A subway traction motor fault diagnosis system and method

Abstract

The present invention relates to the technical field of subway traction motor fault diagnosis, and specifically to a subway traction motor fault diagnosis system and method. In order to solve the problem of the above-mentioned defects in the existing subway traction motor fault diagnosis system, a new subway traction motor fault diagnosis system and method are provided. The present invention is equipped with two wide-range power supply interfaces and multiple types of communication interfaces and is expandable to meet the access requirements of different communication interfaces. At the same time, two CPUs are set, and four acquisition boards are equipped as needed to meet the requirements of different computing power and sampling channels; at the same time, the system algorithm uses vibration and temperature, two signals that can directly characterize the mechanical fault of the motor bearing, as input sources, which improves the accuracy of motor fault diagnosis. At the same time, a modular design concept is adopted, and each module can be used independently, saving the development cycle and improving the development efficiency; in addition, the software part adopts a hierarchical rationalized architecture design to avoid unreasonable functional planning of the architecture design.

Description

Technical Field

[0001] The present invention relates to the technical field of subway traction motor fault diagnosis, and in particular to a subway traction motor fault diagnosis system and method. Background Art

[0002] In recent years, urban rail transit has played a positive role in alleviating urban traffic congestion, helping to optimize urban spatial structure, and improving urban environmental protection. As the power source of urban rail transit trains, the operating status of subway traction motors directly affects the train's power performance and driving safety. Unlike general industrial applications, the service environment of urban rail transit vehicles is relatively complex. The load variation range of subway traction motors is large, the operating temperature variation range is wide, and they are often subjected to non-deterministic impacts on the track surface. At the same time, they frequently work in cyclic conditions such as starting, traction, braking, and parking. Under the long-term action of composite alternating loads such as electricity, magnetism, heat, and force, their performance is degraded and fatigued. Therefore, the monitoring and diagnosis of the status of subway traction motors is of great and positive significance for ensuring the operation safety and driving order of urban rail transit trains.

[0003] The subway traction motor fault diagnosis system uses sensors installed on the subway traction motor to collect traction motor operation data in real time. It realizes real-time monitoring of the subway traction motor operation status through real-time analysis of the collected data, and guides subway train operation and maintenance based on the diagnosis results.

[0004] The first prior art is a fault diagnosis system based on electrical signals, which utilizes the voltage sensor and current sensor equipped in the traction system, and combines the collected voltage data and current data with information such as motor speed and torque through a gradient boosting tree model and a preset fuzzy clustering model for analysis to achieve motor fault diagnosis. This is a machine learning method, but subway traction motor failures are mainly bearing mechanical failures. The electrical signal characterizing a mechanical failure in the traction motor is very weak compared to the vibration signal. The fault judgment may result in a high misjudgment rate due to insufficient signal sampling accuracy, and there is a lack of an effective acquisition and diagnosis system.

[0005] The second prior art is a fault diagnosis system based on dimension reduction data, which is to extract the low-frequency components in the data after obtaining the multi-dimensional running time sequence data of the traction motor to be identified, and divide these time sequence data into multiple different data segments according to the correlation between the data to form multiple time series segments, and then perform dimension reduction processing on each time series segment to obtain the time series segment after dimension reduction, and finally, perform optimal solution on these dimension reduction data, identify the fault state of each sequence segment, and output the identification result of the fault state, which is a statistical method. This method may cause the loss of important high-frequency information when extracting low-frequency components; secondly, the standard of data segment division may affect the reliability and consistency of the results; thirdly, the dimension reduction process may cause information loss, affecting the accuracy of subsequent fault identification; finally, the complexity of solving the optimal solution may lead to a long calculation time, making it difficult to apply in real time, and lacking a complete system architecture system; at the same time, the subway traction motor fault diagnosis system cannot only consider data collection and calculation, but also consider the independence, compatibility and scalability of the system, so as to facilitate the interaction with other equipment data of subway trains. Summary of the invention

[0006] In order to solve the above-mentioned defects of the existing subway traction motor fault diagnosis system, the present invention provides a new subway traction motor fault diagnosis system and method.

[0007] The present invention is achieved by adopting the following technical solutions:

[0008] A subway traction motor fault diagnosis system, comprising:

[0009] The hardware part includes power board, acquisition board, communication board, CPU board and backplane;

[0010] The input of the power board adopts a wide range of DC24V and DC110V power supply that meets the railway industry standards, providing stable and reliable power for each hardware interface of the system to ensure that the power quality and performance of the system are not affected during the process of adding or removing boards, thereby maintaining the normal operation of the system and data integrity;

[0011] The acquisition board is developed based on FPGA to improve the speed and accuracy of data acquisition. It is equipped with signal conditioning circuits and can adapt to various types of sensors. It uses F48 connector interface to monitor the temperature and vibration parameters of the motor in real time by connecting sensors to obtain dynamic operation data, thus providing a basis for subsequent fault diagnosis and performance evaluation. The acquisition board can be increased or decreased according to actual on-site needs;

[0012] The communication board fully considers the rail transit network communication mode and is equipped with MVB and Ethernet communication interfaces to meet the actual networking needs;

[0013] The CPU board is developed based on a high-performance ARM chip and is responsible for real-time analysis and processing of collected real-time data, applying corresponding algorithms and models to extract meaningful information and generate results that can be used for fault diagnosis and decision support;

