that has an analog crosspoint switch for collecting data with a variable group of analog sensor inputs and that has a radio frequency identification function (RF identification) and an inclinometer.

[0233] In embodiments, a data collection and processing system is provided that uses an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has continuous ultrasonic monitoring capabilities. In embodiments, a data collection and processing system is provided that uses an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has cloud-based machine pattern recognition capabilities based on remote analog industrial sensor fusion. In embodiments, a data collection and processing system is provided that uses an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has cloud-based machine pattern analysis capabilities of status information from multiple analog industrial sensors to provide predicted status information for an industrial system. In embodiments, a data collection and processing system is provided that uses an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has a cloud-based policy automation engine for IoT, including IoT device creation, deployment, and management. In embodiments, a data collection and processing system is provided that uses an analog crosspoint switch to collect data with a variable group of analog sensor inputs and has on-device sensor fusion capabilities and data storage for industrial IoT devices. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data having a variable group of analog sensor inputs and has a self-organizing data marketplace for industrial IoT data. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data having a variable group of analog sensor inputs and has the ability to self-organize a data pool based on utilization and / or yield. In embodiments, a data collection and processing system is provided that has the use of analog crosspoint switches to collect data having a variable group of analog sensor inputs and trains AI models based on industry-specific feedback.In embodiments, a data collection and processing system is provided that includes the use of analog crosspoint switches to collect data with a variable group of analog sensor inputs and includes an autonomous swarm of industrial data collectors. In embodiments, a data collection and processing system is provided that includes the use of analog crosspoint switches to collect data with a variable group of analog sensor inputs and includes an IoT distributed ledger. In embodiments, a data collection and processing system is provided that includes the use of analog crosspoint switches to collect data with a variable group of analog sensor inputs and includes a self-organizing collector. In embodiments, a data collection and processing system is provided that includes the use of analog crosspoint switches to collect data with a variable group of analog sensor inputs and includes a network-aware collector. In embodiments, a data collection and processing system is provided that includes the use of analog crosspoint switches to collect data with a variable group of analog sensor inputs and includes a remotely organized collector. In embodiments, a data collection and proces...

Claims

1. 1. A system for data collection, processing and utilization of signals from machines in an industrial environment, comprising: a local data acquisition system having at least one sensor signal obtained from a machine in an industrial environment; a platform including a computing environment connected to a local data collection system; and a local data collection system sensor configured to be connected to the machine to acquire at least one sensor signal from the machine; the local data collection system includes at least one input and at least one output connected to the sensor; the local data acquisition system is configured to record uninterrupted digital waveform data from at least one input; The local data acquisition system includes a neural net expert system configured to provide intelligent management of data acquisition bandwidth from the recorded uninterrupted digital waveform data, the system.

2. the local data collection system includes a plurality of inputs and a plurality of outputs, with at least one input connected to the sensor and another input connected to another sensor; the plurality of outputs including the at least one output and the other output configured to select between at least one of a state in which the at least one output is configured to select between transmitting the at least one sensor signal and transmitting another sensor signal, or a state in which transmitting the at least one sensor signal from the at least one output and transmitting the other sensor signal from the other output are simultaneous; each of the plurality of inputs is individually assigned to one of the plurality of outputs; The system of claim 1 , wherein the local data acquisition system is configured to record uninterrupted digital waveform data from at least one input and the other input simultaneously.

3. 10. The system of claim 1, wherein the neural net expert system includes machine learning using at least one neural network to provide intelligent management of the data collection bands.

4. 10. The system of claim 1, wherein the neural net expert system includes machine learning configured to adjust at least one of the weights, structure, or rules based on feedback including at least one or more inputs or measures of success for training or improving an initial model.

5. 10. The system of claim 1, wherein the neural net expert system is configured to define smart bands from the data collection bands, and the neural net expert system is further configured to pair the smart bands with at least one neural network for providing machine diagnostics, the smart bands referring to at least one of a specific frequency band or a group of spectral peaks.

6. 6. The system of claim 5, wherein the group of spectral peaks comprises signal attributes further comprising at least one of harmonics of a single peak, a true peak level, a crest factor derived from a time waveform, a global derivation from a vibration envelope spectrum, or a logical combination of signal attributes.

