WO2021097604A1 - 一种基于多信息融合的变流器的故障预警方法及装置 - Google Patents

一种基于多信息融合的变流器的故障预警方法及装置 Download PDF

Info

Publication number
WO2021097604A1
WO2021097604A1 PCT/CN2019/119160 CN2019119160W WO2021097604A1 WO 2021097604 A1 WO2021097604 A1 WO 2021097604A1 CN 2019119160 W CN2019119160 W CN 2019119160W WO 2021097604 A1 WO2021097604 A1 WO 2021097604A1
Authority
WO
WIPO (PCT)
Prior art keywords
fault
converter
early warning
temperature
characteristic parameter
Prior art date
Application number
PCT/CN2019/119160
Other languages
English (en)
French (fr)
Inventor
徐绍龙
刘永江
贺冠强
李华
陈俊
王亮
臧晓斌
万伟伟
彭宣霖
李榆银
吴书舟
曾祥浩
Original Assignee
株洲中车时代电气股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株洲中车时代电气股份有限公司 filed Critical 株洲中车时代电气股份有限公司
Priority to PCT/CN2019/119160 priority Critical patent/WO2021097604A1/zh
Priority to EP19953096.5A priority patent/EP4063875A4/en
Publication of WO2021097604A1 publication Critical patent/WO2021097604A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/005—Testing of electric installations on transport means
    • G01R31/008—Testing of electric installations on transport means on air- or spacecraft, railway rolling stock or sea-going vessels
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/40—Testing power supplies

