Device fault prediction and diagnosis device and method

Abstract

The present invention relates to an equipment failure prediction and diagnosis apparatus and method, the equipment failure prediction and diagnosis apparatus comprising: an equipment data collection unit on which at least one separate board is mounted; the at least one independent board corresponds to the type of to-be-diagnosed equipment and is provided with a plurality of input channels for inputting data of the to-be-diagnosed equipment; the device data processing unit includes a common board that supplies power to the device data collection unit and performs real-time data determination of data of the device to be diagnosed by using a diagnosis model corresponding to the channel, the data being input from at least one of the plurality of input channels, the device data collection unit and the device data processing unit are connected to each other through a serial communication interface and integrally combined to each other.

Description

Technical Field

[0001] The present invention relates to an equipment failure prediction and diagnosis device and method thereof, which are configured to achieve device miniaturization by integrally combining an equipment data collection unit and an equipment data processing unit with each other, and ensure scalability and maintainability of line applications. Background Art

[0002] In the field of equipment, Prognostics and Health Management (PHM) is on the rise and is applied to the current status diagnosis and fault prediction of related equipment. Prognostics and health management are developing rapidly, which can not only reduce costs, but also make the system run stably and achieve cost-effective maintenance without stopping.

[0003] Furthermore, instead of centralized data management in physically separate locations (i.e., where cloud computing data centers are located), edge systems are used to support computing at the edge of the network close to devices and things, enabling each device and thing to analyze and utilize individual data.

[0004] In addition, in ordinary equipment fault prediction and diagnosis devices, the systems usually used to perform edge computing are separated. In order to perform edge computing functions, there are separate devices for collecting data from the equipment to be diagnosed and devices for processing the collected data, and the connection between the two devices uses wired communication such as Universal Serial Bus (USB) or Ethernet. However, USB connections are slow, and Ethernet methods also have the problem of data loss. Ultimately, the connection method between separate devices has the problem of low maintainability and scalability compared to when the device is configured as a single system.

[0005] Accordingly, there is a need for a method of miniaturizing a system for performing edge computing into a single device and transmitting data at high speed while preventing data loss and data noise.

[0006] The foregoing content is only intended to help understand the background of the present invention, and is not intended to mean that the present invention falls within the scope of the relevant technologies known to those skilled in the art. Summary of the invention

[0007] Technical issues

[0008] The object of the present invention is to provide an equipment fault prediction and diagnosis device and method thereof, wherein the device is miniaturized by integrating an equipment data collection unit and an equipment data processing unit with each other to reduce costs, and predictive maintenance is achieved by performing equipment status diagnosis in real time, thereby ensuring the scalability and maintainability of line applications.

[0009] Technical Solution

[0010] In the present invention, as a device for solving the above-mentioned technical problems, there is provided a device for predicting and diagnosing equipment faults, which includes an equipment data collection unit and an equipment data processing unit, the equipment data collection unit is equipped with at least one or more separate boards, at least one or more separate boards are configured to correspond to the type of the diagnostic target device and have multiple input channels for inputting data of the diagnostic target device; the equipment data processing unit is provided with a common board, the common board is configured to supply power to the equipment data collection unit and perform real-time data determination on the data of the diagnostic target device input from at least one of the multiple input channels by utilizing a diagnostic model corresponding to the relevant channel, wherein the equipment data collection unit and the equipment data processing unit can be connected to each other via a serial communication interface and integrated with each other.

[0011] For example, the device data processing unit can send the collected data to an external server, and the external server can include a fault prediction and health management (PHM) server, which is configured to determine the data by performing status diagnosis of the device based on the processed data, predicting the failure or service life of the device, and calculating the replacement time of the device.

[0012] For example, the external server may store a plurality of diagnostic models corresponding to at least one of a device status diagnostic algorithm and a service life prediction algorithm, and the device data processing unit may receive a diagnostic model corresponding to the type of device identified by the common board from the external server.

[0013] For example, the device data collection unit may convert data input through at least one of the multiple input channels into a digital signal, parse the converted data, and store the parsed data in a buffer area.

