Processing machine protection and fault prediction data natively in a distributed control system

Abstract

Machine operating state monitoring module (10) which processes machine vibration data based on vibration signals and provides the machine vibration data to a distributed control system (11), wherein the machine operating state monitoring module (10) comprises: a signal conditioning circuit arrangement with an interface for receiving multiple analog vibration signals from multiple sensors attached to a machine (20), an amplification and filter circuit arrangement for conditioning the multiple analog vibration signals and an analog-to-digital conversion circuit arrangement for converting the multiple analog vibration signals into multiple digital vibration signals; a processing circuit arrangement in electrical connection with the signal conditioning circuit arrangement, wherein the processing circuit arrangement comprises several parallel digital signal processing channels, each channel serving to process one corresponding of the several digital vibration signals to generate several scalar vibration values ​​per channel and at least one vibration time signal shape per channel, wherein the several parallel digital processing channels include: a first channel for processing a vibration time signal waveform for use in machine condition prediction processing; and a second channel for processing a vibration time signal waveform for use in machine protection processing, wherein the processing performed in the second channel is independent of the processing performed in the first channel; and a logic generator circuit arrangement in electrical connection with the processing circuit arrangement, wherein the logic generator circuit arrangement is operable to receive the multiple scalar oscillation values ​​and the at least one oscillation time signal shape and is operable to format at least the multiple scalar oscillation values ​​in accordance with an input / output data protocol that is native to the distributed control system.

Description

AREA

[0001] This invention relates to the field of machine control and machine condition monitoring. In particular, this invention relates to a system for integrating a machine vibration data acquisition and machine vibration data analysis module directly into a distributed control system architecture as a native data input device. BACKGROUND

[0002] In earlier machine control and machine vibration monitoring systems, the numerical data generated by a machine vibration monitoring system had to be integrated with data generated by a machine control system using intermediary communication protocols to bridge the systems, such as those based on the OPC standard, Modbus, or Profibus (process fieldbus). Using the conventional integration process, common communication protocols, protocol configuration, data networking, data synchronization, and manual data mapping were required. Testing and troubleshooting were necessary to verify the proper operation of the combined system.

[0003] In conventional machine protection applications, binary relay outputs representing machine alarm states or trip states, and analog current loop outputs of 4 to 20 mA from a vibration monitoring program were used as hard-wired inputs to a distributed control system (DCS) for trip initiation and vibration values.

[0004] In conventional vibration monitoring systems, trigger levels and alarms were configured separately from the control system configuration and display software and displayed by the vibration monitoring systems. Vibration data was only available to the control system if the system integration procedures described above were implemented to acquire the vibration data from the vibration monitoring system.

[0005] Therefore, a system is needed to make numerical data generated by a machine vibration monitoring system available to a machine control system without having to use intermediate communication protocols to connect the systems.

[0006] US 2003 / 0200060 A1 discloses a monitoring device for rotating equipment, comprising a set of sensors, such as temperature, speed, pressure, and vibration sensors, locally connected to a diagnostic unit. The diagnostic unit has an equipment communication interface and a processor that implements one or more diagnostic routines. The diagnostic unit collects data from the sensors and uses this data to perform diagnostics in order to detect actual or potential problems with the rotating equipment on a continuous or semi-continuous basis. Using the communication interface, the diagnostic unit automatically transmits signals indicating the detected states of the rotating equipment to a maintenance or control system via a standard or common communication network, such as a HART or fieldbus network.The diagnostic unit can also be temporarily connected to an offline computer, such as a service provider's computer, which can verify or further define the diagnostic data provided by the diagnostic unit.

[0007] US 5,895,857 A discloses a signal processing device for processing a machine vibration signal and includes a peak detector for determining the peak values ​​of the vibration amplitude during predetermined sampling periods. In one embodiment, the vibration signal is digitally processed by a digital peak detector before being received. In another embodiment, analog processing and peak detection of the vibration signal are employed. The peak values ​​are then provided for further processing. In many situations, the peak values ​​are associated with mechanical defects, which are detected at a periodic rate corresponding to the defect frequency of mechanical components, without requiring additional signal amplification. A further improvement can be achieved by synchronously averaging the peak values ​​with the speed of a rotating element.A signal representing the velocity of a rotating element (target element) can be directly obtained from a velocity sensor that measures the target element's velocity. If the target element is inaccessible, a corrected pseudo-velocity signal can be calculated from the velocity of an accessible rotating element connected to the target element. By transforming the synchronous amplitudes into the frequency domain using a fast Fourier transform, information can be obtained to assess whether or not a fault exists in the rotating machine element.

[0008] US 2012 / 0044015 A1 discloses that a universal digital input module is provided in a process automation controller. The universal digital input module comprises a plurality of digital input channels, each channel receiving a first current at a first voltage level associated with a digital high input, and receiving a second current at a second voltage level associated with the digital high input, the first current being greater than the second current and the first voltage being less than the second voltage.

[0009] US 2012 / 0041695 A1 discloses a vibration data acquisition system that performs a real-time integration or differentiation process on incoming digitized vibration data. The system uses a digital IIR (Infinite Impulse Response) filter that runs at the input data rate to provide the integration or differentiation function. This approach reduces hardware complexity and data storage requirements. Furthermore, the system offers the ability to directly integrate or differentiate stored time waveforms without relying on FFT processing methods.