Definitions

  • the invention relates to a fault warning technology of a converter, in particular to a fault warning method of a converter based on multiple information fusion, a fault warning device of a converter based on multiple information fusion, and a computer readable medium .
  • Safety is the eternal theme of rail transit. With the increase in the number of rail transit operations and construction mileage in my country and the increasingly complex structure of rail transit vehicles, the guarantee of rail transit safety and reliability is facing severe challenges.
  • the traction converter As the core component of the train's power traction system, the traction converter is directly related to the safety and reliability of its operation.
  • the abnormal parking of the vehicle caused by the failure of the traction converter will not only cause vehicle delays and dispatch confusion, but also cause panic among passengers, causing serious economic losses and adverse social impacts. Therefore, the condition monitoring, fault warning and health management of the traction converter can greatly increase its service life, ensure the safe and reliable operation of the converter, and avoid safety accidents caused by failures and vehicle delays. Economic losses are of great engineering significance and economic value.
  • the present invention provides a converter fault warning method based on multi-information fusion, and a converter based on multi-information fusion.
  • the fault early warning device of the traction converter and a computer readable medium are used for real-time monitoring of the operation status of the traction converter, and early warning and precise positioning of the traction converter failure, thereby ensuring the safe operation of the vehicle.
  • the above-mentioned multi-information fusion-based converter fault early warning method includes: establishing a performance parameter database of the converter, and the performance parameter database includes the converter when at least one fault occurs Collected performance parameter sets of multiple functional components of the converter; performing feature extraction on the performance parameter sets in the performance parameter database to obtain a fault characteristic parameter database, the fault characteristic parameter database including the at least one Faults and at least one fault characteristic parameter group corresponding to each fault, each fault characteristic parameter group including a plurality of fault characteristic parameters related to the plurality of functional components of the converter; and based on the fault characteristic parameter database
  • the at least one fault in and the at least one fault characteristic parameter group corresponding to each fault performs neural network modeling to obtain a fault early warning model representing the mapping relationship between the fault and the fault characteristic parameter.
  • the present invention further includes: comparing the fault characteristics in the fault characteristic parameter database based on the fault characteristic threshold corresponding to the fault degree of each type of fault.
  • the parameters are calibrated to determine the fault degree of the fault corresponding to each fault characteristic parameter group, and the execution is performed based on the at least one fault in the fault characteristic parameter database and at least one fault characteristic parameter group corresponding to each fault.
  • the neural network modeling includes: performing neural network modeling based on the fault state of the at least one fault in the fault characteristic parameter database and at least one fault characteristic parameter group corresponding to each fault state, and the fault state includes a fault Type and the degree of failure corresponding to each type of failure.
  • the performing neural network modeling includes performing modeling using a BP neural network model, and the BP neural network model includes an input layer , A hidden layer and an output layer, the input layer includes input nodes related to multiple fault features corresponding to multiple fault feature parameters in the fault feature parameter group, and the output layer includes the at least one fault state The output node.
  • the output of each hidden node in the hidden layer is x i is the input of the input node of the input layer, w ji and ⁇ j are the connection weight and threshold between each hidden node and each input node, i is the index of the input node, and j is The index of the hidden node, the output of each output node in the output layer is w kj and ⁇ k are the connection weights and thresholds between each output node and each hidden node, respectively, and k is the index of the output node.
  • the activation function of the BP neural network is a Sigmoid function.
  • the error between the output value O k of the output node of the output layer and the expected output value t k is
  • the weight correction function in the reverse pass is
  • the threshold correction function is Among them, ⁇ is between 0.01 and 0.8.
  • the expected output value includes a fault number indicating a fault state.
  • the type of the fault includes one or more of the following: bearing inner ring fault, bearing outer ring fault, bearing ball fault , Bearing cage failure, wind motor balance failure failure, fan phase loss failure, fan turn-to-turn short circuit failure, fan phase imbalance failure, fan grounding fault, fan surge failure, capacitor over-temperature failure, capacitor capacitance loss failure, IGBT over temperature fault, IGBT over current fault, filter blockage fault, water pump fault, heat exchanger fault, water cooling plate fault, transformer over temperature fault, transformer insulation damage fault.
  • the fault characteristic corresponding to the fault characteristic parameter includes one or more of the following: the total value characteristic and frequency spectrum of the wind turbine vibration Characteristics, and envelope characteristics, power factors about fan current, unbalance coefficient, negative sequence current, zero sequence current, spectrum characteristics, envelope characteristics, and energy characteristics, temperature effective values and temperature gradients about transformer temperature, about The current effective value and current time domain characteristics of the module current, the harmonic characteristics of the intermediate voltage of the converter, the temperature effective value and the temperature gradient of the IGBT temperature, the total value characteristics and spectral characteristics of the capacitor current, and the converter The effective value of temperature and temperature gradient of the inlet and outlet water temperature, the effective value of the water pressure on the inlet and outlet of the converter, and the effective value of the air temperature on the inlet and outlet of the filter.
  • the present invention further includes: acquiring the measured performance parameters of the multiple functional components of the converter; and based on the measured performance parameters Calculating fault characteristic parameters; and determining the fault of the converter based on the fault characteristic parameters and the fault early warning model.
  • the performance parameter set corresponding to each fault is collected under different working conditions of the converter, so
  • the operating conditions include the operating environment and mileage of the locomotive where the converter is located.
  • this article also provides a converter fault warning device based on multi-information fusion.
  • the above-mentioned computer-readable medium provided by the present invention has computer-executable instructions stored thereon.
  • the computer-executable instructions When executed by the processor, they can implement any of the above-mentioned fault warning methods for converters based on multi-information fusion.
  • Fig. 1 shows a schematic flow chart of a method for pre-warning a fault of a converter according to an aspect of the present invention.
  • Fig. 2 shows a schematic modeling diagram of a fault warning method for a converter provided according to an embodiment of the present invention.
  • Fig. 3 shows a schematic flow chart of a method for pre-warning a fault of a converter according to an embodiment of the present invention.
  • Fig. 4 shows a schematic flow chart of a method for evaluating the state of a converter according to another aspect of the present invention.
  • Fig. 5 shows a schematic diagram of a wind turbine state evaluation model provided according to an embodiment of the present invention.
  • Fig. 6 shows a schematic diagram of a power module state evaluation model provided according to an embodiment of the present invention.
  • Fig. 7 shows a schematic diagram of a capacitance state evaluation model provided according to an embodiment of the present invention.
  • Fig. 8 shows a schematic diagram of a contactor state evaluation model provided according to an embodiment of the present invention.
  • Fig. 9 shows a schematic diagram of a heat dissipation system state evaluation model provided according to an embodiment of the present invention.
  • Fig. 10 shows a schematic diagram of a transformer state evaluation model provided according to an embodiment of the present invention.
  • Fig. 11 shows a schematic diagram of a spider web evaluation system provided according to an embodiment of the present invention.
  • Fig. 12 shows a schematic flowchart of a method for predicting the life of a converter according to an embodiment of the present invention.
  • Fig. 14 shows a schematic structural diagram of a state evaluation device for a converter according to another aspect of the present invention.
  • Fig. 15 shows an intelligent platform for condition monitoring and fault warning of a converter according to an aspect of the present invention.
  • Fig. 16 shows an intelligent monitoring and fault warning software for the working status of the converter according to an aspect of the present invention.
  • the present invention provides a converter fault warning method based on multi-information fusion, and a converter based on multi-information fusion.
  • the fault early warning device of the traction converter and a computer readable medium are used for real-time monitoring of the operation status of the traction converter, and early warning and precise positioning of the traction converter failure, thereby ensuring the safe operation of the vehicle.
  • FIG. 1 shows a schematic flowchart of a method for pre-warning a fault of a converter according to an aspect of the present invention.
  • the aforementioned performance parameter database may include performance parameter sets of multiple functional components of the converter collected when at least one fault occurs.
  • the functional components of the converter may include all or part of the components in a fan, an IGBT power module, a capacitor, a transformer, a cabinet lifting lug, a heat dissipation system, and a contactor.
  • the converter failure may include one or more of fan failure, IGBT power module failure, capacitor failure, transformer failure, cabinet lifting lug failure, heat dissipation system failure, and contactor failure.
  • the performance parameter set of the functional component may include multiple functional parameters of the functional component. Performance parameters include, but are not limited to, one or more of vibration parameters, current parameters, voltage parameters, temperature parameters, and pressure parameters.
  • the performance parameter may use a specific value to indicate the performance of the corresponding functional component.
  • the parameters in the performance parameter database can come from the historical data of the vehicle-mounted data center, the data of the actual operation test of the line, and the test data of the bench in the maintenance base.
  • the parameters in the performance parameter database can also be obtained through fault simulation tests, for example, for on-site fault characteristics of fans, capacitors, contactors, and heat dissipation systems, artificial fault prototypes are manufactured to collect signals under faults.
  • fan faults may further include bearing inner ring faults, bearing outer ring faults, bearing ball faults, bearing cage faults, wind turbine balance failure faults, fan phase loss faults, and fan turn-to-turn short circuits.
  • the IGBT power module failure may further include an IGBT over-temperature failure and/or an IGBT over-current failure.
  • the capacitor fault may further include a capacitor over-temperature fault and/or a capacitor value loss fault.
  • the transformer fault may further include a transformer over-temperature fault and/or a transformer insulation destruction fault.
  • the heat dissipation system failure may further include one or more of filter clogging failure, water pump failure, heat exchanger failure, and water cooling plate failure.
  • the above-mentioned fault early warning method for a converter based on multi-information fusion can be implemented on a fault early warning device for a converter based on multi-information fusion.
  • the converter fault early warning device can use the sensors provided in each functional component of the traction converter to collect the performance parameters of each functional component, so as to establish the performance parameter database of the converter according to the performance parameter sets of multiple functional components.
  • the performance parameter set corresponding to each fault can be collected under different working conditions of the converter.
  • the working conditions of the converter may include the operating environment and mileage of the locomotive where the converter is located.
  • the fault characteristic refers to the characteristic that has a certain directivity to the fault of each component of the converter.
  • the fault characteristic parameter is the specific value of the fault characteristic, which has a certain correlation with the fault state.
  • the fault characteristics may include total value characteristics, frequency spectrum characteristics, and envelope characteristics regarding wind turbine vibration, power factors regarding wind turbine currents, unbalance coefficients, negative sequence currents, zero sequence currents, spectrum characteristics, and envelopes.
  • the effective value of the temperature and temperature gradient of the transformer temperature the effective value of the current and the current time domain characteristics of the module current, the harmonic characteristics of the intermediate voltage of the converter, the effective temperature of the IGBT temperature and the Temperature change gradient, about the total value characteristics and spectral characteristics of the capacitor current, about the temperature effective value and temperature change gradient of the inlet and outlet water temperature of the converter, about the effective value of the water pressure of the inlet and outlet water pressure of the converter, about the inlet and outlet of the filter
  • One or more of the effective temperature values of the air temperature One or more of the effective temperature values of the air temperature.
  • the above-mentioned feature extraction can be implemented through multiple iterative verifications of faulty prototypes and line operation fault feature data.
  • the feature extraction of the performance parameter set in the performance parameter database can be realized by combining the fault characteristics of the traction converter and adopting mature and reliable feature extraction algorithms in the industry, which will not be repeated here.
  • the above-mentioned fault characteristic parameter database may include at least one fault and at least one fault characteristic parameter group corresponding to each fault.
  • each fault characteristic parameter group may include multiple fault characteristic parameters related to multiple functional components of the converter. As shown in Figure 1, in the above-mentioned multi-information fusion-based converter fault early warning method provided by the present invention, the following steps may be further included:
  • 103 Based on at least one fault in the fault characteristic parameter database and at least one fault characteristic parameter group corresponding to each fault, perform neural network modeling to obtain a fault early warning model representing the mapping relationship between the fault and the fault characteristic parameter.
  • the fault early warning device of the converter based on multi-information fusion can be executed based on the fault state of at least one fault in the fault characteristic parameter database and at least one fault characteristic parameter group corresponding to each fault state. Neural network modeling.
  • the above fault status may include the fault type and the fault degree corresponding to each fault type.
  • the fault degree of the fault corresponding to each fault characteristic parameter group can be determined by calibrating the fault characteristic parameters in the fault characteristic parameter database by using one or more fault characteristic thresholds.
  • Each fault characteristic threshold can correspond to a critical point between two fault levels of a fault.
  • the fault feature threshold here may be an individual threshold corresponding to only one fault feature, or a combined threshold corresponding to a combination of fault features.
  • multiple fault characteristic thresholds corresponding to various fault levels of various faults can be determined through various fault experiments and/or expert experience corresponding to various faults.
  • FIG. 2 shows a schematic modeling diagram of a fault warning method for a converter according to an embodiment of the present invention.
  • performing the neural network modeling described above may include using a BP neural network model to perform modeling.
  • the BP neural network model can include an input layer, a hidden layer and an output layer.
  • the aforementioned input layer may include input nodes for multiple fault features corresponding to multiple fault feature parameters in each fault feature parameter group.
  • the input layer is further divided into a signal layer and a feature layer.
  • the signal layer is the original signal measured by the sensor, which can include fan vibration current and voltage, transformer temperature, module current, converter intermediate voltage, IGBT temperature, capacitance current and temperature, converter inlet and outlet water temperature, converter inlet and outlet Water pressure, air temperature at the inlet and outlet of the filter.
  • the feature layer is the feature parameter obtained by feature extraction and feature calculation of each signal, which can include the total value, frequency spectrum and envelope characteristics of fan vibration, fan power factor, three-phase unbalance coefficient, negative sequence current, zero sequence current, frequency spectrum Characteristics, envelope characteristics and energy characteristics, transformer temperature effective value and temperature change gradient, module current effective value and time domain characteristics, intermediate voltage harmonic characteristics, IGBT temperature effective value and temperature change gradient, capacitor temperature effective value and The temperature change gradient, the total value characteristics and frequency spectrum characteristics of the capacitive current device, the effective value and temperature change gradient of the inlet and outlet water temperature, the effective value of the inlet and outlet water pressure, the effective value of the inlet and outlet air temperature and the temperature change gradient.
  • the sample parameters of the input layer can be obtained through fault simulation tests and actual circuit tests.
  • the aforementioned hidden layer may include multiple hidden nodes.
  • the number of hidden layers may be one layer.
  • the number of hidden nodes in the hidden layer can be determined according to the number of input nodes in the input layer and the number of output nodes in the output layer.
  • the number of hidden nodes may be the sum of the number of nodes in the input layer and the output layer.
  • the output of each hidden node in the hidden layer can be x i is the input of the input node of the input layer, w ji and ⁇ j are the connection weight and threshold between each hidden node and each input node, i is the index of the input node, and j is The index of the hidden node.
  • the above-mentioned output layer may include at least one output node in a fault state.
  • the output layer may be the output corresponding to the fault diagnosis result.
  • the output layer may include fan bearing inner ring failure, outer ring failure, rolling element failure, cage failure, wind motor balance failure failure, fan phase loss failure, inter-turn short circuit failure, phase imbalance failure, Ground fault, surge fault, capacitor over-temperature fault, capacitance loss fault, IGBT over-temperature fault, over-current fault, filter blockage fault, water pump fault, heat exchanger fault, water cooling plate fault, transformer over-temperature fault, insulation fault One or more of.
  • the output of each output node in the output layer can be w kj and ⁇ k are the connection weights and thresholds between each output node and each hidden node, k is the index of the output node, and f is the transfer function.
  • the activation function of the BP neural network can be a Sigmoid function, namely The parameter ⁇ can be determined according to the characteristics of the neural network model.
  • the error between the output value O k of the output node of the output layer and the expected output value t k may be
  • the weight correction function in the reverse transfer can be
  • the threshold correction function can be Among them, ⁇ can be between 0.01 and 0.8.
  • the aforementioned expected output value t k may include a fault number indicating a fault state.
  • the fault warning device of the converter based on multi-information fusion can use binary coding to number the fault status.
  • the fault degree of the fault status can be described by a two-digit binary code, where 00 can represent a good status, 01 can represent a minor fault, 10 can represent a moderate fault, and 11 can represent a serious fault.
  • FIG. 3 shows a schematic flowchart of a method for pre-warning a fault of a converter according to an embodiment of the present invention.
  • the above-mentioned fault early warning method for converters based on multi-information fusion can be further Including steps:
  • the converter fault early warning device may use sensors provided in each functional component of the traction converter to collect the measured performance parameters of each functional component.
  • the measured performance parameter can indicate the performance pros and cons of the corresponding functional component under the current working condition.
  • the working conditions of the converter may include the operating environment and mileage of the locomotive where the converter is located.
  • the converter fault early warning device may perform feature extraction on the acquired set of measured performance parameters to calculate its corresponding fault feature parameters.
  • the fault condition of the converter can be automatically determined, and the fault diagnosis result can be output through the output layer of the fault warning model.
  • the converter fault early warning device may output a fault warning in response to determining that the converter has a fault condition to prompt maintenance personnel to perform maintenance in time. In some embodiments, the converter fault early warning device may also output a corresponding fault warning based on the determined fault condition of the converter to prompt maintenance personnel to check and repair the corresponding functional components in time.
  • this article also provides a method for evaluating the state of a converter based on multi-information fusion.
  • the above-mentioned multi-information fusion-based converter state evaluation method provided by the present invention may be implemented on a multi-information fusion converter-based state evaluation device.
  • FIG. 4 shows a schematic flow chart of a method for evaluating the state of a converter according to another aspect of the present invention.
  • the performance parameter database described above may include a set of historical performance parameters collected about at least one functional component of the converter, where each functional component may be associated with multiple sets of performance parameters stored.
  • the functional components of the converter may include all or part of the components in a fan, an IGBT power module, a capacitor, a transformer, a cabinet lifting lug, a heat dissipation system, and a contactor.
  • the performance parameter set of the functional component may include multiple functional parameters of the functional component. Performance parameters include, but are not limited to, one or more of vibration parameters, current parameters, voltage parameters, temperature parameters, and pressure parameters.
  • the performance parameter may use a specific value to indicate the performance of the corresponding functional component.