[0014] For example, the cache area may include a first cache area and a second cache area, the first cache area is used to store parsed data parsed at a first sampling rate; the second cache area is used to store parsed data downsampled at a second sampling rate lower than the first sampling rate.

[0015] For example, the device data processing unit may send data from the first buffer area to an external server, and store data from the second buffer area in a database.

[0016] For example, the data input from the diagnosis target device may include at least one of speed data, vibration data, current data, and temperature data obtained from the device.

[0017] For example, at least one or more individual panels may be arranged to be detachable.

[0018] For example, when the size of the accumulated data stored in the buffer area is larger than a preset size, the device data collection unit may insert the data into the data queue so that the data is thread-processed in the device data processing unit.

[0019] For example, the device data processing unit may determine whether the size of the data queue is greater than or equal to zero, and extract data from the data queue if the size of the data queue is greater than or equal to zero.

[0020] For example, the serial communication interface may be configured to perform SPI communication.

[0021] In the present invention, as a method for solving the above-mentioned technical problems, a device fault prediction and diagnosis method for controlling a device fault prediction and diagnosis device is provided, in which a device fault prediction and diagnosis device and a device data collection unit (which are equipped with at least one or more separate boards with multiple input channels) are connected to each other via a serial communication interface and are integrated with each other, and the method includes: inputting data of a diagnostic target device; and

[0022] Real-time data determination is performed on data of a diagnosis target device input from at least one of a plurality of input channels by using a diagnosis model corresponding to a relevant channel.

[0023] For example, the method may further include determining the data by performing status diagnosis of the equipment based on the processed data, predicting failure and service life of the equipment, and calculating replacement time of the equipment.

[0024] For example, the method may further include: converting data input through at least one of the multiple input channels into a digital signal, parsing the converted data, and storing the parsed data in a buffer area.

[0025] For example, the method may further include: when the size of the accumulated data stored in the buffer area is greater than a preset size, inserting the data into a data queue so that the data is thread-processed in a device data processing unit.

[0026] For example, the method may further include determining whether the size of the data queue is greater than or equal to zero, and extracting data from the data queue if the size of the data queue is greater than or equal to zero.

[0027] Beneficial effects

[0028] According to the device fault prediction and diagnosis device and method of the present invention, the device is miniaturized by a structure in which the device data collection unit and the device data processing unit are integrated with each other, thereby reducing costs, realizing predictive maintenance by performing device status diagnosis in real time, and ensuring the scalability and maintainability of line applications. In addition, compared with an external connection structure, the design cost is reduced by integrating the structure, and the device can be configured adjacent to the device through miniaturization.

[0029] The effects of the present invention are not limited to the above-mentioned effects, and other objects not mentioned will be clearly understood by those skilled in the art from the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 Schematic diagram showing the multi-way valve device of the present invention.

[0031] Figure 2 It is shown Figure 1 A side cross-sectional view of a multi-way valve device is shown.

[0032] Figure 3 It is shown Figure 1 A front cross-sectional view of a multi-way valve device is shown.

[0033] Figure 4 is with Figure 1 Related schematic diagrams. Figure 1 : is an external perspective view showing the device for predicting and diagnosing equipment failure according to an exemplary embodiment of the present invention.

[0034] Figure 2 It is a schematic diagram showing the configuration of a device data collection unit and a device data processing unit constituting the device failure prediction and diagnosis apparatus.

[0035] Figure 3 It is a table diagram showing the reliability verification results of data input through the input channel.

[0036] Figure 4 is a functional block diagram showing the entire system.

[0037] Figure 5 The present invention is a schematic diagram showing a process in which a device data collection unit analyzes data converted into digital signals and stores the analyzed data in a first buffer area and a second buffer area.

[0038] Figure 6 The present invention is a flowchart showing a method of collecting and storing data in a device for predicting and diagnosing equipment failure. DETAILED DESCRIPTION

[0039] The specific structure and function description of the embodiments of the present invention disclosed herein are only for illustrative purposes of the embodiments of the present invention. The present invention can be implemented in many different forms. Therefore, the embodiments of the present invention should not be interpreted as limiting the present invention.