[0010] It is an object of the invention to provide an improved machine operating state monitoring module, a distributed control system, and a method for processing data in a distributed control system. This object is achieved by the subject matter of the independent claims. Advantageous embodiments are defined in the dependent claims. SUMMARY

[0011] Embodiments of the present invention provide a vibration data acquisition and analysis module that can be operated by being directly inserted into a DCS I / O back panel, in particular a DCS I / O back panel board, so that processed vibration parameters can be sampled directly by the DCS I / O controller. Since the process data and the vibration data are both sampled by the same DCS I / O controller, there is no need to integrate numerical data, binary relay outputs, and analog total vibration level outputs from a separate vibration monitoring system into the process control system.Further advantages of embodiments of the invention include: (1) direct acquisition of vibration data by the control system for machine protection and predictive machine condition analysis; (2) direct integration of vibration information into DCS alarm screens; (3) acquisition and display of real-time vibration data on operator screens; (4) the ability to use vibration data to detect anomalous situations associated with equipment faults; and (5) the ability to use vibration data directly in control applications. Machine protection / machine prediction natively in DCS

[0012] Some embodiments of the invention provide a vibration data acquisition system that collects and processes machine protection data and machine prediction data in a software format that is, for example, native to the DCS. As the term is used here, software is "native" to a platform if it is designed to run on that platform, where the platform may be an operating system or a device such as a DCS controller. The system includes vibration modules that calculate total scalar vibration parameters from vibration waveforms in a DCS machine operating state monitoring module. These total scalar values ​​are preferably calculated using parallel digital signal processing in a field card field programmable gate array (FPGA). These processed scalar vibration values ​​are transmitted to the DCS I / O controller via the conventional serial DCS I / O bus.

[0013] The scalar vibration values ​​are sampled and processed by logic routines (referred to here as "control sheets") that are executed at a deterministic rate in the DCS controller. The output of the control sheet logic is preferably transmitted to DCS output modules to execute a machine shutdown or other control functions. In various embodiments, control sheets can be optimized for maximum protection (stricter machine protection) or for maximum availability (relaxed protection to minimize nuisance or false tripping events).

[0014] Preferred embodiments of the invention enable the transmission of time-wave waveform data from the vibration data acquisition system, for example via Ethernet, and the viewing of the waveform data on a machine function management analysis computer. Some embodiments also enable the movement of time-wave waveform blocks on the DCS I / O bus. In these embodiments, the time-wave waveform block data can be transferred to the DCS controller via the serial DCS I / O bus backplane (board) using the DCS Remote Desktop Protocol (DCS-RDP). Separation of protection and prediction

[0015] Preferred embodiments of the invention also provide a separation between machine protection functions and machine prediction functions. In particular, prediction data acquisition and processing do not interfere with protection data acquisition and processing, since the two data streams are processed in the signal processing FPGA via separate, independent data paths. Furthermore, preferred embodiments implement separate physical ports for protection data and prediction data. Prediction data is accessed by machine function state management (MHM) software via a dedicated Ethernet port, which can be disabled by the DCS configuration software. Protection data is transferred to the DCS controller via the DCS I / O backplane or is made available via a separate dedicated Ethernet port.Furthermore, prediction components cannot "write" to protection components, and separate configuration and data storage is provided.

[0016] In preferred embodiments, all protection hardware configuration functions are handled solely by DCS software, although MHM prediction software can access the protection configuration data to determine a sensor and measurement configuration.

[0017] Furthermore, the MHM prediction software can only control the prediction time signal waveform. Although the prediction software can read overall protection values, it cannot influence the configuration of the overall level measurements used for the protection functions. Additionally, in preferred embodiments, the prediction tasks are executed with lower priority in the real-time operating system (RTOS) using preemptive multitasking.

[0018] Preferably, there are separate software hosts that consume the protection and prediction data. A DCS software host processes the protection data, while an MHM software host processes the prediction data.

[0019] One embodiment of the invention relates to a machine function monitoring module (MHM module) that processes machine vibration data based on vibration signals and provides this data to a distributed control system (DCS). The MHM module comprises a signal conditioning circuit, a processing circuit, and a logic generator circuit. The signal conditioning circuit has an interface for receiving analog vibration signals from sensors mounted on a machine, an amplification and filtering circuit for conditioning the analog vibration signals, and an analog-to-digital conversion circuit for converting the analog vibration signals into digital vibration signals. The processing circuit includes multiple parallel digital signal processing channels.Each channel processes a corresponding digital oscillation signal to generate multiple scalar oscillation values ​​per channel and at least one oscillation time waveform per channel. The logic generator circuitry receives the multiple scalar oscillation values ​​and the oscillation time waveform and formats them according to an input / output data protocol, such as the one native to the DCS.

[0020] In some embodiments, the logic generator circuit arrangement includes a backplane interface, in particular a backplane board interface, configured to connect and disconnect electrically and mechanically from an input / output bus of the DCS. In these embodiments, the multiple scalar oscillation values ​​are transmitted via the backplane interface and the input / output bus of the DCS using the input / output data protocol that is native to the DCS, for example.