[0014] The backplane adopts a bus structure and provides high-bandwidth and low-latency interfaces;

[0015] The software part is designed based on the Linux operating system, and mainly includes the system peripheral interface layer, the operating system interface layer, the function module layer and the business function logic layer;

[0016] The system peripheral interface layer is responsible for receiving and preprocessing sensor data in real time;

[0017] The operating system interface layer is responsible for realizing data communication and interaction between the system peripheral interface layer and the business function logic layer;

[0018] The functional module layer is designed in C language and includes system management business unit, data collection business unit, data management business unit, data calculation business unit, and data interaction business unit;

[0019] The system management business unit includes three functional modules: system information configuration, system time management and log management, which respectively manage system parameter information, time information and record information;

[0020] The data acquisition business unit includes an AD data acquisition module, which can collect data according to a preset sampling frequency, and the sampling frequency range is adjustable. The sampling frequency of vibration data should be no less than 10kHz, and the sampling frequency of temperature data should be no less than 1Hz, so as to realize the real-time collection and analog-to-digital conversion of vibration and temperature data. The acquisition module is equipped with a conditioning information configuration interface, which can realize the collection of voltage and current sensor data through configuration. The AD data acquisition module classifies the received acquisition board data to provide a reliable data source for subsequent functional modules;

[0021] The data management business unit includes three data management function modules: data cache management module, data processing module and data storage module. The data cache management module not only caches the received collected data, but also caches the inter-board communication data, vehicle communication data and remote transmission data, and provides independent cache space for external algorithms to ensure that data is not lost; the data processing module mainly performs related operations such as filtering and outlier removal on the collected data and other received data to ensure the validity of the data; the data storage module realizes the storage of collected data, communication data, calculation result data and operation abnormality information to ensure that the information is traceable;

[0022] The data calculation business unit includes an algorithm library module and a fault diagnosis module; the algorithm library module performs feature algorithm calculation and nonlinear principal component algorithm calculation based on multi-feature data on the data processed by the data processing module and gives the calculation results; the fault diagnosis module uses the calculation results of the algorithm library module and the a priori theory to compare the inherent characteristics of the motor bearing and the results between the boards, corrects the calculation results, and gives the fault diagnosis results;

[0023] The data calculation business unit includes an algorithm library module and a fault diagnosis module; the algorithm library module of this method covers 16 commonly used feature calculation methods, but in actual applications, the data processed by the data processing module will be subjected to feature calculations with certain characterization significance, including root mean square value, peak value, kurtosis, skewness, etc. In order to effectively improve the accuracy of motor fault identification, this method will trigger the nonlinear principal component algorithm calculation based on the characteristic algorithm threshold condition. The algorithm calculates key characterization indicators by assigning specific weights to the motor vibration root mean square value, skewness and temperature change rate, where the temperature change rate takes into account specific time constraints; when the key characterization indicator exceeds the limit, the fault diagnosis module will perform a Hilbert transform on the vibration data and calculate the envelope, further analyze the spectrum, identify the bearing characteristic frequency in the envelope signal, and use a priori theory combined with the inherent characteristics of the motor bearing and the comparison of results between boards to correct the calculation results and give a fault diagnosis result;

[0024] The data interaction business unit uses a variety of information interaction forms to achieve communication with other systems or devices;

[0025] The business function logic layer starts the functional processes and threads of the system management business unit, data collection business unit, data management business unit, data calculation business unit and data interaction business unit through logical association to complete the functional logic of each specific business unit.

[0026] Furthermore, there are two CPU boards in the hardware part, namely CPU board 1 and CPU board 2, and four acquisition boards, namely acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. CPU board 1 and CPU board 2 are of the same board design and can both realize data interaction with acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. The configuration of CPU board 1 and CPU board 2 depends on the number of acquisition boards and the sampling frequency of the data. When the number of acquisition boards does not exceed two or the sampling frequency of the data is low or medium, one acquisition board is installed at any position of CPU board 1 or CPU board 2, and the other position is replaced by a blind plate; when the demand is high, CPU board 1 and CPU board 2 are installed at the same time, and the two realize information interaction through the backplane bus.

[0027] Furthermore, acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4 are of the same board design, and can realize data interaction with CPU board 1 and CPU board 2. The configuration of the acquisition board depends on the amount of data collected, that is, the number of sensors required to be connected. Each acquisition board is equipped with 8-way vibration data interface and 8-way temperature data interface, wherein the vibration data interface realizes vibration data collection through conditioning circuit configuration.

[0028] Furthermore, each acquisition board is designed with 48 pins. When the actual number of vibration sensors required is less than 8 and the number of temperature sensors required is less than 8, one acquisition board can be used to meet the needs. The acquisition board can be installed at any position in acquisition board 1, acquisition board 2, acquisition board 3 or acquisition board 4, and blind plates can be used instead at other positions.