[0040] Since the embodiments of the present invention can be variously modified in many different forms, specific embodiments will be shown in the drawings and described in detail in the specification or application of the present invention. However, this is not intended to limit the embodiments of the concept according to the present invention to specific disclosed forms. On the contrary, the present invention should be understood to include all various alternative forms, equivalent forms and alternative forms that may be included in the spirit and scope of the present invention.

[0041] Unless otherwise defined, all terms (including technical and scientific terms) used in this article have the same meaning as commonly understood by those skilled in the art to which the present invention belongs. It should be further understood that terms such as those defined in general dictionaries should be interpreted as having a meaning consistent with the contextual meaning in the relevant field, and should not be interpreted as an ideal or overly formal meaning unless explicitly defined in this article.

[0042] Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Throughout the accompanying drawings, the same reference numerals will refer to the same or similar parts.

[0043] Before describing the equipment failure prediction and diagnosis method according to the exemplary embodiment of the present invention, the configuration of the equipment failure prediction and diagnosis apparatus 1000 applied to the exemplary embodiment will first be described.

[0044] Figure 1 and Figure 2 An external perspective view of a device failure prediction and diagnosis apparatus 1000 according to an exemplary embodiment of the present invention and a functional diagram of a device data collection unit 400 and a device data processing unit 500 constituting the device failure prediction and diagnosis apparatus 1000 are respectively shown.

[0045] Figure 11 is an external stereogram showing a device for predicting and diagnosing equipment failures 1000. The device for predicting and diagnosing equipment failures 1000 is also referred to as a condition monitoring system (CMS) module. The device for predicting and diagnosing equipment failures 1000 may include a housing 100, a heat sink 200, a device data collection unit 400, a device data processing unit 500, and an external connection port. First, the housing 100 surrounds the outside of the device for predicting and diagnosing equipment failures 1000 and is formed of an aluminum material, thereby reducing the weight of the device for predicting and diagnosing equipment failures 1000. In addition, since the inside and outside of the device for predicting and diagnosing equipment failures 1000 can be connected to each other by a bolt fastening structure, it is easy to attach and remove the individual plates 410 and 420 described below. In addition, a plurality of heat sinks 200 are provided on the outside of the housing 100 to discharge the heat generated from the CPU 513A and the GPU 513B.

[0046] In addition, compared with a system configuration capable of running edge computing, the equipment failure prediction and diagnosis device 1000 can be manufactured in an integrated structure in which one service is configured in a single large structure. The equipment failure prediction and diagnosis device 1000 is shaped so that the height is low and the width is longer than its front-to-back length, thereby lowering the center of gravity, and thus stability can be ensured.

[0047] In addition, the device 1000 for predicting and diagnosing equipment faults may be provided with an external connection port 300 for connecting to the outside. The types of the external connection port 300 include a Bayonet Neill-Concelman (BNC) connector 411 with 8 channels (8ch) (data can be collected by connecting a sensor cable thereto) and a voltage of common collector (VCC) / ground (GND) port, and these external connection ports 300 may be provided on the front surface of the device 1000 for predicting and diagnosing equipment faults. In addition, a power connector, a high-definition multimedia interface (HDMI) connection port, a universal serial bus (USB) connection port, and a local area network (LAN) connection port may be provided on the rear surface of the device 1000 for predicting and diagnosing equipment faults. In addition, an LED lamp capable of indicating the status of each input channel and an LED lamp capable of indicating the power supply status may be provided on the front surface of the device 1000 for predicting and diagnosing equipment faults. Preferably, in terms of data input / output, as Figure 1As shown, the external connection port 300 described so far is provided on the side of the device data collection unit 400 .

[0048] Reference Figure 2 , it can be confirmed that the functional block diagram shows the device data collection unit 400 and the device data processing unit 500 constituting the device fault prediction and diagnosis apparatus 1000. The device data collection unit 400 may be equipped with at least one or more separate boards 410 and 420 corresponding to the type of the diagnosis target device and having a plurality of input channels. In this case, each of the separate boards 410 and 420 is configured to be detachable, so that various types of data can be collected according to customization of the separate boards 410 and 420.