[0021] In some embodiments, the logic generator circuit arrangement includes a first network communication interface that is independent of the backplane interface. The first network communication interface transmits the oscillation time signal waveform to a machine function state prediction data analysis computer via a communication network for processing.

[0022] In some embodiments, the logic generator circuit arrangement includes a second network communication interface that is independent of the backplane interface. This second network communication interface can be used to transmit the oscillation time signal waveform via a communication network to a DCS operator computer for machine protection processing.

[0023] In preferred embodiments, the first and second network communication interfaces are Ethernet interface ports.

[0024] In some embodiments, the logic generator circuit arrangement transmits the oscillation time signal shape in data blocks via the backplane interface and the input / output bus of the DCS using a block data transfer protocol.

[0025] The processing circuit arrangement's multiple parallel digital processing channels include a first channel for processing a vibration time waveform for use in machine condition prediction processing and a second channel for processing a vibration time waveform for use in machine protection processing. The processing performed in the second channel is independent of the processing performed in the first channel.

[0026] Another embodiment of the invention relates to a distributed control system (DCS) comprising an input / output bus, one or more multi-function memory modules (MHMs), one or more DCS input modules, a DCS controller, and one or more output modules. The input / output bus transmits data in accordance with a data communication protocol, which is, for example, native to the DCS. Each MHM module contains a signal conditioning circuit, a processing circuit, and a logic generator circuit, as described above. The DCS input modules receive sensor signals from process sensors mounted on a machine, generate scalar process values ​​based on the sensor signals, and provide the scalar process values ​​to the input / output bus. The process sensors include sensors other than vibration sensors. The DCS controller comprises an interface circuit and a logic circuit.The interface circuitry samples the input / output bus at a predefined rate to receive the scalar vibration values ​​and the scalar process values. The logic circuitry executes control logic routines that generate control signals based on the logic processing of one or more of the scalar vibration values, one or more of the scalar process values, or a combination of scalar vibration values ​​and scalar process values. The DCS output modules receive the control signals from the input / output bus and generate machine operating output signals based on these control signals.

[0027] In some embodiments, the logic circuit arrangement of the DCS controller executes the control logic routines at the same predetermined rate and generates the control signals at the same predetermined rate at which the interface circuit arrangement samples the scalar vibration values ​​and the scalar process values ​​from the input / output bus.

[0028] In some embodiments, the DCS input modules generate scalar temperature values, scalar pressure values, scalar flow values, and scalar speed values.

[0029] In some embodiments, each MHM module generates scalar vibration values ​​that include an RMS value, a peak value, a peak-to-peak value, a DC value, an absolute ±peak value, and a PeakVue value.

[0030] In some embodiments, the logic circuitry of the DCS controller selectively executes logic control routines (also referred to here as control sheets) that are optimized for different purposes. Some logic control routines are optimized for maximum machine protection using first trip threshold levels. Other logic control routines, optimized for maximum machine availability, use second trip threshold levels that are higher than the first trip threshold levels.

[0031] In one aspect, an embodiment of the invention provides a method for processing data in a DCS. The data is based on sensor signals generated by sensors mounted on one or more machines controlled by the DCS. The method includes: (a) the reception of analog vibration signals from sensors attached to one or more machines; (b) the conversion of analog vibration signals into digital vibration signals; (c) the simultaneous processing of the digital vibration signals in multiple parallel digital signal processing channels, wherein the processing includes: (c1) the processing of the digital vibration signals in one or more of the parallel digital signal processing channels that are solely intended for machine condition prediction processing; and (c2) the processing of digital vibration signals in one or more of the parallel digital signal processing channels which are intended solely for machine condition protection processing, wherein the digital signal processing channels intended for machine function state prediction processing are separate and independent from the digital signal processing channels intended for machine function state protection processing; (d) the generation of machine operating output signals based on machine function status protection processing; (e) the generation of machine performance signals based on machine operating state prediction processing; (f) using the machine operating output signals in the DCS to shut down one or more of the machines to prevent damage; and (g) using the machine performance signals to monitor trends in machine performance or to predict how much longer one or more of the machines can operate before they are taken out of service for repair or replacement.

[0032] In some embodiments, the processing steps (c1) and (c2) are performed in separate and independent parallel channels of a Field Programmable Gate Array (FPGA).

[0033] In some embodiments, processing step (c1) may not affect processing step (c2).

[0034] In some embodiments, data associated with processing step (c1) are stored in first memory locations and data associated with processing step (c2) are stored in second memory locations, wherein the second memory locations can be read by processing step (c1) but cannot be written to by processing step (c1).

[0035] In some embodiments, the processing step (c1) in the real-time operating system that controls the task priority in the DCS has a lower priority than the processing step (c2).