  • the converter state evaluation device may use sensors provided in each functional component of the traction converter to collect the performance parameters of each functional component, so as to establish the converter based on the performance parameter sets of multiple functional components. Database of performance parameters.
  • the parameters in the performance parameter database can come from the historical data of the vehicle-mounted data center, the data of the actual operation test of the line, and the test data of the bench in the maintenance base.
  • these performance parameters are collected under different working conditions of the converter.
  • the working conditions of the converter may include the operating environment and mileage of the locomotive where the converter is located.
  • the performance parameters of the fan may further include fan vibration parameters, fan current parameters, and fan voltage parameters.
  • the performance parameters related to the IGBT power module may further include module current parameters and IGBT temperature parameters.
  • the performance parameters related to the capacitor may further include a capacitor temperature parameter and a capacitor current parameter.
  • the performance parameters related to the contactor may further include contactor coil voltage parameters and coil current parameters.
  • the performance parameters of the heat dissipation system may further include the inlet and outlet water temperature parameters of the converter, the inlet and outlet water pressure parameters, and the inlet and outlet air temperature parameters of the filter.
  • the performance parameters related to the transformer may further include transformer temperature parameters and transformer vibration parameters.
  • each performance characteristic parameter group may include multiple performance characteristic parameters related to multiple functional components of the converter.
  • Performance characteristics refer to the characteristics that have a certain directivity to the performance of each component of the converter.
  • Performance characteristic parameters are specific values of performance characteristics, which are related to the degree of performance.
  • the performance characteristics corresponding to the performance characteristic parameters may include one or more of vibration intensity, frequency spectrum characteristics, envelope characteristics, power factors, unbalance coefficients, negative sequence currents, and zero sequence currents of the wind turbine.
  • the performance characteristic corresponding to the performance characteristic parameter may include one or more of the current effective value, the current time domain characteristic, and the temperature effective value of the IGBT power module.
  • the performance characteristic corresponding to the performance characteristic parameter may include one or more of the effective temperature value, the capacitance value, and the equivalent series resistance (ESR) value of the capacitor.
  • ESR equivalent series resistance
  • the performance characteristic corresponding to the performance characteristic parameter may include one or more of the contactor coil resistance and the breaking time of the contactor.
  • the performance characteristic corresponding to the performance characteristic parameter may include one or more of the effective value of the temperature, the effective value of the vibration acceleration, and the spectral characteristic of the vibration acceleration with respect to the heat dissipation system.
  • the above-mentioned feature extraction can be implemented through multiple iteration verifications of the converter prototype and the line operating performance feature data.
  • the feature extraction of the performance parameter set in the performance parameter database can be specifically realized by combining the performance characteristics of the traction converter and adopting mature and reliable feature extraction algorithms in the industry, which will not be repeated here.
  • multiple sets of performance characteristic parameters of each functional component may be calibrated based on performance characteristic thresholds corresponding to different performances of each functional component.
  • Each performance characteristic threshold can correspond to a critical point between two pros and cons of a performance.
  • the performance characteristic threshold here may be an individual threshold corresponding to only one performance characteristic, or a combined threshold corresponding to a combination of performance characteristics.
  • multiple performance characteristic thresholds corresponding to the respective pros and cons of various performances may be determined through performance experiments and/or expert experience of various corresponding performances.
  • the calibrated performance characteristic parameter may determine a key performance index value according to the threshold range to which it belongs. The key performance index value can be between 0 and 1, which is used to indicate the degree of the corresponding performance.
  • the above-mentioned performing neural network modeling may include using a BP neural network model to perform modeling.
  • the BP neural network model can include an input layer, a hidden layer and an output layer.
  • the aforementioned input layer may include input nodes corresponding to multiple performance characteristics corresponding to each set of performance characteristic parameters of the functional component.
  • the aforementioned hidden layer may include multiple hidden nodes.
  • the foregoing output layer may include output nodes corresponding to the key performance index values of the functional components.
  • FIG. 5 shows a schematic diagram of a wind turbine state evaluation model provided according to an embodiment of the present invention.
  • the input layer is further divided into a signal layer and a feature layer.
  • the signal layer is the original signal measured by the sensor, which can include fan vibration parameters, fan current parameters, and fan voltage parameters.
  • the feature layer is the feature parameter obtained by feature extraction and feature calculation of each signal. It can include vibration intensity features, frequency spectrum features, and envelope features corresponding to the vibration parameters of the fan; it can include power factors corresponding to the fan current parameters and the fan voltage parameters. Characteristics, three-phase unbalance coefficient characteristics, negative sequence current characteristics and zero sequence current characteristics.
  • the sample parameters of the input layer can be obtained through performance simulation tests and actual circuit tests.
  • FIG. 6 shows a schematic diagram of a power module state evaluation model provided according to an embodiment of the present invention.
  • the signal layer may include module current parameters and IGBT temperature parameters.
  • the characteristic layer may include current effective value characteristics and current time domain characteristics corresponding to the module current parameters; and temperature effective value characteristics corresponding to the IGBT temperature parameters.
  • FIG. 7 shows a schematic diagram of a capacitance state evaluation model provided according to an embodiment of the present invention.
  • the signal layer may include a capacitance temperature parameter and a capacitance current parameter.
  • the characteristic layer may include a temperature effective value characteristic corresponding to a capacitance temperature parameter; and a capacitance capacitance value characteristic and a capacitance ESR value characteristic corresponding to the capacitance current parameter.
  • FIG. 8 shows a schematic diagram of a contactor state evaluation model provided according to an embodiment of the present invention.
  • the signal layer may include contactor coil voltage parameters and contactor coil current parameters.
  • the characteristic layer may include contactor coil resistance characteristics corresponding to contactor coil voltage parameters and contactor coil current parameters; and contactor breaking time characteristics.
  • FIG. 9 shows a schematic diagram of a heat dissipation system state evaluation model provided according to an embodiment of the present invention.
  • the signal layer may include the inlet and outlet water temperature parameters of the converter, the inlet and outlet water pressure parameters of the converter, and the inlet and outlet air temperature parameters of the converter.
  • the characteristic layer may include the characteristics of the outlet water temperature exceeding the target value magnitude corresponding to the inlet and outlet water temperature parameters of the converter, the inlet and outlet water pressure parameters of the converter, and the inlet and outlet air temperature parameters of the converter, and the outlet air temperature exceeding the target value magnitude. feature.
  • FIG. 10 shows a schematic diagram of a transformer state evaluation model provided according to an embodiment of the present invention.
  • the error between the output value O k of the output node of the output layer and the expected output value t k may be
  • the weight correction function in the reverse transfer can be
  • the threshold correction function can be Among them, ⁇ can be between 0.01 and 0.8.
  • FIG. 11 shows a schematic diagram of a spider web evaluation system according to an embodiment of the present invention.
  • the spider web evaluation system can display the key performance index values of fans, IGBTs, capacitors, transformers, cabinet lifting eyes, heat dissipation systems, and contactors at the same time.
  • 1.0 can indicate that the performance of the corresponding functional component is good; 0.8 can indicate that the performance of the corresponding functional component is slightly degraded; 0.6 can indicate that the performance of the corresponding functional component is degraded; 0.4 can indicate that the performance of the corresponding functional component is severely degraded; 0.2 can indicate the corresponding function The component is critically damaged; 0.0 can indicate that the corresponding functional component is damaged.
  • the state evaluation device of the converter based on multi-information fusion may also use at least one functional component Obtain the performance characteristic parameter group from the actual measured performance parameters, and determine the key performance index value of at least one functional component based on the obtained performance characteristic parameter group and the performance evaluation model.
  • the aforementioned at least one functional component may be all functional components of the traction converter.
  • the state evaluation device of the converter may use sensors provided in each functional component of the traction converter to collect the measured performance parameters of each functional component.
  • the state evaluation device of the converter may perform feature extraction on the acquired set of measured performance parameters to calculate its corresponding performance characteristic parameters. By further inputting the calculated performance characteristic parameters into the hidden layer of the performance evaluation model, the performance status of each functional component of the converter can be automatically determined, and the status evaluation results can be output through the output layer of the performance evaluation model.
  • FIG. 12 shows a schematic flowchart of a method for predicting the life of a converter according to an embodiment of the present invention.
  • the above-mentioned multi-information fusion-based converter state evaluation method can further include the steps:
  • the remaining life t remaining of the functional component can be solved by solving the equation to make sure.
  • ⁇ I is the difference between the current key performance index value of the functional component to be tested and the preset key performance index value threshold.
  • the fitting curve obtained by this fitting and its function expression h(t) can simulate the downward trend of the key performance index value of the functional component to be tested, so as to calculate the future trend of the key performance index value of the functional component to be tested, and predict The remaining life t of the functional component to be tested remains . It can be understood that the remaining life t remaining of the functional component to be tested refers to the time required for the current key performance index value of the functional component to be tested to drop to the aforementioned preset key performance index value threshold.
  • the device for evaluating the state of the converter based on multi-information fusion may determine the shortest life of at least one functional component as the life of the converter.
  • the above-mentioned fault warning method of the converter based on multi-information fusion and the state evaluation method of the converter based on multi-information fusion proposed by the present invention can be achieved by analyzing the operation of the wind turbine, capacitor, and IGBT during the operation of the converter.
  • the above-mentioned fault warning method for converters based on multi-information fusion can also establish a multi-dimensional status assessment system for the above-mentioned key materials, systems and structures by fully extracting and mining existing standards, expert experience and past data, so as to achieve Intelligent monitoring and evaluation of the working status of the converter.
  • the above-mentioned method for early warning of converter faults based on multi-information fusion can conduct in-depth learning of the key performance indicators and fault evolution trends of the key materials of the converter through the entire life cycle, and establish the current state of the key materials, the mileage and the mileage of operation. Environment, predict the failure time and mode of key materials, and predict the life of key materials of the converter.
  • this article also provides a converter fault warning device based on multi-information fusion.
  • FIG. 13 shows a schematic structural diagram of a fault early warning device for a converter according to another aspect of the present invention.
  • the above-mentioned fault warning device of the converter based on multi-information fusion provided by the present invention may include a memory 131 and a processor 132.
  • the processor 132 may be coupled to the memory 131, and configured to implement the fault warning method of a converter based on multi-information fusion provided by any one of the above embodiments, so as to check the operating status of the traction converter. Carry out real-time monitoring, and provide early warning and precise positioning of the failure of the traction converter.
  • a computer-readable medium is also provided herein.
  • the above-mentioned computer-readable medium provided by the present invention has computer-executable instructions stored thereon.
  • the computer-executable instructions can implement any of the above-mentioned fault warning methods for converters based on multi-information fusion, so as to monitor the operating status of the traction converter in real time, and perform real-time monitoring of the traction converter. Early warning and precise positioning of faults.
  • this article also provides a converter state assessment device based on multi-information fusion.
  • FIG. 14 shows a schematic structural diagram of a state evaluation device for a converter according to another aspect of the present invention.
  • the above-mentioned multi-information fusion-based converter state evaluation device may include a memory 141 and a processor 142.
  • the processor 142 may be coupled to the memory 141 and configured to implement the state evaluation method of a converter based on multiple information fusion provided by any one of the above embodiments, so as to evaluate the operating state of the traction converter. Carry out real-time monitoring and predict the service life of the traction converter.
  • a computer-readable medium is also provided herein.
  • the intelligent platform may mainly include a state perception layer, a data collection and preprocessing layer, and a state evaluation and fault warning layer.
  • the state sensing layer can be composed of various sensors to realize the vibration, current and voltage of the fan, support the temperature and current of the capacitor, the temperature of the IGBT, the voltage and current of the contactor, the vibration of the cabinet lifting lug, the vibration of the transformer, the temperature, and the temperature of the flow channel of the converter.
  • the perception of the sensor, and the physical quantity of perception is transferred to the acquisition system in the form of an analog quantity.
  • the measurement frequency range must cover the entire fault characteristic frequency, the measurement accuracy must meet the requirements of feature extraction, the use conditions of the converter must be met, and the installation must be firm and reliable.
  • the core of the data acquisition and preprocessing layer is used for the A/D acquisition module for analog data acquisition and the data processing unit for front-end edge calculation, which can realize the acquisition of analog data such as current, voltage, vibration and temperature in the state sensing layer.
  • analog data such as current, voltage, vibration and temperature in the state sensing layer.
  • TCU Through optical fiber communication with TCU, the collection of intermediate voltage, speed, power, finishing module input current and inverter module output current, fan speed and contact status of the traction system is realized.
  • Collect the analog signals transmitted in the state sensing layer and TCU perform segmentation and selection operations on the data structure, perform simple Fourier analysis and calculation of state characteristic quantities on the operated data, and transmit the processed data through the optical fiber Give status assessment and fault early warning layer.
  • the hardware core of the state assessment and fault early warning layer is the phm (prediction and health management) data processing unit, and the software core is the intelligent monitoring and fault early warning software for the working status of the converter.
  • the main function is to extract the characteristics of the transmitted data and calculate the state evaluation index.
  • the established state evaluation system and fault diagnosis model accurately evaluate the operating state of the fan, capacitor, IGBT, contactor, auxiliary transformer, and heat dissipation system.
  • Fault early warning evaluate and early warning of the vibration load, dynamic stress and environmental temperature of the converter cabinet, and predict the life of IGBT, fan, capacitor and contactor, and pass the status evaluation information and fault early warning information through the Ethernet
  • the network is passed to the on-board intelligent center.
  • the computing performance of the Phm data analysis unit can be configurable and selected according to the requirements of computing and storage capabilities.
  • the working mode of the intelligent converter state monitoring and fault early warning system is that the key state characteristics are picked up by the sensors of the state sensing layer and converted into analog quantity for online transmission.
  • the data acquisition and preprocessing layer collects the analog signal from the sensing layer, according to the advance
  • the established rules and models intercept, select, and front-end computing the data into digital signals for transmission to the upper layer.
  • the state evaluation and fault early warning layer receives the data from the collection layer, performs state feature index calculation and fault feature extraction, and then uses the existing state evaluation system and fault model and life prediction model to perform fault early warning, state evaluation and life prediction, and finally The information is passed to the vehicle intelligent center and ground operation and maintenance center for the formulation of intelligent operation and maintenance strategies.
  • the software can include a condition monitoring and evaluation module, a logic analysis and diagnosis module, a life prediction module, and a visualization module.
  • the core of the state monitoring and evaluation module is the state feature calculation rules and evaluation system established for the entire converter cabinet and each key material. Monitor the fan vibration, current and voltage to evaluate the vibration intensity, power factor and dust accumulation of the fan; monitor the temperature and current characteristics of the capacitor, realize the evaluation of the capacitance value, EST value and temperature, and monitor the water temperature of the inverter module and the temperature of the IGBT And the inverter module current to realize the evaluation of IGBT junction temperature, life and module operating temperature; monitor the current and voltage of the contactor coil to realize the evaluation of the contactor coil temperature and coil insulation; monitor the transformer, reactor, capacitor and internal
  • the temperature of the cavity can be used to evaluate the ambient temperature of the converter and the operating status of the transformer and reactor; the vibration of the transformer can be monitored to realize the evaluation of the vibration of the transformer; the vibration of the lifting lug of the converter can be monitored to realize the load environment and the lifting lug of the converter Assessment of dynamic stress.
  • the core of the logic analysis and diagnosis module is the fault diagnosis and early warning model established for fans, capacitors, contactors, IGBTs, heat dissipation systems, and cabinet lifting ears.
  • Monitor fan vibration realize fault diagnosis and early warning of bearing, rotor eccentricity and dynamic balance damage, monitor fan current and voltage, realize fault diagnosis and early warning of fan phase loss, grounding, stator insulation and inter-turn short circuit, and monitor capacitor temperature and pressure
  • monitor capacitor temperature and pressure realize the early warning of capacitor over-temperature, over-voltage and excessive drop in capacitance.
  • Monitor the contactor current, voltage and control commands to realize the diagnosis and early warning of contactor action faults, coil over-temperature and discharge faults.
  • the core of the life prediction module is the life prediction model for IGBTs, capacitors, fans and contactors.
  • the service life of IGBTs, capacitors, fans and contactors are predicted by monitoring IGBT turn-off characteristics, capacitance value, fan bearing and coil insulation, and contactor coil aging evolution rules.
  • the visualization module is mainly for displaying real-time fault waveforms, operating status evaluation graphs, early warning information and historical process data, with good display effects.
  • the operating status evaluation graphs are planned to use normalized spider web graphs.
  • DSP digital signal processors
  • ASIC application-specific integrated circuits
  • FPGA field programmable gate arrays
  • a general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine.
  • the processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.
  • the steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be directly embodied in hardware, in a software module executed by a processor, or in a combination of the two.
  • the software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
  • An exemplary storage medium is coupled to the processor such that the processor can read information from and write information to the storage medium.
  • the storage medium may be integrated into the processor.
  • the processor and the storage medium may reside in the ASIC.
  • the ASIC may reside in the user terminal.
  • the processor and the storage medium may reside as discrete components in the user terminal.
  • the described functions may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function can be stored as one or more instructions or codes on a computer-readable medium or transmitted through it.
  • Computer-readable media includes both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another.
  • the storage medium may be any available medium that can be accessed by a computer.
  • such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices, or can be used to carry or store instructions or data in the form Any other medium that agrees with the program code and can be accessed by a computer.
  • any connection is also properly called a computer-readable medium.
  • the software is transmitted from a web site, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave .
  • coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the medium.
  • Disks and discs as used in this article include compact discs (CD), laser discs, optical discs, digital versatile discs (DVD), floppy disks and Blu-ray discs, among which disks are often reproduced in magnetic Data, and a disc (disc) optically reproduces the data with a laser. Combinations of the above should also be included in the scope of computer-readable media.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Power Conversion In General (AREA)
  • Inverter Devices (AREA)