[0049] In addition, a module package that directly collects and controls data can be designed as each of the separate boards 410 and 420, and the circuit is configured with multiple channels, for example, 8 channels (8ch), so that data of the diagnostic target device can be input. Data can be input from the diagnostic target device through at least one of the multiple channels provided in the BNC connector 411 (which is provided in the separate board 410), and the obtained data can be converted from an analog signal to a digital signal through an analog to digital converter (ADC) 412.

[0050] In addition, the data converted into a digital signal is sent to the attenuator 413 and attenuated according to the device management voltage, and the attenuated data can be collected by the microcontroller unit (MCU) 414. The collected data can be parsed and stored in the buffer area of ​​the controller (MCU, 414). Here, data parsing refers to the task of assembling data and refining these data into required data.

[0051] On the other hand, the data of the diagnostic target device input to the plurality of input channels may include at least one of speed data, vibration data, current data, and temperature data obtained from a sensor provided for each device. The types of the above data are exemplary and are not necessarily limited thereto, and any data that can be collected from each device is not limited to the above types.

[0052] The following describes a system for ensuring the reliability of data of the diagnostic target device input through each input channel. The system for ensuring reliability can be configured in the form of a single-ended GND by connecting VCC and GND to both the BNC connector 411 and the terminal block of the device failure prediction diagnosis apparatus 1000 through a power supply. When a voltage is applied by the power supply in a manner of increasing the voltage from 1V to 24V in units of 1V, the value output from the ADC 412 can be stored, and the average value of the received data 10 times can be calculated for each of 8ch (i.e., 8 channels).

[0053] The results obtained by verifying the reliability of this configuration are as follows Figure 3 shown.

[0054] During the verification process, the bit resolution of the equipment fault prediction and diagnosis device was set to 24 bits, so that the value 2 23 It is represented as the value of ADC 412 when 24V is applied.

[0055] Specifically, the reference voltage is 0.475V, and in order to check whether the high-speed sampling rate is reproduced, the sampling rate is divided into 8kS / s and 16kS / s, and the data collected within 10 seconds are checked respectively. First, in the case where only the equipment failure prediction and diagnosis apparatus 1000 exists, the standard deviation / mean of the output value of ADC 412 (which is expressed in floating form) is checked, and in the case where the equipment failure prediction and diagnosis apparatus 1000 is installed with a vibration sensor, the standard deviation / mean of the output value of ADC 412 is checked, thereby performing a comparison on the results.

[0056] like Figure 3 As shown, it can be confirmed that the noise generated by the equipment fault prediction and diagnosis apparatus 1000 itself can be ignored by the fact that the floating value exists almost constantly, and it can also be confirmed that there is no significant difference in the change between the sampling rates. In the form of the equipment fault prediction and diagnosis apparatus 1000 combined with the vibration sensor, it can be confirmed that the ratio of the standard deviation to the mean is 0.19% at 8kS / s, and the ratio of the standard deviation to the mean is 0.14% at 16kS / s, thereby showing almost no deviation. As a result, it can be confirmed that the reliability of the data collected by connection with the vibration sensor is ensured.

[0057] One of the individual boards 410 and 420 may be provided with an Ethernet module 420. The controller (MCU, 414) of the Ethernet module may also be supplied with power in a manner utilizing POE 511B, and may exchange data with Gigabit Ethernet 512H. In addition, the controller (MCU, 414) may be connected to a plurality of Ethernet connectors via an Ethernet hub chipset. Each Ethernet connector may be connected to an external device via an Ethernet cable.

[0058] Return to reference Figure 2 The device data processing unit 500 may be provided with a common board 510 for performing real-time data determination on data of a diagnosis target device input from at least one of a plurality of input channels by using a diagnosis model corresponding to a relevant channel. Here, the common board 510 may be configured with a power section 511, an interface section 512, and a processing device section 513.