[0029] Furthermore, the data interaction business unit includes three data interaction function modules, namely, inter-board communication module, vehicle communication module and remote transmission module; the inter-board communication module includes data communication between CPU board 1 and CPU board 2, including data communication such as calculation results, configuration information, and data flow direction; CPU board 1 and CPU board 2 are connected to other external devices with the communication board through MVB or Ethernet networking to transmit information interaction; CPU board 1 and CPU2 communicate with acquisition channels, acquisition signal data and other data between acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4 respectively; the vehicle communication module has three communication modes, namely, Ethernet, RS485 and CAN, to realize the communication of diagnostic results, vehicle information and information data between other equipment units and the vehicle maintenance network; the remote transmission module transmits diagnostic result data, acquisition data and other data to the service platform, or receives information from the service platform and distributes it to other business function modules;

[0030] A method for diagnosing a subway traction motor fault comprises the following steps:

[0031] (1) System loading initialization: The hardware unit is powered on to complete the loading of the embedded system, and complete the initialization of parameter configuration information, cache space, and communication connections. If the initialization fails, the information will be recorded through the log.

[0032] (2) Configuration information acquisition: after the system records initialization, the system configuration information needs to be read. This configuration information reflects the current hardware status information of the system and the operation information to be subsequently executed. When the configuration information acquisition fails, the CPU master-slave relationship division mechanism will be started. After the master-slave relationship is clarified, steps (1) and (2) will be re-executed to re-acquire the configuration information and ensure the integrity of the configuration information.

[0033] (3) CPU master-slave relationship division: The master-slave relationship is divided according to the number of CPUs configured in the configuration information. When there is only one CPU, the current CPU is the master. When there are two CPUs, CPU 1 is the master and CPU 2 is the slave. The master CPU performs the overall system scheduling function and receives diagnostic information from the slave CPUs.

[0034] (4) Time calibration, which provides a time basis for ensuring subsequent data synchronization;

[0035] (5) Obtaining collected data, judging the initial state of the sampled data according to the collected information configured by the configuration information, and triggering log recording when the collected information is abnormal;

[0036] (6) Synchronous sampling: To ensure the consistency of data analysis and the accuracy of judgment results, a synchronous sampling signal is set to trigger the synchronous acquisition of data from each channel and record the collected information through a diary;

[0037] (7) Data processing: After the acquired data is processed, the data is stored in a specific amount of data according to the parameters set in the configuration information. On the other hand, the data is placed in the cache space as the input condition of the algorithm for use in executing the algorithm.

[0038] (8) Algorithm execution: extracting calculated data from the cache space to perform feature algorithm calculations, inputting the calculated feature data into a nonlinear principal component algorithm based on multi-feature data to perform nonlinear correlation analysis of the data, and placing the feature calculation and correlation analysis results into the data cache space.

[0039] (9) Fault diagnosis: The algorithm execution results are modified by combining the priori theory with the inherent characteristics of the motor bearings to give one of the four fault diagnosis results: normal, warning, first-level alarm, and second-level alarm. The result data is stored in the cache space.

[0040] (10) Data interaction: extracting raw data, calculation result data, and diagnosis result data from the cache space and transmitting them to the vehicle system and ground service platform through vehicle communication and remote transmission for further data utilization; performing inter-board data comparison and analysis on raw data, calculation result data, diagnosis result data, configuration parameter information, and comparison result data through inter-board communication, and updating cache space data in real time to improve the accuracy of fault diagnosis.

[0041] The beneficial effects produced by the present invention are as follows:

[0042] (1) The general system is only equipped with one input power interface, one type of communication interface and a single acquisition board. The present invention is equipped with two wide-range power supply interfaces and multiple types of communication interfaces and is expandable to meet the access requirements of different communication interfaces. At the same time, two CPUs are set and four acquisition boards are equipped as needed to meet the requirements of different computing power and sampling channels, broaden the application scenarios, and can not only realize single motor fault diagnosis, but also realize multiple motor fault diagnosis by increasing the number of acquisition boards according to the number of diagnosed motors when the existing acquisition channels are insufficient. It has system independence and scalability, and saves hardware costs according to needs;

[0043] (2) The system is equipped with multiple commonly used communication interfaces such as Ethernet, MVB, CAN and RS485. At the same time, the CPU board is expandable. Each CPU board is equipped with an independent cache space that can be compatible with the communication data of other devices as needed. The diagnostic algorithms of other devices can be transplanted and deployed into this system, making full use of hardware resources and saving costs.

[0044] (3) The system algorithm uses vibration and temperature, two signals that can directly represent the mechanical fault of the motor bearing, as input sources, which improves the accuracy of motor fault diagnosis. At the same time, it adopts a modular design concept, and each module can be used independently, saving development cycle and improving development efficiency.

[0045] (4) The architecture design of the software part mainly starts from the bottom interface level to the system level and then to the application level. The hierarchical architecture design is rationalized and five major business functions are planned. Each business contains specific functional modules to avoid unreasonable functional planning of the architecture design, which leads to unclear system logic and affects the stability and execution efficiency of the system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 It is a schematic diagram of the hardware units in the subway traction motor fault diagnosis system of the present invention;

[0049] Figure 2 A schematic diagram of software units in the subway traction motor fault diagnosis system of the present invention;

[0050] Figure 3This is a schematic diagram of the power board panel;

[0051] Figure 4 This is a schematic diagram of the CPU board panel;

[0052] Figure 5 This is a schematic diagram of the communication board panel;

[0053] Figure 6 This is a schematic diagram of the acquisition board panel;

[0054] Figure 7 This is a flow chart of the subway traction motor fault diagnosis method described in the present invention. DETAILED DESCRIPTION

[0055] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the scheme of the present invention will be further described below. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0056] In the description, it should be noted that the terms "first" and "second" are only used for descriptive purposes and should not be understood as indicating or implying relative importance. It should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be internal communication between two elements. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present invention, rather than all of the embodiments.