Abstract

一种基于多信息融合的变流器的故障预警方法及装置，以及一种计算机可读介质。该故障预警方法包括：建立变流器的性能参数数据库（101），性能参数数据库中包括变流器在至少一种故障发生时采集的变流器的多个功能部件的性能参数集；对性能参数数据库中的性能参数集执行特征提取以获得故障特征参数数据库（102），故障特征参数数据库包括至少一种故障以及与每一故障对应的至少一个故障特征参数组，每个故障特征参数组包括关于变流器的多个功能部件的多个故障特征参数；以及基于故障特征参数数据库中的至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模以获得表现故障与故障特征参数之间的映射关系的故障预警模型（103）。

Description

一种基于多信息融合的变流器的故障预警方法及装置 技术领域
本发明涉及变流器的故障预警技术，尤其涉及一种基于多信息融合的变流器的故障预警方法、一种基于多信息融合的变流器的故障预警装置，以及一种计算机可读介质。
背景技术
安全是轨道交通永恒的主题。随着我国轨道交通运营、建设里程数的增加以及轨道交通车辆结构的越发复杂，轨道交通安全与可靠性的保障面临严峻挑战。牵引变流器作为列车动力牵引系统的核心部件，其运行的安全性及可靠性，直接关系到行车安全。因牵引变流器故障导致的车辆异常停车，不但会造成车辆延误和调度的混乱，而且会造成乘客的恐慌，造成严重的经济损失和恶劣的社会影响。因此，对牵引变流器的状态监测、故障预警和健康管理，能极大提高其使用寿命，保证变流器安全可靠的运行，避免因故障失效而导致的安全事故及因车辆延误带来的经济损失，具备重大的工程意义和经济价值。
目前，轨道交通领域对牵引变流器的故障监控还处于故障报警阶段，往往需要在故障发生后进行故障的排查和检修。这种故障报警方式的运维成本较高而效率较低，非常不利于车辆的安全运行。
为了克服现有技术存在的上述缺陷以满足轨道交通领域产品对安全和可靠的需求，本领域亟需一种高效的变流器的故障预警技术，用于对牵引变流器运行状态进行实时监测、对牵引变流器的故障进行提前预警和精准定位，并对牵引变流器的使用寿命进行预测，从而保障车辆的安全运行。
发明内容
以下给出一个或多个方面的简要概述以提供对这些方面的基本理解。此概述不是所有构想到的方面的详尽综览，并且既非旨在指认出所有方面的关键性或决定性要素亦非试图界定任何或所有方面的范围。其唯一的目的是要以简化形式给出一个或多个方面的一些概念以为稍后给出的更加详细的描述之序。
为了克服现有技术存在的上述缺陷以满足轨道交通领域产品对安全和可靠的需求，本发明提供了一种基于多信息融合的变流器的故障预警方法、一种基于多信息融 合的变流器的故障预警装置，以及一种计算机可读介质，用于对牵引变流器运行状态进行实时监测，并对牵引变流器的故障进行提前预警和精准定位，从而保障车辆的安全运行。
本发明提供的上述基于多信息融合的变流器的故障预警方法，包括：建立所述变流器的性能参数数据库，所述性能参数数据库中包括所述变流器在至少一种故障发生时采集的所述变流器的多个功能部件的性能参数集；对所述性能参数数据库中的性能参数集执行特征提取以获得故障特征参数数据库，所述故障特征参数数据库包括所述至少一种故障以及与每一故障对应的至少一个故障特征参数组，每个故障特征参数组包括关于所述变流器的所述多个功能部件的多个故障特征参数；以及基于所述故障特征参数数据库中的所述至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模以获得表现故障与故障特征参数之间的映射关系的故障预警模型。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，还包括：基于对应每种故障的故障程度的故障特征阈值对所述故障特征参数数据库中的故障特征参数进行标定以确定每一故障特征参数组所对应的故障的故障程度，所述基于所述故障特征参数数据库中的所述至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模包括：基于所述故障特征参数数据库中的所述至少一种故障的故障状态及与每一故障状态对应的至少一个故障特征参数组执行神经网络建模，所述故障状态包括故障类型和对应每种故障类型的故障程度。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述执行神经网络建模包括使用BP神经网络模型执行建模，所述BP神经网络模型包括输入层、隐含层和输出层，所述输入层包括关于所述故障特征参数组中的多个故障特征参数所对应的多个故障特征的输入节点，所述输出层包括所述至少一种故障状态的输出节点。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述隐含层中每个隐含节点的输出为
Figure PCTCN2019119160-appb-000001
x i为所述输入层的输入节点的输入，w ji和θ j分别为每个隐含节点和每个输入节点之间的连接权值和阈值，i为所述输入节点的索引，j为所述隐含节点的索引，所述输出层中每个输出节点的输出为
Figure PCTCN2019119160-appb-000002
w kj和θ k分别为每个输出节点和每个隐含节点之间的连接权值和阀值，k为所述输出节点的索引。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述BP神经网络的激活函数为Sigmoid函数。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述输出层的输出节点的输出值O k与期望输出值t k之间的误差为
Figure PCTCN2019119160-appb-000003
反向传递中的权值修正函数为
Figure PCTCN2019119160-appb-000004
阀值修正函数为
Figure PCTCN2019119160-appb-000005
其中η取0.01～0.8之间。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述期望输出值包括指示故障状态的故障编号。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述故障的类型包括以下一者或多者：轴承内圈故障、轴承外圈故障、轴承滚珠故障、轴承保持架故障、风机动平衡破坏故障、风机缺相故障、风机匝间短路故障、风机相不平衡故障、风机接地故障、风机喘振故障、电容超温故障、电容容值损失故障、IGBT超温故障、IGBT过流故障、滤网堵塞故障、水泵故障、热交换器故障、水冷板故障、变压器超温故障、变压器绝缘破坏故障。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，所述故障特征参数对应的故障特征包括以下一者或多者：关于风机振动的总值特征、频谱特征、和包络特征、关于风机电流的功率因素、不平衡系数、负序电流、零序电流、频谱特征、包络特征、和能量特征、关于变压器温度的温度有效值和温度变化梯度、关于模块电流的电流有效值和电流时域特征、关于变流器中间电压的谐波特征、关于IGBT温度的温度有效值和温度变化梯度、关于电容电流的总值特征和频谱特征、关于变流器进出口水温的温度有效值和温度变化梯度、关于变流器进出口水压的水压有效值、关于滤网进出口空气温度的温度有效值。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，还包括：获取所述变流器的所述多个功能部件实测性能参数；基于所述实测性能参数计算故障特征参数；以及基于所述故障特征参数和所述故障预警模型确定所述变流器的故障。
可选地，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，对应每一故障的所述性能参数集是在所述变流器的不同工况下采集的，所述工况包括所述变流器所在的机车的运行环境和里程。
根据本发明的另一方面，本文还提供了一种基于多信息融合的变流器的故障预警装置。
根据本发明的另一方面，本文还提供了一种计算机可读介质。
本发明提供的上述计算机可读介质，其上存储有计算机可执行指令。所述计算机可执行指令在由处理器执行时，可以实施上述任意一种基于多信息融合的变流器的故障预警方法。
附图说明
在结合以下附图阅读本公开的实施例的详细描述之后，能够更好地理解本发明的上述特征和优点。在附图中，各组件不一定是按比例绘制，并且具有类似的相关特性或特征的组件可能具有相同或相近的附图标记。
图1示出了根据本发明的一方面提供的变流器的故障预警方法的流程示意图。
图2示出了根据本发明的一个实施例提供的变流器的故障预警方法的建模示意图。
图3示出了根据本发明的一个实施例提供的变流器的故障预警方法的流程示意图。
图4示出了根据本发明的另一方面提供的变流器的状态评估方法的流程示意图。
图5示出了根据本发明的一个实施例提供的风机状态评估模型的示意图。
图6示出了根据本发明的一个实施例提供的功率模块状态评估模型的示意图。
图7示出了根据本发明的一个实施例提供的电容状态评估模型的示意图。
图8示出了根据本发明的一个实施例提供的接触器状态评估模型的示意图。
图9示出了根据本发明的一个实施例提供的散热系统状态评估模型的示意图。
图10示出了根据本发明的一个实施例提供的变压器状态评估模型的示意图。
图11示出了根据本发明的一个实施例提供的蛛网评估体系的示意图。
图12示出了根据本发明的一个实施例提供的预测变流器寿命的方法的流程示意图。
图13示出了根据本发明的另一方面提供的变流器的故障预警装置的结构示意图。
图14示出了根据本发明的另一方面提供的变流器的状态评估装置的结构示意图。
图15示出了根据本发明的一方面的变流器的状态监测及故障预警智能平台。
图16示出了根据本发明的一方面的变流器的工作状态智能监测及故障预警软件。
附图标记
101-103       变流器的故障预警方法的步骤；
301-306       变流器的故障预警方法的步骤；
401-404       变流器的状态评估方法的步骤；
131           存储器；
132           处理器；
141           存储器；
142           处理器。
具体实施方式
以下由特定的具体实施例说明本发明的实施方式，本领域技术人员可由本说明书所揭示的内容轻易地了解本发明的其他优点及功效。虽然本发明的描述将结合优选实施例一起介绍，但这并不代表此发明的特征仅限于该实施方式。恰恰相反，结合实施方式作发明介绍的目的是为了覆盖基于本发明的权利要求而有可能延伸出的其它选择或改造。为了提供对本发明的深度了解，以下描述中将包含许多具体的细节。本发明也可以不使用这些细节实施。此外，为了避免混乱或模糊本发明的重点，有些具体细节将在描述中被省略。
为了克服现有技术存在的上述缺陷以满足轨道交通领域产品对安全和可靠的需求，本发明提供了一种基于多信息融合的变流器的故障预警方法、一种基于多信息融合的变流器的故障预警装置，以及一种计算机可读介质，用于对牵引变流器运行状态进行实时监测，并对牵引变流器的故障进行提前预警和精准定位，从而保障车辆的安全运行。
请参考图1，图1示出了根据本发明的一方面提供的变流器的故障预警方法的流程示意图。
如图1所示，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，可以包括步骤：
101：建立变流器的性能参数数据库。
在一些实施例中，上述性能参数数据库中可以包括变流器在至少一种故障发生时 采集的变流器的多个功能部件的性能参数集。在一些实施例中，变流器的功能性部件可以包括风机、IGBT功率模块、电容器、变压器、柜体吊耳、散热系统、接触器中的全部部件或部分部件。在一些实施例中，变流器故障可以包括风机故障、IGBT功率模块故障、电容器故障、变压器故障、柜体吊耳故障、散热系统故障、接触器故障中的一种或多种。在一些实施例中，功能部件的性能参数集可以包括该功能部件的多个功能参数。性能参数包括但不限于振动参数、电流参数、电压参数、温度参数、压强参数中的一种或多种。在一些实施例中，性能参数可以通过具体的取值来指示对应功能部件的性能优劣。
实践中，性能参数数据库中的参数可以来自于车载数据中心的历史数据、线路实际运行测试的数据以及在检修基地的台架测试数据。此外，性能参数数据库中的参数还可以通过故障模拟试验获得，例如针对风机、电容、接触器、散热系统的现场故障特性，人为制造故障样机，进行故障下的信号采集。
具体来说，在一些实施例中，风机故障可以进一步包括轴承内圈故障、轴承外圈故障、轴承滚珠故障、轴承保持架故障、风机动平衡破坏故障、风机缺相故障、风机匝间短路故障、风机相不平衡故障、风机接地故障、风机喘振故障中的一种或多种。在一些实施例中，IGBT功率模块故障可以进一步包括IGBT超温故障和/或IGBT过流故障。在一些实施例中，电容器故障可以进一步包括电容超温故障和/或电容容值损失故障。在一些实施例中，变压器故障可以进一步包括变压器超温故障和/或变压器绝缘破坏故障。在一些实施例中，散热系统故障可以进一步包括滤网堵塞故障、水泵故障、热交换器故障、水冷板故障中的一种或多种。
在一些实施例中，本发明提供的上述基于多信息融合的变流器的故障预警方法，可以在一个基于多信息融合的变流器的故障预警装置上实施。该变流器故障预警装置可以使用设于牵引变流器的各功能部件的传感器来采集各功能部件的性能参数，从而根据多个功能部件的性能参数集来建立变流器的性能参数数据库。较优地，对应每一故障的性能参数集可以是在变流器的不同工况下采集的。该变流器的工况可以包括变流器所在的机车的运行环境和里程。
如图1所示，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，还可以包括步骤：
102：对性能参数数据库中的性能参数集执行特征提取，以获得故障特征参数数据库。
故障特征是指对变流器的各个部件的故障有一定指向性的特征。故障特征参数是 故障特征的具体数值，与故障状态有一定的关联性。在一些实施例中，故障特征可以包括关于风机振动的总值特征、频谱特征、和包络特征、关于风机电流的功率因素、不平衡系数、负序电流、零序电流、频谱特征、包络特征、和能量特征、关于变压器温度的温度有效值和温度变化梯度、关于模块电流的电流有效值和电流时域特征、关于变流器中间电压的谐波特征、关于IGBT温度的温度有效值和温度变化梯度、关于电容电流的总值特征和频谱特征、关于变流器进出口水温的温度有效值和温度变化梯度、关于变流器进出口水压的水压有效值、关于滤网进出口空气温度的温度有效值中的一者或多者。
上述特征提取可以通过故障样机和线路运行故障特征数据的多次迭代验证来实施。对性能参数数据库中的性能参数集的特征提取，具体可以结合牵引变流器的故障特性，并采用行业内成熟可靠的特征提取算法来实现，在此不做赘述。
在一些实施例中，上述故障特征参数数据库可以包括至少一种故障，以及与每一故障对应的至少一个故障特征参数组。在一些实施例中，每个故障特征参数组可以包括关于变流器的多个功能部件的多个故障特征参数。如图1所示，在本发明提供的上述基于多信息融合的变流器的故障预警方法中，还可以包括步骤：