[0059] Specifically, by connecting the RJ-45 standard connector port 511A to the power unit 511, the power generated by the Ethernet power (POE) 511B can be sent to the separate boards 410 and 420 of the device data collection unit 400. In addition, the power unit 511 can supply power to the device data collection unit 400 in the form of 1.1V, 1.5V, 1.8V and 3.3V, and can also supply power to the processing device unit 513. In this case, the form of supplying power is not limited to the above form. In addition, the interface unit 512 can be configured to be connected to the processing device unit 513 through the LPDDR4 512A and DDR 32-bit interface. In addition, the interface unit 512 can be configured with an integrated package of a flash memory and a controller, and is connected to the processing device unit 513 through an embedded Multi-Media Controller (embedded Multi-Media Controller, eMMC) 512B and an MMC interface.

[0060] A status LED 512C for providing data collection status alerts for the power or channel section may be connected via a general-purpose input / output (GPIO), and a debugger / downloader 512D for firmware may be connected to the processing device section 513 via USB and a universal asynchronous receiver / transmitter (UART).

[0061] In addition, an audio amplifier (AMP) 512E for issuing a bad alarm after performing data determination can be connected to the processing device part 513 in an integrated circuit built-in audio (I2S) manner. The connection between the controller (MCU, 414) and the CPU 513A in the device data collection unit 400 can be completed through SPI communication 512F to be described below, and can be interfaced in a reduced gigabit media independent interface (RGMII) manner. In addition, the wireless mode supports Bluetooth and WiFi 512G, and Gigabit Ethernet 512H can be configured for the Ethernet module.

[0062] Furthermore, the CPU 513A and the GPU 513B constitute the processing device section 513. The CPU 513A and the GPU 513B may control data collection and preprocessing in real time, and may be loaded with an artificial intelligence (AI) driven algorithm, thereby enabling edge computing to be performed through real-time data determination.

[0063] In addition, the device data processing unit 500 can be connected to the device data collection unit 400 via a serial communication interface. In order to send and process data at high speed without data loss or noise, it is preferred to utilize a serial communication interface as a short-distance communication. For example, serial communication includes a serial peripheral interface (SPI) communication 512F. The device data processing unit 500 can also consider methods such as Ethernet to connect to the device data collection unit 400, but the Ethernet method is not preferred due to the risk of data loss during data transmission. In contrast, the SPI communication 512F has the following advantages: it can send data in real time at high speed over a short distance without data loss. Accordingly, by utilizing the SPI communication 512F, high-speed data processing can be applied, and data determination can be performed in real time in the device data processing unit 500 by an AI-driven algorithm. Accordingly, the SPI communication 512F method can be said to be a communication method suitable for the device fault prediction and diagnosis device 1000, in which the device data collection unit 400 and the device data processing unit 500 are integrated with each other and configured as a single device.

[0064] Figure 4 is a functional block diagram 600 showing the entire system.

[0065] Reference Figure 4, the functional block diagram 600 of the CPU 513A may include a Linux-based operating system 610, a communication protocol 620, and an Internet of Things (IoT) framework 630 as embedded software (SW). The communication protocol 620 includes: HyperText Transfer Protocol (HTTP) 621, Transmission Control Protocol / Internet Protocol (TCP / IP) 622, and Message Queueing Telemetry Transport (MQTT) 623, HTTP 621 is a protocol for hypertext transfer, TCP / IP 622 is an Internet information transfer protocol, and MQTT 623 is a message queue protocol, but these are only examples and are not necessarily limited to this.

[0066] In addition, the IoT framework 630 may be composed of a result monitoring part 631 , a data collecting / transmitting part 632 , a device searching part 633 , a resource managing part 634 , a data storing part 635 , and a managing part 636 .

[0067] First, the device search unit 633 can search for external devices connected to the device fault prediction and diagnosis device 1000. The resource management unit 634 can manage the resources stored in the CPU 513A, and the data storage unit 635 can store the data collected from the device. In addition, the management unit 636 provides information about the loaded diagnostic model. In addition, the resource management unit 634 can perform the following functions: provide a location for storing some data in the device, and manage resources by checking the occupancy of the CPU 513A and GPU513B in the device. In the data collection / transmission unit 632, the data collected from the device through the BNC connector 411 is received through each channel, and can be collected and sent in real time through the internal buffer area of ​​the CPU 513A via the internal buffer area of ​​the controller (MCU, 414) of the device data collection unit 400. The diagnostic model loading and determination unit 650 can perform the following functions: load the diagnostic model corresponding to the device type identified in the common board 510 (the diagnostic model is provided from the external server 2000) onto the virtual machine of the AI ​​framework 640 to provide different diagnostic models through the channel. This enables real-time data determination of the transmitted data fragments, and enables the edge computing function to be performed during the visualization of the results monitored in the result monitoring unit 183D together with the data.