[0036] In some embodiments, step (f) includes providing the machine operating output signals to the distributed control system via an input / output bus, and step (g) includes providing the machine performance signals to a machine function status management data analysis computer via a network communication interface that is independent of the input / output bus. In some embodiments, the method includes generating control signals in the DCS to selectively disable the network communication interface. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Further embodiments of the invention will become apparent with reference to the detailed description in conjunction with the figures, wherein elements are not to scale in order to show the details more clearly, wherein the same reference numerals denote the same elements throughout the multiple views, and wherein: Fig. 1 shows a machine function status monitoring module (MHM module) according to an embodiment of the invention; Fig. 2 shows a digital field FPGA signal processing circuit arrangement according to an embodiment of the invention; and Fig. Figure 3 shows an example of control logic implemented by a DCS controller according to an embodiment of the invention. DETAILED DESCRIPTION

[0038] Embodiments of the present invention provide a vibration data acquisition and analysis module that is directly connected via an interface to an I / O backplane of a distributed control system to enable the direct acquisition of vibration data by the DCS for machine protection and predictive machine condition analysis. As used here, a "distributed control system (DCS)" is a type of automated control system used in a process or plant in which control elements are distributed over an entire machine or over several machines to provide operating instructions for different parts of the machine(s). As used here, "protection" refers to the use of data obtained from one or more sensors (vibration, temperature, pressure, etc.).) are collected to shut down a machine in situations where serious and costly damage could occur if the machine is allowed to continue running. On the other hand, "prediction" refers to the use of data collected by one or more vibration sensors, possibly in combination with data from other types of sensors, to observe trends in machine performance and to predict how much longer a machine can operate before it should be taken out of service for maintenance or replacement.

[0039] Fig. Figure 1 shows a machine function status monitoring module (MHM) 10, which is directly connected to a DCS 11 via an interface. In a preferred embodiment, the module 10 includes an analog field signal conditioning and sensor power card 12, which exemplarily comprises a signal conditioning circuit arrangement and receives and conditions the sensor signals; a digital field FPGA signal processing card 14, which exemplarily comprises a processing circuit arrangement and processes the sensor signals; and a DCS logic generator card (LGC) 16, which exemplarily comprises a logic generator circuit arrangement and provides an interface to a DCS I / O bus 18. Preferably, the field card 12 can accept input from up to eight measuring sensors 20 via a field signal interface connector 22. In a preferred embodiment, two of the sensor input signals can be configured as tachometer channels.

[0040] Preferably, galvanic electrical isolation is provided between the analog field card 12 and the digital field card 14. This electrical isolation prevents unintended current flow, such as that caused by ground loops between the mounting locations of the sensors 20 and the DCS 11.

[0041] The sensor power circuit 24 and the signal conditioning circuit 25 can support a wide range of sensors 20, including piezoelectric accelerometers, piezoelectric ICPs for velocity, piezodynamic sensors for pressure, electrodynamic sensors for velocity, eddy current sensors for displacement, AC sensors for vibration, and DC sensors for displacement. Supported speed sensors include eddy current displacement sensors, passive electromagnetic sensors, Hall-effect tachometer sensors, N-pulse / speed wave encoders, and TTL pulse sensors. As long as they are within the following exemplary voltage input ranges: 0 to +24 V, -24 V to +24 V, -12 V to +12 V, and 0 to -24 V, many additional sensor types are supported across the frequency range from DC to 20 kHz.In the preferred embodiment, up to eight sensor power circuits can be individually programmed for a constant current between 0 and 20 mA, which can also be used as the lifting current for an electrodynamic (passive) velocity sensor. Constant voltage sources (+24 V DC or -24 V DC) can also be selected as the constant current. The input voltage ranges listed above are also individually programmable in each sensor channel. This allows any combination of sensor power and input range configuration between the channels, thereby enabling a combination of supported sensors.

[0042] With timing provided by a clock 26, an 8-channel analog-to-digital converter (8-channel ADC) 28 converts the eight analog signals into a single serial data stream comprising eight simultaneously sampled, interleaved channels of data. In some preferred embodiments, two tachometer trigger circuits 30 convert the signals from the two analog tachometers into tachometer pulses.

[0043] Field card 14 features an 8-channel field-programmable gate array (8-channel FPGA) 36 for processing the vibration data. The FPGA 36 receives the digital 8-channel waveform data and the 2-channel tachometer data and processes the output data in parallel to generate scalar total vibration parameters and total vibration waveforms. The processed waveforms can include low-pass filtered, PeakVue™, order-tracking, high-pass filtered (DC-blocked), and optionally single-integrated (velocity), dual-integrated (displacement), or non-integrated (acceleration) waveforms. Additionally, prediction data channels preferably include an upsampling data block to provide higher-resolution data for time-synchronous averaging (TSA) or order-tracking applications.

[0044] The vibration card configuration circuit 32 of the analog field card 12 preferably includes a set of serial-parallel buffer registers that accept a serial data stream of configuration data from the application firmware of the LGC 16. This data is loaded into a parallel-serial shift register in the interface of the FPGA 36. Subsequently, the FPGA 36 handles the shifting of the serial data to the control buffers using a synchronous SPI format.