103：基于故障特征参数数据库中的至少一种故障及与每一故障对应的至少一个故障特征参数组，执行神经网络建模以获得表现故障与故障特征参数之间的映射关系的故障预警模型。
在一些实施例中，基于多信息融合的变流器的故障预警装置可以基于故障特征参数数据库中的至少一种故障的故障状态，以及与每一故障状态对应的至少一个故障特征参数组来执行神经网络建模。上述故障状态可以包括故障类型和对应每种故障类型的故障程度。在一些实施例中，每一故障特征参数组所对应的故障的故障程度，可以通过采用一个或多个故障特征阈值对故障特征参数数据库中的故障特征参数进行标定来确定。每个故障特征阈值可以对应一种故障的两个故障程度之间的临界点。这里的故障特征阈值可以是仅对应一个故障特征的个体阈值，也可以是对应一个故障特征组合的组合阈值。
在一些实施例中，对应各种故障的各故障程度的多个故障特征阈值，可以通过各种对应故障的故障实验和/或专家经验来确定。
请参考图2，图2示出了根据本发明的一个实施例提供的变流器的故障预警方法的建模示意图。
如图2所示，在一些实施例中，上述执行神经网络建模，可以包括使用BP神经 网络模型来执行建模。该BP神经网络模型可以包括输入层、隐含层和输出层。
具体来说，上述输入层可以包括关于各故障特征参数组中的多个故障特征参数所对应的多个故障特征的输入节点。在一些实施例中，输入层进一步分为信号层和特征层。信号层为传感器测量得到的原始信号，可以包括风机振动电流和电压、变压器温度、模块电流、变流器中间电压、IGBT温度、电容电流和温度、变流器进出口水温、变流器进出口水压、滤网进出口空气温度。特征层为对各信号进行特征提取和特征运算得到的特征参量，可以包括风机振动的总值、频谱和包络特征，风机功率因素、三相不平衡系数、负序电流、零序电流、频谱特征、包络特征和能量特征，变压器温度有效值和温度变化梯度、模块电流有效值和时域特征、中间电压的谐波特征、IGBT的温度有效值和温度变化梯度、电容的温度有效值和温度变化梯度、电容电流器的总值特征和频谱特征、进出口水温的有效值和温度变化梯度、进出口水压的有效值、进出口空气温度的有效值和温度变化梯度。在一些实施例中，输入层的样本参量可以通过故障模拟试验和线路实际测试得到。
上述隐含层可以包括多个隐含节点。在一些实施例中，隐含层的层数可以为一层。隐含层的隐含节点的数量可以根据输入层的输入节点数量和输出层的输出节点数量来确定。在一些实施例种，隐含节点的数量可以取输入层和输出层节点数的总和。在一些实施例中，隐含层中每个隐含节点的输出可以为
Figure PCTCN2019119160-appb-000006
x i为所述输入层的输入节点的输入，w ji和θ j分别为每个隐含节点和每个输入节点之间的连接权值和阈值，i为所述输入节点的索引，j为所述隐含节点的索引。
上述输出层可以包括至少一种故障状态的输出节点。在一些实施例中，输出层可以为对应故障诊断结果的输出。在一些实施例中，输出层可以包括风机轴承内圈故障、外圈故障、滚动体故障、保持架故障、风机动平衡破坏故障、风机缺相故障、匝间短路故障、相不平衡故障、接地故障、喘振故障、电容超温故障、容值损失故障、IGBT超温故障、过流故障、滤网堵塞故障、水泵故障、热交换器故障、水冷板故障、变压器超温故障、绝缘故障中的一种或多种。在一些实施例中，输出层中每个输出节点的输出可以为
Figure PCTCN2019119160-appb-000007
w kj和θ k分别为每个输出节点和每个隐含节点之间的连接权值和阀值，k为所述输出节点的索引，f为传递函数。
在一些实施例中，BP神经网络的激活函数为可以Sigmoid函数，即
Figure PCTCN2019119160-appb-000008
参数β可以根据神经网络模型的特性进行确定。
在一些实施例中，输出层的输出节点的输出值O k与期望输出值t k之间的误差可以为
Figure PCTCN2019119160-appb-000009
在一些实施例中，反向传递中的权值修正函数可以为
Figure PCTCN2019119160-appb-000010
在一些实施例中，阀值修正函数可以为
Figure PCTCN2019119160-appb-000011
其中η可以取0.01～0.8之间。
在一些实施例中，上述期望输出值t k可以包括指示故障状态的故障编号。如图2所示，基于多信息融合的变流器的故障预警装置可以采用二进制编码来对故障状态进行编号。在一些实施例中，故障状态的故障程度可以由两位二进制编码来描述，其中，00可以代表状态良好，01可以代表轻微故障，10可以代表中度故障，11可以代表严重故障。
请参考图3，图3示出了根据本发明的一个实施例提供的变流器的故障预警方法的流程示意图。
如图3所示，在本发明的一个实施例中，在获得表现故障与故障特征参数之间的映射关系的故障预警模型之后，上述基于多信息融合的变流器的故障预警方法还可以进一步包括步骤：
304：获取变流器的多个功能部件实测性能参数；
305：基于实测性能参数计算故障特征参数；以及
306：基于故障特征参数和故障预警模型确定变流器的故障。
在一些实施例中，变流器故障预警装置可以使用设于牵引变流器的各功能部件的传感器来采集各功能部件的实测性能参数。该实测性能参数可以指示当前工况下的对应功能部件的性能优劣。该变流器的工况可以包括变流器所在的机车的运行环境和里程。
在一些实施例中，变流器故障预警装置可以对获取的实测性能参数集执行特征提取，以计算其对应的故障特征参数。通过将计算获得的故障特征参数进一步输入到故障预警模型的隐含层，即可自动确定变流器存在的故障情况，并通过故障预警模型的输出层输出故障诊断结果。
在一些实施例中，变流器故障预警装置可以响应于确定变流器存在故障情况，而输出故障预警以提示维护人员及时进行检修。在一些实施例中，变流器故障预警装置还可以根据确定变流器存在的故障情况，而输出相应的故障预警以提示维护人员及时对相应的功能性部件进行检修。
根据本发明的另一方面，本文还提供了一种基于多信息融合的变流器的状态评估方法。在一些实施例中，本发明提供的上述基于多信息融合的变流器的状态评估方法，可以在一个基于多信息融合的变流器的状态评估装置上实施。
请参考图4，图4示出了根据本发明的另一方面提供的变流器的状态评估方法的流程示意图。
如图4所示，在本发明提供的上述基于多信息融合的变流器的状态评估方法中，可以包括步骤：
401：建立变流器的性能参数数据库。
在一些实施例中，上述性能参数数据库中可以包括关于变流器的至少一个功能部件所采集的历史性能参数集，其中每个功能部件可以相关联地存储有多组性能参数。
在一些实施例中，变流器的功能性部件可以包括风机、IGBT功率模块、电容器、变压器、柜体吊耳、散热系统、接触器中的全部部件或部分部件。在一些实施例中，功能部件的性能参数集可以包括该功能部件的多个功能参数。性能参数包括但不限于振动参数、电流参数、电压参数、温度参数、压强参数中的一种或多种。在一些实施例中，性能参数可以通过具体的取值来指示对应功能部件的性能优劣。在一些实施例中，变流器状态评估装置可以使用设于牵引变流器的各功能部件的传感器来采集各功能部件的性能参数，从而根据多个功能部件的性能参数集来建立变流器的性能参数数据库。
实践中，性能参数数据库中的参数可以来自于车载数据中心的历史数据、线路实际运行测试的数据以及在检修基地的台架测试数据。较优地，这些性能参数是在变流器的不同工况下采集的。该变流器的工况可以包括变流器所在的机车的运行环境和里程。
具体来说，在一些实施例中，关于风机的性能参数可以进一步包括风机振动参数、风机电流参数和风机电压参数。在一些实施例中，关于IGBT功率模块的性能参数可以进一步包括模块电流参数和IGBT温度参数。在一些实施例中，关于电容器的性能参数可以进一步包括电容温度参数和电容电流参数。在一些实施例中，关于接触器的性能参数可以进一步包括接触器线圈电压参数和线圈电流参数。在一些实施例中，关于散热系统的性能参数可以进一步包括变流器进出口水温参数、进出口水压参数和滤网进出口空气温度参数。在一些实施例中，关于变压器的性能参数可以进一步包括变压器温度参数和变压器振动参数。
如图4所示，在本发明提供的上述基于多信息融合的变流器的状态评估方法中， 还可以包括步骤：
402：对性能参数数据库中每个功能部件的性能参数执行特征提取，以获得关于每个功能部件的多组性能特征参数。
在一些实施例中，每个性能特征参数组可以包括关于变流器的多个功能部件的多个性能特征参数。性能特征是指对变流器的各个部件的性能有一定指向性的特征。性能特征参数是性能特征的具体数值，与性能的优劣程度具有关联性。
在一些实施例中，性能特征参数对应的性能特征可以包括关于风机的振动烈度、频谱特征、包络特征、功率因素、不平衡系数、负序电流和零序电流中的一者或多者。在一些实施例中，性能特征参数对应的性能特征可以包括关于IGBT功率模块的电流有效值、电流时域特征和温度有效值中的一者或多者。在一些实施例中，性能特征参数对应的性能特征可以包括关于电容器的温度有效值、电容容值和电容等效串联电阻(Equivalent series resistance，ESR)值中的一者或多者。在一些实施例中，性能特征参数对应的性能特征可以包括关于接触器的接触器线圈电阻和分断时间中的一者或多者。在一些实施例中，性能特征参数对应的性能特征可以包括关于散热系统的温度有效值、振动加速度有效值和振动加速度频谱特征中的一者或多者。
上述特征提取可以通过变流器样机和线路运行性能特征数据的多次迭代验证来实施。对性能参数数据库中的性能参数集的特征提取，具体可以结合牵引变流器的性能特性，并采用行业内成熟可靠的特征提取算法来实现，在此不做赘述。
如图4所示，在本发明提供的上述基于多信息融合的变流器的状态评估方法中，还可以包括步骤：
403：基于对应每个功能部件的不同性能的性能特征阈值，对每个功能部件的多组性能特征参数进行标定，以确定每组性能特征参数所对应的关键性能指标值。
在一些实施例中，每个功能部件的多组性能特征参数可以基于对应每个功能部件的不同性能的性能特征阈值来进行标定。每个性能特征阈值可以对应一种性能的两个优劣程度之间的临界点。这里的性能特征阈值可以是仅对应一个性能特征的个体阈值，也可以是对应一个性能特征组合的组合阈值。
在一些实施例中，对应各种性能的各优劣程度的多个性能特征阈值，可以通过各种对应性能的性能实验和/或专家经验来确定。在一些实施例中，经过标定的性能特征参数可以根据所属的阈值范围而确定一个关键性能指标值。该关键性能指标值可以介于0～1之间，用于指示对应性能的优劣程度。
如图4所示，在本发明提供的上述基于多信息融合的变流器的状态评估方法中，还可以包括步骤：
404：基于每个功能部件的多组性能特征参数及其对应的关键性能指标值执行神经网络建模，以获得表现各功能部件的关键性能指标值与性能特征参数之间的映射关系的性能评估模型。
在一些实施例中，上述执行神经网络建模可以包括使用BP神经网络模型来执行建模。该BP神经网络模型可以包括输入层、隐含层和输出层。上述输入层可以包括对应功能部件的每组性能特征参数所对应的多个性能特征的输入节点。上述隐含层可以包括多个隐含节点。上述输出层可以包括对应功能部件的关键性能指标值的输出节点。
请参考图5，图5示出了根据本发明的一个实施例提供的风机状态评估模型的示意图。
如图5所示，在一些实施例中，输入层进一步分为信号层和特征层。信号层为传感器测量得到的原始信号，可以包括风机振动参数、风机电流参数和风机电压参数。特征层为对各信号进行特征提取和特征运算得到的特征参量，可以包括对应于风机振动参数的振动烈度特征、频谱特征、包络特征；可以包括对应于风机电流参数和风机电压参数的功率因素特征、三相不平衡系数特征、负序电流特征和零序电流特征。在一些实施例中，输入层的样本参量可以通过性能模拟试验和线路实际测试得到。
请参考图6，图6示出了根据本发明的一个实施例提供的功率模块状态评估模型的示意图。
如图6所示，在一些实施例中，信号层可以包括模块电流参数和IGBT温度参数。特征层可以包括对应于模块电流参数的电流有效值特征和电流时域特征；以及对应于IGBT温度参数的温度有效值特征。
请参考图7，图7示出了根据本发明的一个实施例提供的电容状态评估模型的示意图。
如图7所示，在一些实施例中，信号层可以包括电容温度参数和电容电流参数。特征层可以包括对应于电容温度参数的温度有效值特征；以及对应于电容电流参数的电容容值特征和电容ESR值特征。
请参考图8，图8示出了根据本发明的一个实施例提供的接触器状态评估模型的示意图。
如图8所示，在一些实施例中，信号层可以包括接触器线圈电压参数和接触器线 圈电流参数。特征层可以包括对应于接触器线圈电压参数和接触器线圈电流参数的接触器线圈电阻特征；以及接触器分断时间特征。
请参考图9，图9示出了根据本发明的一个实施例提供的散热系统状态评估模型的示意图。
如图9所示，在一些实施例中，信号层可以包括变流器进出口水温参数、变流器进出口水压参数和变流器进出口空气温度参数。特征层可以包括对应于变流器进出口水温参数、变流器进出口水压参数和变流器进出口空气温度参数的出口水温超出目标值量级特征，以及出口风温超出目标值量级特征。
请参考图10，图10示出了根据本发明的一个实施例提供的变压器状态评估模型的示意图。
如图10所示，在一些实施例中，信号层可以包括变压器温度参数和变压器振动参数。特征层可以包括对应于变压器温度参数的温度有效值特征；以及对应于变压器振动参数的振动加速度有效值特征和振动加速度频谱特征。
在一些实施例中，隐含层的层数可以为一层。隐含层的隐含节点的数量可以根据输入层的输入节点数量和输出层的输出节点数量来确定。在一些实施例种，隐含节点的数量可以取输入层和输出层节点数的总和。在一些实施例中，隐含层中每个隐含节点的输出可以为
Figure PCTCN2019119160-appb-000012
x i为所述输入层的输入节点的输入，w ji和θ j分别为每个隐含节点和每个输入节点之间的连接权值和阈值，i为所述输入节点的索引，j为所述隐含节点的索引。
在一些实施例中，输出层可以为对应状态评估结果的输出。在一些实施例中，输出层中每个输出节点的输出可以为
Figure PCTCN2019119160-appb-000013
w kj和θ k分别为每个输出节点和每个隐含节点之间的连接权值和阀值，k为所述输出节点的索引，f为传递函数。
在一些实施例中，BP神经网络的激活函数为可以Sigmoid函数，即
Figure PCTCN2019119160-appb-000014
参数β可以根据神经网络模型的特性进行确定。
在一些实施例中，输出层的输出节点的输出值O k与期望输出值t k之间的误差可以为
Figure PCTCN2019119160-appb-000015
在一些实施例中，反向传递中的权值修正函数可以为
Figure PCTCN2019119160-appb-000016
在一些实施例中，阀值修正函数可以为
Figure PCTCN2019119160-appb-000017
其中η可以取0.01～0.8之间。
在一些实施例中，基于多信息融合的变流器的状态评估装置可以通过蛛网评估体系来进行变流器的状态评估。上述期望输出值t k可以包括介于0～1的关键性能指标值。
请参考图11，图11示出了根据本发明的一个实施例提供的蛛网评估体系的示意图。