[0058] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0059] like Figure 1 and Figure 2 As shown, a subway traction motor fault diagnosis system comprises:

[0060] The hardware part includes power board, acquisition board, communication board, CPU board and backplane;

[0061] The input of the power board adopts a wide range of DC24V and DC110V power supply that meets the railway industry standards, providing stable and reliable power for each hardware interface of the system to ensure that the power quality and performance of the system are not affected during the process of adding or removing boards, thereby maintaining the normal operation of the system and data integrity;

[0062] The acquisition board is developed based on FPGA to improve the speed and accuracy of data acquisition. It is equipped with signal conditioning circuits and can adapt to various types of sensors. It uses F48 connector interface to monitor the temperature and vibration parameters of the motor in real time by connecting sensors to obtain dynamic operation data, thus providing a basis for subsequent fault diagnosis and performance evaluation. The acquisition board can be increased or decreased according to actual on-site needs;

[0063] The communication board fully considers the rail transit network communication mode and is equipped with MVB and Ethernet communication interfaces to meet the actual networking needs;

[0064] The CPU board is developed based on a high-performance ARM chip and is responsible for real-time analysis and processing of collected real-time data, applying corresponding algorithms and models to extract meaningful information and generate results that can be used for fault diagnosis and decision support;

[0065] The backplane adopts a bus structure and provides high-bandwidth and low-latency interfaces;

[0066] The software part is designed based on the Linux operating system, and mainly includes the system peripheral interface layer, the operating system interface layer, the function module layer and the business function logic layer;

[0067] The system peripheral interface layer is responsible for receiving and preprocessing sensor data in real time;

[0068] The operating system interface layer is responsible for realizing data communication and interaction between the system peripheral interface layer and the business function logic layer;

[0069] The functional module layer is designed in C language and includes system management business unit, data collection business unit, data management business unit, data calculation business unit, and data interaction business unit;

[0070] The system management business unit includes three functional modules: system information configuration, system time management and log management, which respectively manage system parameter information, time information and record information;

[0071] The data acquisition business unit includes an AD data acquisition module, which can collect data according to a preset sampling frequency, and the sampling frequency range is adjustable. The sampling frequency of vibration data should be no less than 10kHz, and the sampling frequency of temperature data should be no less than 1Hz, so as to realize the real-time collection and analog-to-digital conversion of vibration and temperature data. The acquisition module is equipped with a conditioning information configuration interface, which can realize the collection of voltage and current sensor data through configuration. The AD data acquisition module classifies the received acquisition board data to provide a reliable data source for subsequent functional modules;

[0072] The data management business unit includes three data management function modules: data cache management module, data processing module and data storage module. The data cache management module not only caches the received collected data, but also caches the inter-board communication data, vehicle communication data and remote transmission data, and provides independent cache space for external algorithms to ensure that data is not lost; the data processing module mainly performs related operations such as filtering and outlier removal on the collected data and other received data to ensure the validity of the data; the data storage module realizes the storage of collected data, communication data, calculation result data and operation abnormality information to ensure that the information is traceable;

[0073] The data calculation business unit includes an algorithm library module and a fault diagnosis module; the algorithm library module performs feature algorithm calculation and nonlinear principal component algorithm calculation based on multi-feature data on the data processed by the data processing module and gives the calculation results; the fault diagnosis module uses the calculation results of the algorithm library module and the a priori theory to compare the inherent characteristics of the motor bearing and the results between the boards, corrects the calculation results, and gives the fault diagnosis results;

[0074] The specific algorithm of the data computing business unit is as follows:

[0075] The algorithm library module covers 16 commonly used feature calculation methods. In the specific implementation of the present invention, the algorithm library module performs feature calculations including root mean square value, peak value, kurtosis, skewness and other features with certain characterization significance on the data processed by the data processing module, as shown in formulas (1), (2), (3) and (4).

[0076]

[0077] Among them, x i is the i-th vibration sampling point, N is the total number of sampling points of the signal;

[0078] Peak P K =max(x i )(2)

[0079]

[0080] Where μ is the mean of the signal, σ is the standard deviation, and x i are the sampling points of the signal.

[0081]

[0082] In order to effectively improve the accuracy of motor fault identification, the present invention will trigger the nonlinear principal component algorithm calculation based on the characteristic algorithm threshold condition at the same time. The algorithm mainly combines the motor vibration and temperature for comprehensive diagnosis. n is the key characterization index, as shown in formula (5):

[0083] P n =W1RMS n +W2S n +W3SS n (5)

[0084]

[0085] Among them, W1, W2, W3 are weights; RMS n is the root mean square value of vibration; S n is the temperature change rate, as shown in formula (6); SS n is the skewness. n is the time when the vibration RMS value exceeds the limit for the nth time, T(t n ) is the corresponding temperature, t n+1 The effective value of vibration exceeds the limit. m min or t s The moment when the vibration RMS value exceeds the limit for the first time after s, T(t n+1 ) is the corresponding temperature, t m No more than 2min, t s Not shorter than 30s but not longer than 60s;

[0086] When P n When the indicator exceeds the limit, the fault diagnosis module will perform Hilbert transform on the vibration data, calculate the envelope, further analyze the spectrum, identify the bearing characteristic frequency in the envelope signal, and use a priori theory combined with the inherent characteristics of the motor bearing and the comparison of the results between the boards to correct the calculation results and give the fault diagnosis results;

[0087] The data interaction business unit uses a variety of information interaction forms to achieve communication with other systems or devices;

[0088] The business function logic layer starts the functional processes and threads of the system management business unit, data collection business unit, data management business unit, data calculation business unit and data interaction business unit through logical association to complete the functional logic of each specific business unit.