[0045] During operation of the preferred embodiment, the MHM module 10 appears to the DCS controller 19 as an analog multi-channel input card with scalar outputs similar to those of a standard DCS input module 21, allowing measured temperature, pressure, or valve position values ​​to be output. As discussed in more detail below, vibration signals are converted into scalar values ​​by the module 10 and passed to the DCS controller 19 via the DCS backplane. An example of a DCS controller 19 is the Ovation™ controller, manufactured by Emerson Process Management (a division of Emerson Electronic Co.). In the typical DCS architecture, only sixteen scalar values ​​are passed to the DCS controller 19 as high-speed samples. In a high-speed sample, the DCS controller 19 can read these sixteen scalar values ​​at a rate of up to 10 ms.

[0046] The time signal shape block data (and some scalar values) can be transferred to the DCS controller 19 via the DCS I / O bus 18 using a block data transfer method such as the Remote Desktop Protocol (RDP) at a rate lower than the sampling rate of the sixteen scalar values.

[0047] Since the scalar values ​​generated by the machine status monitoring module 10 are read by the DCS controller 19, they are processed by software running on the DCS controller 19 in the same way as any other DCS data. A key function of the DCS controller 19 is to compare the scalar values ​​with alarm thresholds. If the thresholds are exceeded, alarms are generated. Logic within the DCS controller 19 can also determine, based on alarm conditions such as the closing of a relay, whether any action should be taken. The DCS controller 19 also performs software operations that include alarm relay logic, tuning, and time delays. Preferably, DCS control outputs, such as relay outputs and proportional 4-20 mA outputs, are driven by standard output modules 23 of the DCS.Predictive mass data is formatted in the LGC host processor 48 and transmitted via an Ethernet port 52a to a machine function management analysis computer (MHM analysis computer) 54 for detailed analysis and display. Protection mass data is also formatted in the LGC host processor 48, but transmitted via a separate Ethernet port 52b to the DCS operator computer 60.

[0048] In preferred embodiments, a DCS operator computer 60 includes an interface for displaying vibration parameters and other machine operating data (pressures, temperatures, speeds, alarm conditions, etc.) output by the DCS controller 19.

[0049] In Fig. Figure 2 shows a functional block diagram of a single channel of the digital field FPGA 36. A preferred embodiment includes seven additional channels with the same arrangement as that of the single channel shown in Figure 2. Fig. 2 channel shown. As described in more detail below, the digital waveform data of the channel can be passed through a variety of digital filters and integration stages before being converted into total vibration values ​​or packaged as "mass" time waveforms for further analysis by software running on the LGC card 16, or for transmission to DCS software or MHM software.

[0050] As in Fig. As shown in Figure 2, the ADC interface 70 receives via connector 34 (in Fig. (1 shown) the eight channels of continuous, simultaneously sampled data from the ADC 28 of the analog field card 12. Preferably, the data is in the format of a multiplexed synchronous serial data stream in Serial Peripheral Interface (SPI) format. The ADC interface 70 demultiplexes the data stream into data streams of eight separate channels.

[0051] Although all eight channels could be used for vibration signal processing, in a preferred embodiment two of the eight channels can be used for tachometer measurement processing. Preferably, each tachometer measurement channel contains: - a monostable flip-flop circuit 110, which is a programmable trigger "blanking" function that provides noise suppression for tachometer pulse sequences with excessive jitter or noise; - a division by N 111, which is a programmable pulse divider that divides the pulse rates of the tachometer signals generated by gears or code wheels; - a reverse rotation detector 112, which determines the direction of shaft rotation by comparing the phase of two tachometer pulse signals; - a speed indicator 115, which calculates the speed of the tachometer pulse current as a scalar total value; - a zero-speed detector 113 which provides a “zero speed” indication when the tachometer has been inactive for a programmable interval such as 0.1 s, 1 s, 10 s or 100 s; and - an overspeed detector 114 which provides an "overspeed" indication when the tachometer exceeds a fixed threshold of 2 kHz or 62 kHz. In alternative embodiments, this threshold may be programmable.

[0052] Further based on Fig. 2 Each of the eight independent parallel channels of signal processing in the FPGA 36 preferably contains the following components: - a high-pass filter 72 for DC blocking, preferably set to 0.01 Hz, 0.1 Hz, 1 Hz or 10 Hz, which can be selected or bypassed for the integrators described below based on the position of a switch 74; - two stages of integration of the digital signal shape including a first integrator 76 and a second integrator 78, which provide for the data unit a conversion of acceleration to velocity, acceleration to distance or velocity to distance; - a digital tracking bandpass filter 82 with a bandpass center frequency set by the tachometer frequency or by multiples of the tachometer frequency, and which receives as input, based on the position of a switch 80, either the "normal" data stream (no integration), the single integration data stream, or the double integration data stream, as described in more detail below; and - Blocks 88-100 for calculating the total scalar measurement value, which, as described below, determine total scalar values ​​of several different signal shapes.

[0053] In the preferred embodiment, the purpose of the digital tracking bandpass filter 82 is to provide a narrow bandpass frequency response (with high Q) with a center frequency determined by the rotational speed of a selected tachometer input. The center frequency can also be a selected integer multiple of the tachometer speed. When a waveform passes through this filter, only vibration components corresponding to multiples of the rotational speed of the monitored machine remain. When the corresponding FPGA computation block (88, 90, or 92) calculates the RMS, peak value, or peak-to-peak scalar value of the resulting waveform, the result is the same as a value that would be returned by an "nX peak" calculation performed in the application firmware of the LGC 16.Since this scalar calculation is performed as a continuous process in the FPGA 36 rather than as a calculation carried out in firmware, it is better suited for a "shutdown parameter" compared to a corresponding value generated at a lower rate in firmware. One application of this measurement is the monitoring of aeroderivative gas turbines, which generally require a tracking filter function for monitoring.