如图11所示，在一些实施例中，蛛网评估体系可以显示同时显示风机、IGBT、电容、变压器、柜体吊耳、散热系统、接触器的关键性能指标值。在一些实施例中，1.0可以指示对应功能部件性能良好；0.8可以指示对应功能部件性能略有衰退；0.6可以指示对应功能部件性能衰退；0.4可以指示对应功能部件性能严重衰退；0.2可以指示对应功能部件临界损坏；0.0可以指示对应功能部件损坏。
在本发明的一个实施例中，在获得表现关键性能指标值与性能特征参数之间的映射关系的性能评估模型之后，基于多信息融合的变流器的状态评估装置还可以使用至少一个功能部件的实测性能参数获得性能特征参数组，并基于获得的性能特征参数组和性能评估模型来确定至少一个功能部件的关键性能指标值。
在一些实施例中，上述至少一个功能部件可以是牵引变流器的所有功能部件。在一些实施例中，变流器的状态评估装置可以使用设于牵引变流器的各功能部件的传感器来采集各功能部件的实测性能参数。在一些实施例中，变流器的状态评估装置可以对获取的实测性能参数集执行特征提取，以计算其对应的性能特征参数。通过将计算获得的性能特征参数进一步输入到性能评估模型的隐含层，即可自动确定变流器各功能部件的性能状态，并通过性能评估模型的输出层输出状态评估结果。
请参考图12，图12示出了根据本发明的一个实施例提供的预测变流器寿命的方法的流程示意图。
如图12所示，在本发明的一个实施例中，在获得表现关键性能指标值与性能特征参数之间的映射关系的性能评估模型之后，上述基于多信息融合的变流器的状态评估方法还可以进一步包括步骤：
1201：基于至少一个功能部件的性能评估模型和持续监测获得的性能参数，确定至少一个功能部件的关键性能指标值的历史数据；
1202：基于至少一个功能部件的关键性能指标值的历史数据执行曲线拟合，以获得关于至少一个功能部件的关键性能指标值退化趋势曲线；以及
1203：基于至少一个功能部件的当前关键性能指标值和对应的关键性能指标值退化趋势曲线，确定至少一个功能部件的寿命。
在一些实施例中，基于多信息融合的变流器的状态评估装置可以持续监测获得待 预测寿命的功能部件的性能参数，并根据该功能部件的性能评估模型确定相应的关键性能指标值的多个历史数据。之后，基于多信息融合的变流器的状态评估装置可以利用解析表达式逼近离散数据的拟合手段得到一条拟合曲线，即该功能部件的关键性能指标值退化趋势曲线，并得到该拟合曲线的函数表达式h(t)。该函数表达式h(t)可以指示该功能部件的关键性能指标值在时刻t的下降梯度函数。
在一些实施例中，该功能部件的剩余寿命t 剩余可以通过求解方程
Figure PCTCN2019119160-appb-000018
来确定。式中，ΔI为待测功能部件的当前关键性能指标值和预设的关键性能指标值阈值的差值。通过该拟合获得的拟合曲线及其函数表达式h(t)可以模拟待测功能部件的关键性能指标值的下降趋势，从而推算待测功能部件的关键性能指标值以后的走向，并预测待测功能部件的剩余寿命t 剩余。可以理解的是，待测功能部件的剩余寿命t 剩余是指从待测功能部件的当前关键性能指标值下降到上述预设的关键性能指标值阈值所需的时间。
在一些实施例中，基于多信息融合的变流器的状态评估装置可以将至少一个功能部件的最短寿命确定为变流器的寿命。
基于以上描述，本发明提出的上述基于多信息融合的变流器的故障预警方法，以及基于多信息融合的变流器的状态评估方法，可以通过对变流器运行过程中风机、电容、IGBT、接触器、变压器、电抗器、散热系统、吊耳的电流、电压、振动及温度多维度信号进行特征提取和特征融合，建立基于特征融合的故障预警模型，从而对变流器各功能性器件进行精准定位。上述基于多信息融合的变流器的故障预警方法，还可以通过对现有标准、专家经验和以往数据的充分提炼和挖掘，建立上述关键物料、系统和结构的多维状态评估体系，从而实现对变流器的工作状态的智能监测和评估。上述基于多信息融合的变流器的故障预警方法，可以通过对变流器关键物料全生命周期关键性能指标及故障演化趋势进行深度学习，并通过建立的根据关键物料当前状态、运行的里程及环境，对关键物料失效时间及方式的预测，并对变流器关键物料进行寿命预测。
尽管为使解释简单化将上述方法图示并描述为一系列动作，但是应理解并领会，这些方法不受动作的次序所限，因为根据一个或多个实施例，一些动作可按不同次序发生和/或与来自本文中图示和描述或本文中未图示和描述但本领域技术人员可以理解的其他动作并发地发生。
根据本发明的另一方面，本文还提供了一种基于多信息融合的变流器的故障预警装置。
请参考图13，图13示出了根据本发明的另一方面提供的变流器的故障预警装置的结构示意图。
如图13所示，本发明提供的上述基于多信息融合的变流器的故障预警装置可以包括存储器131和处理器132。在一些实施例中，处理器132可以耦接于存储器131，并配置用于实施上述任意一个实施例所提供的基于多信息融合的变流器的故障预警方法，从而对牵引变流器运行状态进行实时监测，并对牵引变流器的故障进行提前预警和精准定位。
根据本发明的另一方面，本文还提供了一种计算机可读介质。
本发明提供的上述计算机可读介质，其上存储有计算机可执行指令。该计算机可执行指令在由处理器132执行时，可以实施上述任意一种基于多信息融合的变流器的故障预警方法，从而对牵引变流器运行状态进行实时监测，并对牵引变流器的故障进行提前预警和精准定位。
根据本发明的另一方面，本文还提供了一种基于多信息融合的变流器的状态评估装置。
请参考图14，图14示出了根据本发明的另一方面提供的变流器的状态评估装置的结构示意图。
如图14所示，本发明提供的上述基于多信息融合的变流器的状态评估装置可以包括存储器141和处理器142。在一些实施例中，处理器142可以耦接于存储器141，并配置用于实施上述任意一个实施例所提供的基于多信息融合的变流器的状态评估方法，从而对牵引变流器运行状态进行实时监测，并对牵引变流器的使用寿命进行预测。
根据本发明的另一方面，本文还提供了一种计算机可读介质。
本发明提供的上述计算机可读介质，其上存储有计算机可执行指令。该计算机可执行指令在由处理器142执行时，可以实施上述任意一个实施例所提供的基于多信息融合的变流器的状态评估方法，从而对牵引变流器运行状态进行实时监测，并对牵引变流器的使用寿命进行预测。
尽管上述的实施例所述的处理器132、142可以通过软件与硬件的组合来实现。但是可以理解，处理器132、142也可以单独地在软件或硬件中加以实施。对于硬件实施而言，处理器132、142可以在一个或多个专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理器件(DAPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、处理器、控制器、微控制器、微处理器、用于执行上述功能的 其它电子装置或上述装置的选择组合来加以实施。对软件实施而言，处理器132、142可以通过在通用芯片上运行的诸如程序模块(procedures)和函数模块(functions)等独立的软件模块来加以实施，其中每一个模块可以执行一个或多个本文中描述的功能和操作。
基于上述多信息融合的故障预警、状态评估和寿命预测方案，可开发出变流器的状态监测及故障预警智能平台，如图15所示。该智能平台可主要包括状态感知层、数据采集及预处理层、和状态评估及故障预警层。
状态感知层可由各类传感器组成，实现对风机振动、电流和电压，支撑电容温度和电流，IGBT温度，接触器电压、电流，柜体吊耳振动，变压器振动、温度，变流器流道温度的感知，并将感知物理量以模拟量的形式传递给采集系统。这里需选用可靠性高、稳定性好的传感器，测量频率范围需覆盖整个故障特征频率，测量精度需满足特征提取的要求，需满足变流器使用条件要求，安装需牢固可靠。
数据采集及预处理层的核心用于模拟量采集的A/D采集模块及用于前端边沿计算的数据处理单元，可实现对状态感知层电流、电压、振动和温度等模拟量的采集，同时通过光纤与TCU进行通讯实现对牵引系统中间电压、转速、功率、整理模块输入电流及逆变模块输出电流、风机转速及接触器触点状态的采集。对状态感知层及TCU中传递的模拟信号进行采集，对数据结构进行分割、甄选操作，对操作后的数据进行简单的傅里叶分析及状态特征量的计算，将经过处理的数据通过光纤传输给状态评估和故障预警层。
状态评估及故障预警层的硬件核心为phm(预测与健康管理)数据处理单元，软件核心为变流器工作状态智能监测和故障预警软件。主要功能为对传输过来的数据进行特征提取和状态评估指标计算，根据建立的状态评估体系和故障诊断模型，对风机、电容、IGBT、接触器、辅助变压器、散热系统的运行状态进行准确评估及故障预警，对变流器柜体振动载荷、动态应力及环境温度的状态进行评估和预警，并对IGBT、风机、电容及接触器的寿命进行预测，并将状态评估信息和故障预警信息通过以太网传递给车载智能中心。Phm数据分析单元的运算性能，能根据运算和存储能力的需求进行可配置化选择。
智能变流器状态监测及故障预警系统工作模式为通过状态感知层的传感器对各关键状态特性进行拾取并转化成模拟量网上传输，数据采集及预处理层采集来自感知层的模拟信号，按照预先制定的规则和模式对数据进行截取、甄选及前端运算转化为数字信号往上层传输。状态评估及故障预警层接收来自采集层的数据，进行状态特征 指标计算和故障特征提取，再分别采用已有状态评估体系和故障模型和寿命预测模型进行故障预警、状态评估及寿命预测，最后将信息传递给车载智能中心及地面运维中心进行智能运维策略的制定。
此外，可基于上述多信息融合的故障预警、状态评估和寿命预测方案，开发一套变流器工作状态智能监测及故障预警软件。如图16所示，该软件可包括状态监测及评估模块、逻辑分析及诊断模块、寿命预测模块和可视化模块。
状态监测及评估模块的核心为针对变流器柜体整体及各关键物料所建立起的状态特征运算规则及评估体系。监测风机振动、电流及电压实现对风机振动烈度、功率因素及积灰程度的评估；监测电容温度和电流特征，实现对电容容值、EST值及温度的评估，监测逆变模块水温、IGBT温度及逆变模块电流，实现对IGBT结温、寿命及模块运行温度的评估；监测接触器线圈的电流、电压实现对接触器线圈温度和线圈绝缘情况的评估；监测变压器、电抗器、电容及内部腔体温度，实现对变流器环境温度、变压器及电抗器运行状态的评估；监测变压器振动，实现对变压器振动的评估；监测变流器吊耳振动，实现对变流器载荷环境及吊耳动态应力的评估。
逻辑分析及诊断模块的核心为针对风机、电容、接触器、IGBT、散热系统、柜体吊耳所建立的故障诊断和预警模型。监测风机振动，实现对轴承、转子偏心及动平衡破坏的故障诊断和预警，监测风机电流、电压实现对风机缺相、接地、定子绝缘及匝间短路的故障诊断和预警，监测电容温度、压力及电流及中间电压实现对电容超温、过压及容值过度下降的预警。监测接触器电流、电压及控制指令实现对接触器动作故障、线圈超温及放电故障的诊断和预警。监测IGBT电流、中间电压及水冷板温度实现对IGBT超温、过流及电流异常预警。监测各关键物料及变流器内部腔体温度、模块水温、流量、牵引功率、速度、及风机转速，实现对滤网堵塞故障、水泵故障、热交换器故障、水冷板导热故障的诊断和预警。监测变压器的温度和电流实现变压器超温预警及绝缘破坏预警。监测柜体吊耳振动，实现对吊耳动态载荷过大预警。
寿命预测模块的核心为针对IGBT、电容、风机及接触器的寿命预测模型。分别通过监测IGBT关段特性、电容容值、风机轴承及线圈绝缘、接触器线圈老化的演变规律，预测IGBT、电容、风机及接触器的使用寿命。
可视化模块主要为对实时故障波形、运行状态评估图、预警信息及历史过程数据的显示，具备良好的展示效果，其中运行状态评估图拟采用归一化后的蛛网图。
本领域技术人员将可理解，信息、信号和数据可使用各种不同技术和技艺中的任何技术和技艺来表示。例如，以上描述通篇引述的数据、指令、命令、信息、信号、 位(比特)、码元、和码片可由电压、电流、电磁波、磁场或磁粒子、光场或光学粒子、或其任何组合来表示。
本领域技术人员将进一步领会，结合本文中所公开的实施例来描述的各种解说性逻辑板块、模块、电路、和算法步骤可实现为电子硬件、计算机软件、或这两者的组合。为清楚地解说硬件与软件的这一可互换性，各种解说性组件、框、模块、电路、和步骤在上面是以其功能性的形式作一般化描述的。此类功能性是被实现为硬件还是软件取决于具体应用和施加于整体系统的设计约束。技术人员对于每种特定应用可用不同的方式来实现所描述的功能性，但这样的实现决策不应被解读成导致脱离了本发明的范围。
结合本文所公开的实施例描述的各种解说性逻辑模块、和电路可用通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或其它可编程逻辑器件、分立的门或晶体管逻辑、分立的硬件组件、或其设计成执行本文所描述功能的任何组合来实现或执行。通用处理器可以是微处理器，但在替换方案中，该处理器可以是任何常规的处理器、控制器、微控制器、或状态机。处理器还可以被实现为计算设备的组合，例如DSP与微处理器的组合、多个微处理器、与DSP核心协作的一个或多个微处理器、或任何其他此类配置。
结合本文中公开的实施例描述的方法或算法的步骤可直接在硬件中、在由处理器执行的软件模块中、或在这两者的组合中体现。软件模块可驻留在RAM存储器、闪存、ROM存储器、EPROM存储器、EEPROM存储器、寄存器、硬盘、可移动盘、CD-ROM、或本领域中所知的任何其他形式的存储介质中。示例性存储介质耦合到处理器以使得该处理器能从/向该存储介质读取和写入信息。在替换方案中，存储介质可以被整合到处理器。处理器和存储介质可驻留在ASIC中。ASIC可驻留在用户终端中。在替换方案中，处理器和存储介质可作为分立组件驻留在用户终端中。
在一个或多个示例性实施例中，所描述的功能可在硬件、软件、固件或其任何组合中实现。如果在软件中实现为计算机程序产品，则各功能可以作为一条或更多条指令或代码存储在计算机可读介质上或藉其进行传送。计算机可读介质包括计算机存储介质和通信介质两者，其包括促成计算机程序从一地向另一地转移的任何介质。存储介质可以是能被计算机访问的任何可用介质。作为示例而非限定，这样的计算机可读介质可包括RAM、ROM、EEPROM、CD-ROM或其它光盘存储、磁盘存储或其它磁存储设备、或能被用来携带或存储指令或数据结构形式的合意程序代码且能被计算机访问的任何其它介质。任何连接也被正当地称为计算机可读介质。例如，如果软件是 使用同轴电缆、光纤电缆、双绞线、数字订户线(DSL)、或诸如红外、无线电、以及微波之类的无线技术从web网站、服务器、或其它远程源传送而来，则该同轴电缆、光纤电缆、双绞线、DSL、或诸如红外、无线电、以及微波之类的无线技术就被包括在介质的定义之中。如本文中所使用的盘(disk)和碟(disc)包括压缩碟(CD)、激光碟、光碟、数字多用碟(DVD)、软盘和蓝光碟，其中盘(disk)往往以磁的方式再现数据，而碟(disc)用激光以光学方式再现数据。上述的组合也应被包括在计算机可读介质的范围内。
提供对本公开的先前描述是为使得本领域任何技术人员皆能够制作或使用本公开。对本公开的各种修改对本领域技术人员来说都将是显而易见的，且本文中所定义的普适原理可被应用到其他变体而不会脱离本公开的精神或范围。由此，本公开并非旨在被限定于本文中所描述的示例和设计，而是应被授予与本文中所公开的原理和新颖性特征相一致的最广范围。