[0068] In addition, the external server 2000 may include a fault prediction and health management (Prognostics and Health Management, PHM) server, which can perform status diagnosis of the equipment based on the processed data, predict the failure or service life of the equipment, and calculate the replacement time of the equipment. The external server 2000 can perform predictive maintenance of the equipment based on the data sent from the equipment data processing unit 500. In this case, a plurality of diagnostic models corresponding to at least one of the equipment status diagnosis algorithm and the service life prediction algorithm can be stored in the external server 2000 and provided to the equipment data processing unit 500. Accordingly, the equipment data processing unit 500 can receive a diagnostic model corresponding to the equipment type identified in the common board 510 among the stored diagnostic models, and can perform data determination in real time.

[0069] Figure 5 2 is a schematic diagram showing a process in which the device data collection unit 400 parses the data converted into a digital signal and stores the parsed data in the first buffer area and the second buffer area.

[0070] refer to Figure 5 From the left figure in the figure, the following process can be confirmed: the data of the device data collection unit 400 is divided and stored in the first buffer area and the second buffer area. The device fault prediction and diagnosis device 1000 capable of driving a high-speed sampling rate can receive and parse up to 16,000 samples per second from the BNC connector 411. In order to receive all parsed data without omission, it is necessary to first store the data by placing an intermediate buffer area, and control the data transmission by means of a message queue. In this case, the data can be divided and stored in two buffer areas, thereby dividing and sending to the data transmission thread and the data monitoring thread. In order to send to the data transmission thread, the parsed data parsed at a first sampling rate (i.e., a high-speed sampling rate) can be stored in the first buffer area. In addition, in order to send to the data monitoring thread, the data can be downsampled to a second sampling rate lower than the first sampling rate and stored in the second buffer area, so that there is no delay in the process of real-time data determination.

[0071] In this case, in order to utilize the shared resources (which are called threads) of CPU 513A, a method of accessing data by synchronizing with time control can be utilized. Mutual exclusion (mutex) is used as a synchronization mechanism provided by Pthread, and the amount of data to be inserted into each data queue can be controlled by stopping or sending data through the mutex_lock instruction. Each data queue is implemented as temporary storage through a ring buffer, and data transmission can be performed in a first-in first-out (FIFO) manner.

[0072] refer to Figure 5 In the left figure, the following process can be confirmed: the device data processing unit 500 sends the data of the first buffer area to the external server 2000, and sends the data of the second buffer area to the database (DB) for storage in the database. More specifically, the various data fragments temporarily stored in the first buffer area and the second buffer area are stored in each data queue in a first-in-first-out manner by utilizing the queue method so as to be sent to each thread, and the mutex_lock instruction can be used to control access to shared resources (which are called threads), thereby controlling data blocking and data transmission. Each thread can also autonomously control data transmission and extract required resources by utilizing the mutex_lock instruction, thereby preventing data loss. Accordingly, the data extracted from the thread can be sent to the external server 2000 or the database (DB) 3000 via the TCP / IP protocol, so that it can be managed.

[0073] Based on the configuration of the above-mentioned equipment failure prediction and diagnosis device 1000, refer to Figure 6 A device fault prediction diagnosis method S700 according to an exemplary embodiment is described.

[0074] Figure 6 is a flowchart showing a method S700 for collecting and storing data in a device failure prediction and diagnosis apparatus 1000, in which a device data processing unit 500 and a device data collection unit 400 are connected to each other via a serial communication interface, and the device data collection unit 400 is equipped with at least one or more separate boards 410 and 420 (which have Figure 1 multiple input channels).