[0054] For several of the scalar totals, the individual data type from which the values ​​are calculated can be selected from the normal data stream, the single integrated data stream, the double integrated data stream, the high-pass filtered (DC-blocked) data stream, or the tracking filter data stream, based on the positions of the switches 84a-84d. Furthermore, several of the scalar total channels have individually programmable low-pass filters 88a-88d. In the preferred embodiment, these scalar totals are generated independently of and in parallel with the time signal waveforms used for prediction or protection. The blocks for calculating the scalar total include: - an RMS block 88 which determines the RMS block of the time signal shape, wherein the RMS integration time can preferably be set to 0.01 s, 0.1 s, 1 s or 10 s; - a peak value block 90, which determines the larger of the positive or negative signal shape peak value relative to the average value of the signal shape, which is preferably measured over a period of time determined either by the tachometer duration or by a programmable time delay; - a peak-to-peak block 92 that determines the signal shape peak-to-peak value over a period of time determined either by the tachometer duration or by a programmable time delay; - a block 94 for the absolute ± peak value, which determines the value of the furthest positive signal shape migration and the value of the furthest negative signal shape migration relative to the zero point of the measuring range, which is preferably measured over a time period determined either by the tachometer duration or by a programmable time delay; - a DC block 96 which determines the DC value of the time signal shape, which has a measuring range which is preferably set to 0.01 Hz, 0.1 Hz, 1 Hz or 10 Hz; and - a PeakVue™ block 100 that determines the scalar value representing the peak value of the filtered and full-wave rectified PeakVue™ waveform as described in U.S. Patent No. 5,895,857 to Robinson et al. (here inserted by reference), preferably measured over a time period determined either by the tachometer duration or by a programmable time delay. The full-wave rectification and peak-hold functions are implemented in function block 98. The PeakVue™ waveform from block 98 is also made available as a selectable input to the predictive time waveform and guard time waveform processing described herein.

[0055] The prediction time waveform processing section 116 of the FPGA 36 provides a continuous filtered time waveform for use by any prediction monitoring functions. An independent low-pass filter / decimator 104a is provided so that the prediction time waveform can have a different bandwidth than the protection time waveform. A waveform upsampling block 106 provides data rate multiplication for analysis types such as time-synchronous averaging (TSA) and order tracking. The input to the prediction time waveform processing section 116 can be selected from the normal data stream, the single-integrated data stream, the double-integrated data stream, the high-pass filtered (DC-blocked) data stream, or the PeakVue™ data stream, based on the positions of switch 102a.

[0056] The protection time waveform section 118 of the FPGA 36 provides a continuous filtered time waveform for use by protection monitoring functions. An independent low-pass filter / decimator 104b is provided, allowing the protection time waveform to have a different bandwidth than the prediction time waveform. The input to the protection time waveform processing section 118 can be selected from the normal data stream, the single integrated data stream, the double integrated data stream, the high-pass filtered (DC-blocked) data stream, or the PeakVue™ data stream, based on the positions of switch 102b.

[0057] Preferred embodiments provide transition data acquisition, wherein continuous, parallel time signal waveforms can be acquired from each signal processing channel for transmission to an external data storage system. Preferably, transition signal waveforms have a fixed bandwidth and are acquired from the protection time signal waveform data stream.

[0058] As in Fig. As shown in Figure 1, the scalar total values ​​and the digitally filtered time signal shapes are transmitted via the LGC interface 38 to the LGC logic board 16 for further processing and transport to the DCS controller 19 via the DCS I / O backplane board 18 or via the Ethernet port 52 to external software applications running on the MHM data analysis computer 54.

[0059] Fig.Figure 3 shows an example of a control logic routine (here also referred to as a control sheet) executed by the DCS controller 19. In preferred embodiments, a control sheet is scheduled for execution at a predetermined rate, such as 1 s, 0.1 s, or 0.01 s, by the DCS software running in the controller 19. While the control sheet, which controls the vibration process, is executing, total scalar vibration values ​​are sampled from the DCS I / O bus 18, and output values ​​are generated at the execution rate of the control sheet.

[0060] Preferably, the logic functions performed by the control sheets include: - a voting logic such as a logic to determine that a warning condition exists if 2 out of 2 scalar values ​​are above the threshold or if 2 out of 3 are above the threshold. - Combining vibration data with other DCS process parameter data (such as pressure and temperature). - Trigger multiplication, which is a temporary condition determined by the current machine state or by a manual input that increases an alarm level. Trigger multiplication is commonly used during the start-up of a rotating machine such as a turbine. As the turbine accelerates, it typically passes through at least one mechanical resonant frequency. Because higher than normal vibration conditions are measured during this resonance, trigger multiplication is used to temporarily raise some or all of the alarm levels to prevent false alarms. The trigger multiplication input can be set manually with an operator input or can be set automatically based on the rotational speed or another machine state input. - Trip override, which is typically a manual input to suppress the operation of the output logic to lock trip functions, such as during machine startup. Trip override is a function that suppresses either all generated vibration alarms, any outputs that would be used as a trip control, or both. The trip override input can be set manually with an operator input or can be set automatically based on a "machine state" input.