Claims (23)

  1. 一种基于多信息融合的变流器的故障预警方法，包括：
    建立所述变流器的性能参数数据库，所述性能参数数据库中包括所述变流器在至少一种故障发生时采集的所述变流器的多个功能部件的性能参数集；
    对所述性能参数数据库中的性能参数集执行特征提取以获得故障特征参数数据库，所述故障特征参数数据库包括所述至少一种故障以及与每一故障对应的至少一个故障特征参数组，每个故障特征参数组包括关于所述变流器的所述多个功能部件的多个故障特征参数；以及
    基于所述故障特征参数数据库中的所述至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模以获得表现故障与故障特征参数之间的映射关系的故障预警模型。
  2. 如权利要求1所述的故障预警方法，其特征在于，还包括：
    基于对应每种故障的故障程度的故障特征阈值对所述故障特征参数数据库中的故障特征参数进行标定以确定每一故障特征参数组所对应的故障的故障程度，
    所述基于所述故障特征参数数据库中的所述至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模包括：基于所述故障特征参数数据库中的所述至少一种故障的故障状态及与每一故障状态对应的至少一个故障特征参数组执行神经网络建模，所述故障状态包括故障类型和对应每种故障类型的故障程度。
  3. 如权利要求2所述的故障预警方法，其特征在于，所述执行神经网络建模包括使用BP神经网络模型执行建模，所述BP神经网络模型包括输入层、隐含层和输出层，所述输入层包括关于所述故障特征参数组中的多个故障特征参数所对应的多个故障特征的输入节点，所述输出层包括所述至少一种故障状态的输出节点。
  4. 如权利要求3所述的故障预警方法，其特征在于，所述隐含层中每个隐含节点的输出为
    Figure PCTCN2019119160-appb-100001
    x i为所述输入层的输入节点的输入，w ji和θ j分别为每个隐含节点和每个输入节点之间的连接权值和阈值，i为所述输入节点的索引，j为所述隐含节点的索引，
    所述输出层中每个输出节点的输出为
    Figure PCTCN2019119160-appb-100002
    w kj和θ k分别为每个输出节点和每个隐含节点之间的连接权值和阀值，k为所述输出节点的索引。
  5. 如权利要求4所述的故障预警方法，其特征在于，所述BP神经网络的激活函数为Sigmoid函数。
  6. 如权利要求4所述的故障预警方法，其特征在于，所述输出层的输出节点的输出值O k与期望输出值t k之间的误差为
    Figure PCTCN2019119160-appb-100003
    反向传递中的权值修正函数为
    Figure PCTCN2019119160-appb-100004
    阀值修正函数为
    Figure PCTCN2019119160-appb-100005
    其中η取0.01～0.8之间。
  7. 如权利要求6所述的故障预警方法，其特征在于，所述期望输出值包括指示故障状态的故障编号。
  8. 如权利要求1所述的故障预警方法，其特征在于，所述故障的类型包括以下一者或多者：轴承内圈故障、轴承外圈故障、轴承滚珠故障、轴承保持架故障、风机动平衡破坏故障、风机缺相故障、风机匝间短路故障、风机相不平衡故障、风机接地故障、风机喘振故障、电容超温故障、电容容值损失故障、IGBT超温故障、IGBT过流故障、滤网堵塞故障、水泵故障、热交换器故障、水冷板故障、变压器超温故障、变压器绝缘破坏故障。
  9. 如权利要求1所述的故障预警方法，其特征在于，所述故障特征参数对应的故障特征包括以下一者或多者：关于风机振动的总值特征、频谱特征、和包络特征、关于风机电流的功率因素、不平衡系数、负序电流、零序电流、频谱特征、包络特征、和能量特征、关于变压器温度的温度有效值和温度变化梯度、关于模块电流的电流有效值和电流时域特征、关于变流器中间电压的谐波特征、关于IGBT温度的温度有效值和温度变化梯度、关于电容电流的总值特征和频谱特征、关于变流器进出口水温的温度有效值和温度变化梯度、关于变流器进出口水压的水压有效值、关于滤网进出口空气温度的温度有效值。
  10. 如权利要求1所述的故障预警方法，其特征在于，还包括：
    获取所述变流器的所述多个功能部件实测性能参数；
    基于所述实测性能参数计算故障特征参数；以及
    基于所述故障特征参数和所述故障预警模型确定所述变流器的故障。
  11. 如权利要求1所述的故障预警方法，其特征在于，对应每一故障的所述性能参数集是在所述变流器的不同工况下采集的，所述工况包括所述变流器所在的机车的运行环境和里程。
  12. 一种基于多信息融合的变流器的故障预警装置，包括：
    存储器；以及
    处理器，所述处理器配置为：
    建立所述变流器的性能参数数据库，所述性能参数数据库中包括所述变流器在至少一种故障发生时采集的所述变流器的多个功能部件的性能参数集；
    对所述性能参数数据库中的性能参数集执行特征提取以获得故障特征参数数据库，所述故障特征参数数据库包括所述至少一种故障以及与每一故障对应的至少一个故障特征参数组，每个故障特征参数组包括关于所述变流器的所述多个功能部件的多个故障特征参数；以及
    基于所述故障特征参数数据库中的所述至少一种故障及与每一故障对应的至少一个故障特征参数组执行神经网络建模以获得表现故障与故障特征参数之间的映射关系的故障预警模型。
  13. 如权利要求12所述的故障预警装置，其特征在于，所述处理器进一步配置为：
    基于对应每种故障的故障程度的故障特征阈值对所述故障特征参数数据库中的故障特征参数进行标定以确定每一故障特征参数组所对应的故障的故障程度；以及
    基于所述故障特征参数数据库中的所述至少一种故障的故障状态及与每一故障状态对应的至少一个故障特征参数组执行神经网络建模，所述故障状态包括故障类型和对应每种故障类型的故障程度。
  14. 如权利要求13所述的故障预警装置，其特征在于，所述执行神经网络建模包括使用BP神经网络模型执行建模，所述BP神经网络模型包括输入层、隐含层和输出层，所述输入层包括关于所述故障特征参数组中的多个故障特征参数所对应的多个故障特征的输入节点，所述输出层包括所述至少一种故障状态的输出节点。
  15. 如权利要求14所述的故障预警装置，其特征在于，所述隐含层中每个隐含节点的输出为
    Figure PCTCN2019119160-appb-100006
    x i为所述输入层的输入节点的输入，w ji和θ j分别为每个隐含节点和每个输入节点之间的连接权值和阈值，i为所述输入节点的索引，j为所述隐含节点的索引，
    所述输出层中每个输出节点的输出为
    Figure PCTCN2019119160-appb-100007
    w kj和θ k分别为每个输出节点和每个隐含节点之间的连接权值和阀值，k为所述输出节点的索引。
  16. 如权利要求15所述的故障预警装置，其特征在于，所述BP神经网络的激活函数为Sigmoid函数。
  17. 如权利要求15所述的故障预警装置，其特征在于，所述输出层的输出节点的输出值O k与期望输出值t k之间的误差为
    Figure PCTCN2019119160-appb-100008
    反向传递中的权值修正函数为
    Figure PCTCN2019119160-appb-100009
    阀值修正函数为
    Figure PCTCN2019119160-appb-100010
    其中η取0.01～0.8之间。
  18. 如权利要求17所述的故障预警装置，其特征在于，所述期望输出值包括指示故障状态的故障编号。
  19. 如权利要求12所述的故障预警装置，其特征在于，所述故障的类型包括以下一者或多者：轴承内圈故障、轴承外圈故障、轴承滚珠故障、轴承保持架故障、风机动平衡破坏故障、风机缺相故障、风机匝间短路故障、风机相不平衡故障、风机接地故障、风机喘振故障、电容超温故障、电容容值损失故障、IGBT超温故障、IGBT过流故障、滤网堵塞故障、水泵故障、热交换器故障、水冷板故障、变压器超温故障、 变压器绝缘破坏故障。
  20. 如权利要求12所述的故障预警装置，其特征在于，所述故障特征参数对应的故障特征包括以下一者或多者：关于风机振动的总值特征、频谱特征、和包络特征、关于风机电流的功率因素、不平衡系数、负序电流、零序电流、频谱特征、包络特征、和能量特征、关于变压器温度的温度有效值和温度变化梯度、关于模块电流的电流有效值和电流时域特征、关于变流器中间电压的谐波特征、关于IGBT温度的温度有效值和温度变化梯度、关于电容电流的总值特征和频谱特征、关于变流器进出口水温的温度有效值和温度变化梯度、关于变流器进出口水压的水压有效值、关于滤网进出口空气温度的温度有效值。
  21. 如权利要求12所述的故障预警装置，其特征在于，所述处理器还配置为：
    获取所述变流器的所述多个功能部件实测性能参数；
    基于所述实测性能参数计算故障特征参数；以及
    基于所述故障特征参数和所述故障预警模型确定所述变流器的故障。
  22. 如权利要求12所述的故障预警装置，其特征在于，对应每一故障的所述性能参数集是在所述变流器的不同工况下采集的，所述工况包括所述变流器所在的机车的运行环境和里程。
  23. 一种计算机可读介质，其上存储有计算机可执行指令，所述计算机可执行指令在由处理器执行时，实施如权利要求1-11中任一项所述的方法。
PCT/CN2019/119160 2019-11-18 2019-11-18 一种基于多信息融合的变流器的故障预警方法及装置 WO2021097604A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2019/119160 WO2021097604A1 (zh) 2019-11-18 2019-11-18 一种基于多信息融合的变流器的故障预警方法及装置
EP19953096.5A EP4063875A4 (en) 2019-11-18 2019-11-18 INFORMATION COMBINATION METHOD AND DEVICE FOR EARLY WARNING OF FAULTS FOR AN INVERTER