[0089] In specific implementation, the power board includes two modes: DC 9~36V and DC 77~137.5V, which can effectively meet the common power supply requirements of the railway industry, such as DC 12V, DC 24V and DC 110V. By integrating multiple power supply modules, the power board not only provides stable internal power support for the entire hardware unit of the subway traction motor fault diagnosis system, but also provides reliable working power for the normal operation of external sensors.

[0090] When implementing it, Figure 3As shown, the power board is designed with five power interfaces and are defined as A1 port - DC 9 ~ 36V positive pole, A2 port - DC7 ~ 137.5V positive pole, A3 port - power interface positioning hole, A4 port - common power negative pole. The power board is also designed with ten system power working status indicator lights. The meanings of the system power working status indicator lights are as follows: indicator light A reflects the external input power, that is, the entire hardware unit receives the external power supply input status, indicator light B reflects the power working status of the power board itself, indicator lights C1 and C2 respectively reflect the power and working status of CPU board 1 and CPU board 2, indicator light C3 reflects the power and working status of the communication board, indicator lights D1, D2, D3, D4 respectively reflect the power and working status of acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4, indicator light D5 is a reserved indicator light, the indicator light is always green, indicating that the power and working status are normal, the indicator light is always red, indicating that the power supply is abnormal, the indicator light is always green or always red, indicating that the board is abnormal, and the indicator light is off, indicating that there is no power input.

[0091] During specific implementation, there are two CPU boards in the hardware part, namely CPU board 1 and CPU board 2, and four acquisition boards, namely acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. CPU board 1 and CPU board 2 are designed as the same board and can both realize data interaction with acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. The configuration of CPU board 1 and CPU board 2 depends on the number of acquisition boards and the sampling frequency of the data. When the number of acquisition boards does not exceed two or the sampling frequency of the data is low or medium, one acquisition board is installed at any position of CPU board 1 or CPU board 2, and the other position is replaced by a blind plate; when the demand is high, CPU board 1 and CPU board 2 are installed at the same time, and the two realize information interaction through the backplane bus.

[0092] In specific implementation, the CPU board is designed with two Ethernet interfaces and a set of DB9 interfaces, such as Figure 4 As shown, the two Ethernet interfaces are multiplexed interface designs, taking into account the current vehicle Ethernet interface form. One Ethernet interface is a 100M Ethernet interface, and the other Ethernet interface is a Gigabit Ethernet interface, which can realize communication data interaction with subway trains, data remote transmission and program debugging, etc. The DB9 interface is a multiplexed interface for RS232, RS485 and CAN communications. RS232 mainly provides Debug debugging functions, and RS485 and CAN are used to realize RS485 or CAN communication data interaction with other peripheral devices. A, B, C, and D are Ethernet working indicators. A and B are 100M Ethernet working indicators, C and D are Gigabit Ethernet working indicators, A1 is the RS485 communication indicator, and A2 is the CAN communication indicator.

[0093] Two CPU boards can be selected as needed to realize the expandable functions of the board. The CPU board is equipped with multiple common communication interfaces such as Ethernet, RS485 and CAN to realize communication between other equipment in the subway train. At the same time, the CPU runs an embedded operating system. In addition to independently realizing the software functions of the subway traction motor fault diagnosis system, it can also provide operating resource space for other equipment software functions by opening up independent space, thereby reducing the hardware equipment of the subway train and realizing the compatibility of equipment in terms of equipment communication and system software operation.

[0094] In specific implementation, acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4 are designed as the same board, and can realize data interaction with CPU board 1 and CPU board 2. The configuration of the acquisition board depends on the amount of data collected, that is, the number of sensors required to be connected. Each acquisition board is equipped with 8-way vibration data interface and 8-way temperature data interface, among which the vibration data interface realizes vibration data collection through conditioning circuit configuration.

[0095] In specific implementation, each acquisition board is designed with 48 pins, such as Figure 6 When the actual number of vibration sensors is less than 8 and the number of temperature sensors is less than 8, one acquisition board can be used to meet the requirement. The acquisition board can be installed at any position of acquisition board 1, acquisition board 2, acquisition board 3 or acquisition board 4, and blind plates can be used in other positions.