[0075] First, in step S701, the data collected from the BNC connector 411 is converted into a digital signal by the ADC 412, temporarily stored in the controller (MCU, 414), and then sent to the CPU 513A of the platform module through the internal SPI communication 512F. Subsequently, in step S702, a large amount of data parsed by the CPU 513A at a high speed (i.e., 16kS / s) is parsed, and then stored in a first buffer area (which is used to store parsed data parsed at a first sampling rate) and a second buffer area (which is used to store parsed data downsampled at a second sampling rate lower than the first sampling rate) and accumulated. The data can be divided and stored in two buffer areas so as to be divided and sent to the data transmission thread and the data monitoring thread.

[0076] In step S703, the device data collection unit 400 may determine whether the size of the accumulated data stored in the first buffer area is greater than the preset size N to check whether the data is at a level that can be stored in the buffer area. This is to check the amount of data accumulated in the buffer area and determine whether the amount of data is at a level that can be stored in the buffer area. When it is determined that the size of the accumulated data stored in the first buffer area is not greater than the preset size N (i.e., no in step S703), additional data can be received. However, when it is determined that the size of the accumulated data stored in the first buffer area is greater than the preset size N (i.e., yes in step S703), in step S704, the data may be inserted into the data queue so that the data may be threaded in the device data processing unit 500. In step S705, when all data is accumulated in the data queue, the first buffer area is cleared to receive subsequent data. In steps S701 to S705, the process described so far may be performed by the device data collection unit 400.

[0077] The following process is described: after the buffer area is cleared, the device data processing unit 500 sends data to the external server 2000 and the database (DB) 3000.

[0078] In step S710, the device data processing unit 500 can check the current size of the data queue and prepare to collect data in CPU513A. In step S711, it can be determined whether to extract data from the data queue by determining whether the size of the data queue is greater than or equal to zero. In the case where the size of the data queue is less than zero (i.e., no in step S711), the size of the data queue continues to be checked. However, in the case where the size of the data queue is greater than or equal to zero (i.e., yes in step S711), in step S712, data can be extracted from the data queue. In this case, when there is data in the data queue, the mutex_lock instruction can be used simultaneously to perform time-synchronized access control according to the data queue and extract data.

[0079] Then, in step S713, the extracted data is sent to the server via the TCP / IP protocol. In step S714, the device data processing unit 500 receives the data reception result from the external server 2000 and prepares to extract new data from the data queue again. According to the data processing process as described above, the device data processing unit 500 can smoothly perform real-time data determination on the data of the diagnosis target device by using the diagnosis model corresponding to the relevant channel.

[0080] Ultimately, the device can be miniaturized by integrating the device data acquisition unit and the device data processing unit into a structure that can reduce costs, and real-time device status diagnosis can be performed through AI algorithms to achieve predictive maintenance, and the scalability and maintainability of line applications can be ensured.

[0081] In addition, the present invention described above can be implemented as a computer-readable code in a medium having a program recorded therein. Computer-readable media include all types of recording devices storing data that can be read by a computer system. Examples of computer-readable media include hard disk drives (HDD), solid-state disks (SSD), silicon disk drives (SDD), ROM, RAM, CD-ROM, magnetic tapes, floppy disks, optical data storage devices, etc. Accordingly, the above detailed description should not be interpreted as being restrictive in all aspects, but exemplary. The scope of the present invention should be determined by a reasonable interpretation of the appended claims, and all modifications within the equivalent scope of the present invention are included within the scope of the present invention.

[0082] <Description of Reference Symbols in the Drawings>

[0083] 100: Shell 200: Heat sink

[0084] 300: External connection port 400: Device data collection unit

[0085] 500: Equipment data processing unit 410, 420: Separate board

[0086] 411: BNC connector 412: ADC

[0087] 413: Attenuator 414: Controller (MCU)

[0088] 510: Public Board 511: Ministry of Electricity

[0089] 511A: Connector port 511B: POE

[0090] 512: Interface unit 512A: LPDDR4

[0091] 512B: eMMC 512C: Status LED

[0092] 512D: Debugger / Downloader 512E: Audio AMP

[0093] 512F: SPI communication 512G: Bluetooth, WiFi

[0094] 512H: Gigabit Ethernet 513: Processing device unit

[0095] 513A: CPU 513B: GPU

[0096] 610: Operating system 620: Communication protocol

[0097] 621:HTTP 622:TCP / IP

[0098] 623:MQTT 630:IOT Framework

[0099] 631: Result Monitoring Department 632: Data Collection / Transmission Department

[0100] 633: Device search unit 634: Resource management unit

[0101] 635: Data storage unit 636: Management unit

[0102] 640: AI framework 650: Diagnostic model loading and determination

[0103] 1000: Equipment failure prediction and diagnosis device 2000: External server

[0104] 3000: Database (DB).