7. 2. The system of claim 1, wherein the neural net expert system is configured to define smart bands from the data collection bands, and wherein the neural net expert system uses the smart bands in a neural approach utilizing weighted triggering of multiple input stimuli to a neuron group that supplies a weighted output to another neuron group, the output of the neuron group being classified as a smart band that supplies the other neuron group.

8. The system of claim 1 , wherein the local data collection system is configured to create data collection paths based on hierarchical templates that each include a data collection band associated with a machine associated with the data collection path.

9. 9. The system of claim 8, wherein at least one of the hierarchical templates is associated with similar elements associated with the machine and another machine, and at least one of the hierarchical templates is associated with the machine being located in proximity to the other machine.

10. 10. The system of claim 1, wherein the local data collection system includes a graphical user interface system configured to use the neural net expert system to manage data collection bandwidth to provide diagnostics of the machine.

11. The system of claim 1 , wherein the platform is configured to determine relative phase changes based on recorded uninterrupted digital waveform data.

12. The system of claim 11 , wherein the platform is configured to determine an operational deflection shape based on changes in relative phase and recorded uninterrupted digital waveform data.

13. 10. The system of claim 1, wherein the local data collection system is configured to acquire the recorded uninterrupted digital waveform data from the machine while the machine and the other machine are both operating, and the local data collection system is configured to characterize the contributions of the machine and the other machine in the recorded uninterrupted digital waveform data.

14. 10. The system of claim 1, wherein uninterrupted digital waveform data is obtained simultaneously from each sensor having a maximum resolvable frequency large enough to capture periodic and transient impact events.

15. 1. A computer-implemented method for data collection, processing, and utilization of signals from machines in an industrial environment, comprising: connecting sensors of the local data collection system to machines in the industrial environment; acquiring at least one sensor signal from the machine; recording uninterrupted digital waveform data from at least one input of a sensor connected to the machine; and The method includes the step of: a neural net expert system intelligently managing data collection bandwidth from the recorded uninterrupted digital waveform data.

16. 16. The method of claim 15, wherein the neural net expert system includes machine learning using at least one neural network to provide intelligent management of the data collection bandwidth.

17. 16. The method of claim 15, wherein providing the intelligent management further includes defining a smart band from the data collection band and pairing the smart band with at least one neural network for providing diagnostics of the machine, the smart band referencing at least one of a specific frequency band or a set of spectral peaks.

18. 16. The method of claim 15, wherein providing the intelligent management further comprises defining a smart band from the data collection band and using the smart band in a neural approach utilizing weighted triggering of multiple input stimuli to a neuron group that provides a weighted output to another neuron group, wherein the output of the neuron group is classified as a smart band that provides another neuron group.

19. acquiring another sensor signal from a machine in an industrial environment, the another sensor signal being distinct from the at least one sensor signal; Among the plurality of outputs, a first output and a second output are the first output is configured to select between delivery of at least one sensor signal and the other sensor signal; or a state in which at least one sensor signal from the first output and another sensor signal from the second output are simultaneously transmitted; and individually assigning one or more inputs to any of the plurality of outputs; 16. The method of claim 15, wherein the recorded uninterrupted digital waveform data is provided from at least one input and another input simultaneously.

20. A non-transitory computer-readable storage medium having stored thereon a plurality of instructions, the plurality of instructions, when executed by one or more processors, causing at least a portion of the one or more processors to: Connecting sensors of a local data collection system to machines in an industrial environment; acquiring at least one sensor signal from the machine; recording uninterrupted digital waveform data from at least one input of a sensor connected to the machine; A transient computer-readable storage medium that causes a neural net expert system to perform operations including intelligently managing data collection bandwidth from recorded uninterrupted digital waveform data.

21. 21. The computer-readable storage medium of claim 20, wherein the neural net expert system includes machine learning that uses at least one neural network to provide intelligent management of the data collection bandwidth.

22. 21. The computer-readable storage medium of claim 20, wherein the operations performed by the one or more processors further include obtaining recorded uninterrupted digital waveform data from the machine while the machine and another machine are operating together, and the local data collection system is configured to characterize contributions from the machine and the another machine in the recorded uninterrupted digital waveform data.

Images