[0096] In specific implementation, the communication board is used as a communication function expansion board to achieve data interaction with CPU board 1 and CPU board 2. When the CPU communication interface does not meet the needs, such as when it is necessary to access the MVB network or when it is necessary to access more external communication devices, the communication board can be inserted to achieve communication interaction with external devices. When there is no need to expand communication, a blind plate can be used instead. The communication board is designed with 1 Ethernet port and two sets of DB9 interfaces, such as Figure 5 As shown. The Ethernet interface is a 100M multiplexed interface design, taking into account the current vehicle Ethernet interface form, and can realize communication data interaction with subway trains, data remote transmission and program debugging and other functions. The two groups of DB9 are MVB communication interfaces, which are redundantly designed. One is a male connector and the other is a female connector. Each group contains two redundant channels, A and B. A, B, C, and D are working indicators. A and B are Ethernet working indicators, C is the MVB male connector communication indicator, and D is the MVB female connector communication indicator.

[0097] In specific implementation, the backplane adopts a unified interface design, and each functional board can be installed in any slot of the chassis corresponding to the functional board, thereby avoiding board burning caused by mismatch between the functional board and the slot.

[0098] During specific implementation, the data interaction business unit includes three data interaction function modules: inter-board communication module, vehicle communication module and remote transmission module; the inter-board communication module includes data communication between CPU board 1 and CPU board 2, including data communication such as calculation results, configuration information, and data flow direction; CPU board 1 and CPU board 2 are connected to other external devices with the communication board through MVB or Ethernet networking to transmit information interaction; CPU board 1 and CPU2 communicate with acquisition channels, acquisition signal data and other data between acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4 respectively; the vehicle communication module has three communication modes of Ethernet, RS485 and CAN, and realizes the communication of diagnostic results, vehicle information and information data with the vehicle maintenance network and other equipment units; the remote transmission module transmits data such as diagnostic results and acquisition data to the service platform, or receives service platform information and distributes it to other business function modules. The service platform information contains relevant information such as system parameter configuration information and time information;

[0099] In specific implementation, the diagnosis results in the fault diagnosis module include four levels: normal, warning, first-level alarm, and second-level alarm.

[0100] A method for diagnosing subway traction motor faults, such as Figure 7 As shown, the following steps are included:

[0101] (1) System loading initialization: the hardware unit is powered on to complete the loading of the embedded system, and complete the initialization of parameter configuration information, cache space, and communication connection. If the initialization fails, the information will be recorded through the log;

[0102] (2) Configuration information acquisition: after the system records initialization, the system configuration information needs to be read. This configuration information reflects the current hardware status information of the system and the operation information to be subsequently executed. When the configuration information acquisition fails, the CPU master-slave relationship division mechanism will be started. After the master-slave relationship is clarified, steps (1) and (2) will be re-executed to re-acquire the configuration information and ensure the integrity of the configuration information.

[0103] (3) CPU master-slave relationship division: The master-slave relationship is divided according to the number of CPUs configured in the configuration information. When there is only one CPU, the current CPU is the master. When there are two CPUs, CPU 1 is the master and CPU 2 is the slave. The master CPU performs the overall system scheduling function and receives diagnostic information from the slave CPUs.

[0104] (4) Time calibration, which provides a time basis for ensuring subsequent data synchronization;

[0105] (5) Obtaining collected data, judging the initial state of the sampled data according to the collected information configured by the configuration information, and triggering log recording when the collected information is abnormal;

[0106] (6) Synchronous sampling: To ensure the consistency of data analysis and the accuracy of judgment results, a synchronous sampling signal is set to trigger the synchronous acquisition of data from each channel and record the collected information through a diary;

[0107] (7) Data processing: After the acquired data is processed, the data is stored in a specific amount of data according to the parameters set in the configuration information, and the data is placed in the cache space as the input condition of the algorithm for the algorithm to execute;

[0108] (8) Algorithm execution: extracting calculated data from the cache space to perform feature algorithm calculation, inputting the calculated feature data into a nonlinear principal component algorithm based on multi-feature data to perform nonlinear correlation analysis calculation of the data, and placing the feature calculation and correlation analysis results into the data cache space;

[0109] The specific algorithm is as follows:

[0110] The processed data is subjected to feature calculations including root mean square value, peak value, kurtosis, and skewness, which have certain representative significance, as shown in equations (1), (2), (3), and (4).

[0111]

[0112] Among them, x i is the i-th vibration sampling point, N is the total number of sampling points of the signal;

[0113] Peak P K =max(x i )(2)

[0114]

[0115] Where μ is the mean of the signal, σ is the standard deviation, and x i are the sampling points of the signal.

[0116]

[0117] In order to effectively improve the accuracy of motor fault identification, the present invention will trigger the nonlinear principal component algorithm calculation based on the characteristic algorithm threshold condition at the same time. The algorithm mainly combines the motor vibration and temperature for comprehensive diagnosis. n is the key characterization index, as shown in formula (5):

[0118] P n =W1RMS n +W2S n +W3SS n (5)

[0119]

[0120] Among them, W1, W2, W3 are weights; RMS n is the root mean square value of vibration; S n is the temperature change rate, as shown in formula (6); SS n is the skewness, t n is the time when the vibration RMS value exceeds the limit for the nth time, T(t n ) is the corresponding temperature, t n+1 The effective value of vibration exceeds the limit. m min or t s The moment when the vibration RMS value exceeds the limit for the first time after s, T(t n+1 ) is the corresponding temperature, t m No more than 2min, t s Not shorter than 30s but not longer than 60s;

[0121] (9) Fault diagnosis: the algorithm execution results are modified by combining the priori theory with the inherent characteristics of the motor bearing, and one of the four fault diagnosis results, normal, warning, first-level alarm, and second-level alarm, is given, and the result data is stored in the cache space;

[0122] When P n When the indicator exceeds the limit, the vibration data is Hilbert transformed and the envelope is calculated. The spectrum is further analyzed to identify the bearing characteristic frequency in the envelope signal. The calculation results are corrected by combining the prior theory with the inherent characteristics of the motor bearing and the comparison of the results between the boards to give the fault diagnosis results.