Claims

1. A device for predicting and diagnosing equipment failures, comprising: a device data collection unit equipped with at least one or more separate boards, the at least one or more separate boards being configured to correspond to the type of the diagnosis target device and having a plurality of input channels for inputting data of the diagnosis target device; and an equipment data processing unit provided with a common board configured to supply power to the equipment data collecting unit and to perform real-time data determination on data of a diagnosis target equipment input from at least one of the plurality of input channels by utilizing a diagnosis model corresponding to a relevant channel; The device data acquisition unit and the device data processing unit are connected to each other via a serial communication interface and are integrated with each other.

2. The device according to claim 1, wherein: The device data processing unit sends the collected data to an external server, and The external server includes a fault prediction and health management (PHM) server, which is configured to determine data by performing status diagnosis of the equipment based on processed data, predicting the failure or service life of the equipment, and calculating the replacement time of the equipment.

3. The device according to claim 2, wherein: The external server stores a plurality of diagnosis models corresponding to at least one of an equipment status diagnosis algorithm and a service life prediction algorithm, and The device data processing unit receives a diagnosis model corresponding to a type of device recognized by the common board from an external server.

4. The device according to claim 2, wherein: The device data collection unit converts data input through at least one of the multiple input channels into a digital signal, analyzes the converted data, and stores the analyzed data in a buffer area.

5. The device according to claim 4, wherein: The cache area includes: A first buffer area for storing parsed data parsed at a first sampling rate; and The second buffer area is used to store the parsed data downsampled at a second sampling rate lower than the first sampling rate.

6. The device according to claim 5, wherein: The device data processing unit sends the data from the first buffer area to an external server, and stores the data from the second buffer area in a database.

7. The device according to claim 1, wherein: The data input from the diagnosis target device includes at least one of speed data, vibration data, current data, and temperature data obtained from the device.

8. The device according to claim 1, wherein: The at least one or more separate panels are configured to be detachable.

9. The device according to claim 4, wherein: When the size of the data stored and accumulated in the buffer area is larger than a preset size, the device data collection unit inserts the data into the data queue so that the data is thread-processed in the device data processing unit.

10. The device according to claim 9, wherein: The device data processing unit determines whether the size of the data queue is greater than or equal to zero, and extracts data from the data queue if the size of the data queue is greater than or equal to zero.

11. The device according to claim 1, wherein: The serial communication interface is configured to perform SPI communication.

12. A method for predicting and diagnosing equipment failures for controlling an equipment fault predicting and diagnosing device, in which an equipment data processing unit and an equipment data collecting unit are connected to each other via a serial communication interface and are integrally combined with each other, and the equipment data collecting unit is equipped with at least one or more separate boards having a plurality of input channels, the method comprising: Enter the data of the diagnostic target device; as well as Real-time data determination is performed on data of a diagnosis target device input from at least one of a plurality of input channels by using a diagnosis model corresponding to a relevant channel.

13. The method according to claim 12, further comprising: The data is determined by performing equipment status diagnosis based on the processed data, predicting equipment failure and service life, and calculating equipment replacement time.

14. The method according to claim 12, further comprising: Data inputted through at least one of the plurality of input channels is converted into a digital signal, the converted data is parsed, and the parsed data is stored in a buffer area.

15. The method according to claim 14, further comprising: When the size of the data stored and accumulated in the buffer area is larger than a preset size, the data is inserted into the data queue so that the data is thread-processed in the device data processing unit.

16. The method according to claim 15, further comprising: It is determined whether the size of the data queue is greater than or equal to zero, and if the size of the data queue is greater than or equal to zero, data is extracted from the data queue.

Images