[0061] A time delay is a delay typically programmed to ensure that triggering conditions have persisted for a specified period before a machine trip is allowed to occur. As recommended by API 670, trip time delays are usually set between 1 and 3 seconds. The purpose of this delay is to reject false alarms caused by mechanical or electrical spikes or disturbances.

Claims

[1] Machine operating state monitoring module (10) which processes machine vibration data based on vibration signals and provides the machine vibration data to a distributed control system (11), wherein the machine operating state monitoring module (10) comprises: a signal conditioning circuit arrangement with an interface for receiving multiple analog vibration signals from multiple sensors attached to a machine (20), an amplification and filter circuit arrangement for conditioning the multiple analog vibration signals and an analog-to-digital conversion circuit arrangement for converting the multiple analog vibration signals into multiple digital vibration signals; a processing circuit arrangement in electrical connection with the signal conditioning circuit arrangement, wherein the processing circuit arrangement comprises several parallel digital signal processing channels, each channel serving to process one corresponding of the several digital vibration signals to generate several scalar vibration values ​​per channel and at least one vibration time signal shape per channel, wherein the several parallel digital processing channels include: a first channel for processing a vibration time signal waveform for use in machine condition prediction processing; and a second channel for processing a vibration time signal waveform for use in machine protection processing, wherein the processing performed in the second channel is independent of the processing performed in the first channel; and a logic generator circuit arrangement in electrical connection with the processing circuit arrangement, wherein the logic generator circuit arrangement is operable to receive the multiple scalar oscillation values ​​and the at least one oscillation time signal shape and is operable to format at least the multiple scalar oscillation values ​​in accordance with an input / output data protocol that is native to the distributed control system. [2] Machine operating state monitoring module (10) according to claim 1, wherein the logic generator circuit arrangement includes a backplane interface (44) configured to electrically and mechanically connect to and disconnect from an input / output bus (18) of the distributed control system (11), and wherein the multiple scalar vibration values ​​are transmitted via the backplane interface and the input / output bus (18) of the distributed control system (11) using the input / output data protocol native to the distributed control system (11). [3] Machine function state monitoring module (10) according to claim 2, wherein the logic generator circuit arrangement includes at least one network communication interface which is operable to transmit the at least one oscillation time signal shape via a communication network to a machine function state management data analysis computer (54) for machine function state prediction processing, wherein the at least one network communication interface is independent of the back panel interface (44). [4] Machine function status monitoring module (10) according to claim 3, wherein the at least one network communication interface comprises an Ethernet interface port (52b). [5] Machine function status monitoring module (10) according to claim 2, wherein the logic generator circuit arrangement includes at least one network communication interface which is operable to transmit the at least one oscillation time signal shape via a communication network to an operator computer (60) of the distributed control system for machine protection processing, wherein the at least one network communication interface is independent of the back panel interface (44). [6] Machine function status monitoring module (10) according to claim 5, wherein the at least one network communication interface comprises an Ethernet interface port (52a). [7] Machine operating state monitoring module (10) according to claim 2, wherein the logic generator circuit arrangement comprises: a first network communication interface capable of transmitting at least one oscillation time signal waveform via a first communication network to a machine function state management data analysis computer (54) for machine function state prediction processing; and a second network communication interface capable of transmitting at least one oscillation time signal waveform via a second communication network to an operator computer (60) of the distributed control system for machine protection processing, wherein the first and second network communication interfaces are independent of the rear panel interface (44). [8] Machine operating status monitoring module (10) according to claim 7, wherein the first and second network communication interfaces comprise Ethernet interface ports (52a, 52b). [9] Machine function status monitoring module (10) according to claim 2, wherein the logic generator circuit arrangement is operable to transmit the at least one oscillation time signal shape in data blocks via the backplane interface (44) and via the input / output bus of the distributed control system (11) using a block data transmission protocol. [10] Distributed tax system (11) which includes: an input / output bus (18) over which data is transmitted in accordance with a data communication protocol that is native to the distributed control system; one or more machine operating state monitoring modules (10) in electrical connection with the input / output bus (18), each of the one or more machine operating state monitoring modules (10) comprising: a signal conditioning circuit arrangement with an interface for receiving multiple analog vibration signals from multiple sensors attached to a machine (20), an amplification and filter circuit arrangement for conditioning the multiple analog vibration signals and an analog-to-digital conversion circuit arrangement for converting the multiple analog vibration signals into multiple digital vibration signals; a processing circuit arrangement in electrical connection with the signal conditioning circuit arrangement, wherein the processing circuit arrangement comprises several parallel digital signal processing channels, each channel serving to process one corresponding of the several digital vibration signals to generate several scalar vibration values ​​per channel and at least one vibration time signal shape per channel, wherein the several parallel digital processing channels include: a first channel for processing a vibration time signal waveform for use in machine condition prediction processing; and a second channel for processing a vibration time signal waveform for use in machine protection processing, wherein the processing performed in the second channel is independent of the processing performed in the first channel; and a logic