[0123] (10) Data interaction: extracting raw data, calculation result data, and diagnosis result data from the cache space and transmitting them to the vehicle system and ground service platform through vehicle communication and remote transmission for further data utilization; performing inter-board data comparison and analysis on raw data, calculation result data, diagnosis result data, configuration parameter information, and comparison result data through inter-board communication, and updating cache space data in real time to improve the accuracy of fault diagnosis.

[0124] The above is only a specific implementation of the present invention, which enables those skilled in the art to understand or implement the present invention. Although detailed descriptions are given with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments, and they should all be covered by the protection scope of the claims.

Claims

1. A subway traction motor fault diagnosis system, characterized in that: include: The hardware part includes power board, acquisition board, communication board, CPU board and backplane; The input of the power board adopts DC24V and DC110V wide range power supply that meets the railway industry standards; The acquisition board is developed based on FPGA and equipped with signal conditioning circuits. It can adapt to various types of sensors and uses an F48 connector interface. It monitors the temperature and vibration parameters of the motor in real time by connecting sensors to obtain dynamic operation data. The communication board is provided with MVB and Ethernet communication interfaces; The CPU board is developed based on a high-performance ARM chip and is responsible for real-time analysis and processing of collected real-time data, applying corresponding algorithms and models to extract meaningful information and generate results that can be used for fault diagnosis and decision support; The backplane adopts a bus structure and provides high-bandwidth and low-latency interfaces; The software part is designed based on the Linux operating system, and mainly includes the system peripheral interface layer, the operating system interface layer, the function module layer and the business function logic layer; The system peripheral interface layer is responsible for receiving and preprocessing sensor data in real time; The operating system interface layer is responsible for realizing data communication and interaction between the system peripheral interface layer and the business function logic layer; The functional module layer is designed in C language and includes system management business unit, data collection business unit, data management business unit, data calculation business unit, and data interaction business unit; The system management business unit includes three functional modules: system information configuration, system time management and log management, which respectively manage system parameter information, time information and record information; The data acquisition business unit includes an AD data acquisition module, which can collect data according to a preset sampling frequency, and the sampling frequency range is adjustable. The sampling frequency of vibration data should be no less than 10kHz, and the sampling frequency of temperature data should be no less than 1Hz, so as to realize the real-time collection and analog-to-digital conversion of vibration and temperature data. The acquisition module is equipped with a conditioning information configuration interface, which can realize the collection of voltage and current sensor data through configuration. The AD data acquisition module classifies the received acquisition board data to provide a reliable data source for subsequent functional modules; The data management business unit includes three data management function modules: data cache management module, data processing module and data storage module. The data cache management module not only caches the received collected data, but also caches the inter-board communication data, vehicle communication data and remote transmission data, and provides independent cache space for external algorithms to ensure that data is not lost; the data processing module mainly performs related operations such as filtering and outlier removal on the collected data and other received data to ensure the validity of the data; the data storage module realizes the storage of collected data, communication data, calculation result data and operation abnormality information to ensure that the information is traceable; The data calculation business unit includes an algorithm library module and a fault diagnosis module; the algorithm library module performs feature algorithm calculation and nonlinear principal component algorithm calculation based on multi-feature data on the data processed by the data processing module and gives the calculation results; the fault diagnosis module uses the calculation results of the algorithm library module and the a priori theory to compare the inherent characteristics of the motor bearing and the results between the boards, corrects the calculation results, and gives the fault diagnosis results; The data interaction business unit uses a variety of information interaction forms to achieve communication with other systems or devices; The business function logic layer starts the functional processes and threads of the system management business unit, data collection business unit, data management business unit, data calculation business unit and data interaction business unit through logical association to complete the functional logic of each specific business unit.

2. A subway traction motor fault diagnosis system according to claim 1, characterized in that: The power board includes DC9~36V and DC There are two modes: 77~137.5V. The power board is designed with five power interfaces and they are defined as A1 port - DC9~36V positive pole, A2 port - DC7~137.5V positive pole, A3 port - power interface positioning hole, A4 port - common power negative pole. The power board is also designed with ten system power working status indicators. The meanings of each system power working status indicator are as follows: indicator A reflects the external input power, that is, the entire hardware unit receives the external power supply input status, indicator B reflects the power working status of the power board itself, indicators C1 and C2 respectively reflect the power and working status of CPU board 1 and CPU board 2, indicator C3 reflects the power and working status of the communication board, indicators D1, D2, D3, D4 respectively reflect the power and working status of acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4, indicator D5 is a reserved indicator, the indicator light is always green, indicating that the power and working status are normal, the indicator light is always red, indicating that the power supply is abnormal, the indicator light is always green or always red, indicating that the board is abnormal, and the indicator light is off, indicating that there is no power input.

3. A subway traction motor fault diagnosis system according to claim 2, characterized in that: There are two CPU boards in the hardware part, namely CPU board 1 and CPU board 2, and four acquisition boards, namely acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. CPU board 1 and CPU board 2 are of the same board design and can both realize data interaction with acquisition board 1, acquisition board 2, acquisition board 3 and acquisition board 4. The configuration of CPU board 1 and CPU board 2 depends on the number of acquisition boards and the sampling frequency of the data. When the number of acquisition boards does not exceed two or the sampling frequency of the data is low or medium, one acquisition board is installed at any position of CPU board 1 or CPU board 2, and the other position is replaced by a blind plate; when the demand is high, CPU board 1 and CPU board 2 are installed at the same time, and the two realize information interaction through the backplane bus.

4. A subway traction motor fault diagnosis system according to claim 3, characterized in that: The CPU board is designed with two Ethernet interfaces and a group of DB9 interfaces. The two Ethernet interfaces are multiplexed interface designs, taking into account the current vehicle Ethernet interface forms. One Ethernet interface is a 100M Ethernet interface, and the other Ethernet interface is a Gigabit Ethernet interface. The DB9 interface is a multiplexed interface for RS232, RS485 and CAN communications. RS232 mainly provides Debug debugging functions, and RS485 and CAN are used to realize RS485 or CAN communication data interaction with other peripheral devices.

5. A subway traction motor fault diagnosis system according to claim 4, characterized in that: Each acquisition board is equipped with 8-channel vibration data interfaces and 8-channel temperature data interfaces, among which the vibration data interface realizes vibration data acquisition through conditioning circuit configuration.

6. A subway traction motor fault diagnosis system according to claim 5, characterized in that: Each acquisition board is designed with 48 pins. When the actual number of vibration sensors required is less than 8 and the number of temperature sensors required is less than 8, one acquisition board can be used to meet the needs. The acquisition board can be installed at any position in acquisition board 1, acquisition board 2, acquisition board 3 or acquisition board 4, and blind plates can be used at other positions.

7. A subway traction motor fault diagnosis system according to claim 6, characterized in that: The back panel adopts a unified interface design.

8. A subway traction motor fault diagnosis system according to claim 7, characterized in that: The data interaction business unit includes three data interaction functional modules: inter-board communication module, vehicle communication module and remote transmission module; the inter-board communication module includes data communication between CPU board 1 and CPU board 2; the vehicle communication module has three communication modes: Ethernet, RS485 and CAN, to realize the communication of diagnostic results, vehicle information and information data with the vehicle maintenance network and other equipment units; the remote transmission module transmits diagnostic result data, collected data and other data to the service platform, or receives information from the service platform and distributes it to other business functional modules.

9. A subway traction motor fault diagnosis system according to claim 8, characterized in that: The diagnostic results in the fault diagnosis module include four levels: normal, warning, level one alarm, and level two alarm.

10. A method for diagnosing subway traction motor faults, characterized in that: The steps include: (1) System loading initialization: the hardware unit is powered on to complete the loading of the embedded system, and complete the initialization of parameter configuration information, cache space, and communication connection. If the initialization fails, the information will be recorded through the log; (2) Configuration information acquisition: after the system records initialization, the system configuration information needs to be read. This configuration information reflects the current hardware status information of the system and the operation information to be subsequently executed. When the configuration information acquisition fails, the CPU master-slave relationship division mechanism will be started. After the master-slave relationship is clarified, steps (1) and (2) will be re-executed to re-acquire the configuration information to ensure the integrity of the configuration information. (3) CPU master-slave relationship division: The master-slave relationship is divided according to the number of CPUs configured according to the configuration information. When there is only one CPU, the current CPU is the master. When there are two CPUs, the CPU numbered 1 is the master and the CPU numbered 2 is the slave. The master CPU performs the overall system scheduling function and receives diagnostic information from the slave CPU. (4) Time calibration, which provides a time basis for ensuring subsequent data synchronization; (5) Obtain the collected data and make an initial state judgment based on the collected data configured in the configuration information. When the collected data is abnormal, a log record is triggered; (6) Synchronous sampling: To ensure the consistency of data analysis and the accuracy of judgment results, a synchronous sampling signal is set to trigger the synchronous acquisition of data from each channel and record the acquired information through a diary; (7) Data processing: After the acquired data is processed, the data is stored in a specific amount of data according to the parameters set in the configuration information. On the other hand, the data is placed in the cache space as the input condition of the algorithm for use in executing the algorithm. (8) Algorithm execution: extracting calculated data from the cache space to perform feature algorithm calculation, inputting the calculated feature data into a nonlinear principal component algorithm based on multi-feature data to perform nonlinear correlation analysis of the data, and placing the feature calculation and correlation analysis results into the data cache space; (9) Fault diagnosis: The algorithm execution results are modified by combining the priori theory with the inherent characteristics of the motor bearings, and one of the four fault diagnosis results, normal, warning, first-level alarm, and second-level alarm, is given. The result data is stored in the cache space; (10) Data interaction: extracting raw data, calculation result data, and diagnosis result data from the cache space and transmitting them to the vehicle system and ground service platform through vehicle communication and remote transmission for further data utilization; performing inter-board data comparison and analysis on raw data, calculation result data, diagnosis result data, configuration parameter information, and comparison result data through inter-board communication, and updating cache space data in real time to improve the accuracy of fault diagnosis.

Images