generator circuit arrangement electrically connected to the processing circuit arrangement, wherein the logic generator circuit arrangement is operable to receive the multiple scalar vibration values ​​and the at least one vibration time signal shape, and is operable to format at least the multiple scalar vibration values ​​in accordance with an input / output communication protocol native to the distributed control system (11), wherein the logic generator circuit arrangement includes a backplane interface (44) configured to connect and disconnect electrically and mechanically from the input / output bus (18), wherein the multiple scalar vibration values ​​are transmitted via the backplane interface (44) and via the input / output bus (18) in accordance with an input / output communication protocol native to the distributed control system (11); one or more input modules (21) of the distributed control system in electrical connection with the input / output bus (18), each of the one or more input modules of the distributed control system being operable to receive sensor signals from process sensors attached to a machine, to generate scalar process values ​​on the basis of the sensor signals from the process sensors and to provide the scalar process values ​​to the input / output bus (18), wherein the process sensors include sensors other than vibration sensors; a controller (19) of the distributed control system, wherein the controller comprises: an interface circuit arrangement that samples the input / output bus (18) at a predetermined rate to receive the scalar vibration values ​​and the scalar process values ​​thereof; and a logic circuit arrangement for executing control logic routines that generate control signals based on logic processing of one or more of the scalar vibration values, one or more of the scalar process values, or a combination of one or more of the scalar vibration values ​​and one or more of the scalar vibration values; and one or more output modules (23) of the distributed control system in electrical connection with the input / output bus (18), each of the one or more output modules of the distributed control system being capable of receiving the control signals from the input / output bus (18) and generating machine operating output signals on the basis of the control signals. [11] Distributed control system (11) according to claim 10, wherein the logic circuit arrangement of the controller (19) of the distributed control system executes the control logic routines and generates the control signals at the same predetermined rate at which the interface circuit arrangement samples the scalar vibration values ​​and the scalar process values ​​from the input / output bus (18). [12] Distributed control system (11) according to claim 10, wherein the one or more input modules (21) of the distributed control system generate the scalar process values, which include scalar temperature values ​​and / or scalar pressure values ​​and / or scalar flow values ​​and / or scalar rotational speed values. [13] Distributed control system (11) according to claim 10, wherein the one or more machine operating state monitoring modules (10) generate the scalar vibration values ​​which include an RMS value and / or a peak value and / or a peak-to-peak value and / or a DC value and / or an absolute ±-peak value and / or a PeakVue value. [14] Distributed control system (11) according to claim 10, wherein the logic circuit arrangement optionally performs the following: first control logic routines optimized for maximum machine protection, using the first trip threshold levels; or Second logic control routines optimized for maximum machine availability, using second trip threshold levels that are higher than the first trip threshold levels. [15] Method for processing data in a distributed control system (11), wherein the data are based on sensor signals generated by sensors (20) attached to one or more machines under the control of the distributed control system (11), the method comprising: (a) the reception of multiple analog vibration signals from multiple sensors (20) attached to the one or multiple machines; (b) the conversion of the multiple analog vibration signals into multiple digital vibration signals; (c) the simultaneous processing of the multiple digital vibration signals in multiple parallel digital signal processing channels, the processing comprising: (c1) the processing of one or more of the multiple digital vibration signals in one or more of the multiple parallel digital signal processing channels which are solely intended for machine condition prediction processing; and (c2) processing one or more of the multiple digital vibration signals in one or more of the multiple parallel digital signal processing channels which are intended solely for machine condition protection processing, wherein the one or more parallel digital signal processing channels which are intended for machine condition prediction processing are separate from and independent of the one or more parallel digital signal processing channels which are intended for machine condition protection processing; (d) the generation of machine operating output signals based on machine function status protection processing; (e) the generation of machine performance signals based on machine operating state prediction processing; (f) wherein the distributed control system (11) shuts down one or more of the machines based on the machine operating output signals in order to prevent damage; and (g) the display of machine performance data on a machine condition management computer based on the machine performance signals. [16] Method according to claim 15, wherein the processing steps (c1) and (c2) are performed in separate and independent parallel channels of a Field Programmable Gate Array (FPGA) (36). [17] Method according to claim 15, wherein the processing step (c1) cannot influence the processing step (c2). [18] Method according to claim 15, wherein the data assigned to the processing step (c1) are stored in first storage locations and wherein the data assigned to the processing step (c2) are stored in second storage locations, and wherein the second storage locations can be read during the processing step (c1) but cannot be written to during the processing step (c1). [19] Method according to claim 15, wherein the processing step (c1) in a real-time operating system that controls the task priority in the distributed control system (11) has a lower priority than the processing step (c2). [20] Method according to claim 15, wherein Step (f) further includes providing the machine output signals to the distributed control system (11) via an input / output bus (18), and Step (g) further includes providing the machine capability signals to a machine function status management data analysis computer (54) via a network communication interface that is independent of the input / output bus (18). [21] Method according to claim 20, further comprising generating control signals in the distributed control system (11) to selectively deactivate the network communication interface.

Images