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/119160 WO2021097604A1 (zh) 2019-11-18 2019-11-18 一种基于多信息融合的变流器的故障预警方法及装置

Publications (1)

Publication Number Publication Date
WO2021097604A1 true WO2021097604A1 (zh) 2021-05-27

Family

ID=75979915

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/119160 WO2021097604A1 (zh) 2019-11-18 2019-11-18 一种基于多信息融合的变流器的故障预警方法及装置

Country Status (2)

Country Link
EP (1) EP4063875A4 (zh)
WO (1) WO2021097604A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113686922A (zh) * 2021-06-28 2021-11-23 深圳爱莫科技有限公司 一种反应釜风险在线智能监测装置与方法
CN113777924A (zh) * 2021-09-10 2021-12-10 福州大学 一种接触器的直接吸力闭环控制方法及系统
CN113792487A (zh) * 2021-09-14 2021-12-14 中国铁道科学研究院集团有限公司 动车组牵引系统超温故障预测方法及装置
CN114091296A (zh) * 2021-10-29 2022-02-25 国网冀北电力有限公司电力科学研究院 高压直流断路器的故障预警方法及装置
CN114151293A (zh) * 2021-11-30 2022-03-08 甘肃中电瓜州风力发电有限公司 风机变桨系统的故障预警方法、系统、设备及存储介质
EP4170364A1 (en) * 2021-10-22 2023-04-26 General Electric Renovables España S.L. Power converters
CN116561707A (zh) * 2023-07-11 2023-08-08 南京大全电气研究院有限公司 一种变压器故障校验、预警方法及系统
CN116840671A (zh) * 2023-06-09 2023-10-03 东禾电气有限公司 一种基于智能融合断路器的故障预警与精确定位方法
CN117808456A (zh) * 2023-12-30 2024-04-02 龙坤(无锡)智慧科技有限公司 一种基于智慧运营管理的设备故障预警方法及装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103439658A (zh) * 2013-08-12 2013-12-11 深圳市禾望电气有限公司 一种变流器的故障诊断方法和系统
CN104502758A (zh) * 2014-12-17 2015-04-08 西北工业大学 航空静止变流器故障诊断方法
CN105044497A (zh) * 2015-06-30 2015-11-11 株洲南车时代电气股份有限公司 一种牵引变流器故障智能分析方法
US20180119677A1 (en) * 2015-12-31 2018-05-03 Beijing Goldwind Science & Creation Windpower Equipment Co., Ltd. Computer storage medium, computer program product, method for monitoring fault of wind power generator set, and device
CN109754110A (zh) * 2017-11-03 2019-05-14 株洲中车时代电气股份有限公司 一种牵引变流器故障的预警方法及系统
CN109917200A (zh) * 2017-12-12 2019-06-21 株洲中车时代电气股份有限公司 列车牵引变流器故障诊断方法、系统、设备及存储介质
CN109973331A (zh) * 2019-05-05 2019-07-05 内蒙古工业大学 一种基于bp神经网络的风力发电机组风轮叶片故障诊断算法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109116150A (zh) * 2018-08-03 2019-01-01 福州大学 一种基于小脑模型神经网络的电力电子变换器故障诊断方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103439658A (zh) * 2013-08-12 2013-12-11 深圳市禾望电气有限公司 一种变流器的故障诊断方法和系统
CN104502758A (zh) * 2014-12-17 2015-04-08 西北工业大学 航空静止变流器故障诊断方法
CN105044497A (zh) * 2015-06-30 2015-11-11 株洲南车时代电气股份有限公司 一种牵引变流器故障智能分析方法
US20180119677A1 (en) * 2015-12-31 2018-05-03 Beijing Goldwind Science & Creation Windpower Equipment Co., Ltd. Computer storage medium, computer program product, method for monitoring fault of wind power generator set, and device
CN109754110A (zh) * 2017-11-03 2019-05-14 株洲中车时代电气股份有限公司 一种牵引变流器故障的预警方法及系统
CN109917200A (zh) * 2017-12-12 2019-06-21 株洲中车时代电气股份有限公司 列车牵引变流器故障诊断方法、系统、设备及存储介质
CN109973331A (zh) * 2019-05-05 2019-07-05 内蒙古工业大学 一种基于bp神经网络的风力发电机组风轮叶片故障诊断算法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4063875A4 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113686922A (zh) * 2021-06-28 2021-11-23 深圳爱莫科技有限公司 一种反应釜风险在线智能监测装置与方法
CN113777924A (zh) * 2021-09-10 2021-12-10 福州大学 一种接触器的直接吸力闭环控制方法及系统
CN113777924B (zh) * 2021-09-10 2023-07-07 福州大学 一种接触器的直接吸力闭环控制方法及系统
CN113792487B (zh) * 2021-09-14 2024-02-02 中国国家铁路集团有限公司 动车组牵引系统超温故障预测方法及装置
CN113792487A (zh) * 2021-09-14 2021-12-14 中国铁道科学研究院集团有限公司 动车组牵引系统超温故障预测方法及装置
EP4170364A1 (en) * 2021-10-22 2023-04-26 General Electric Renovables España S.L. Power converters
CN114091296A (zh) * 2021-10-29 2022-02-25 国网冀北电力有限公司电力科学研究院 高压直流断路器的故障预警方法及装置
CN114091296B (zh) * 2021-10-29 2024-05-10 国网冀北电力有限公司电力科学研究院 高压直流断路器的故障预警方法及装置
CN114151293A (zh) * 2021-11-30 2022-03-08 甘肃中电瓜州风力发电有限公司 风机变桨系统的故障预警方法、系统、设备及存储介质
CN114151293B (zh) * 2021-11-30 2022-11-22 甘肃中电瓜州风力发电有限公司 风机变桨系统的故障预警方法、系统、设备及存储介质
CN116840671A (zh) * 2023-06-09 2023-10-03 东禾电气有限公司 一种基于智能融合断路器的故障预警与精确定位方法
CN116840671B (zh) * 2023-06-09 2024-01-19 东禾电气有限公司 一种基于智能融合断路器的故障预警与精确定位方法
CN116561707B (zh) * 2023-07-11 2023-09-05 南京大全电气研究院有限公司 一种变压器故障校验、预警方法及系统
CN116561707A (zh) * 2023-07-11 2023-08-08 南京大全电气研究院有限公司 一种变压器故障校验、预警方法及系统
CN117808456A (zh) * 2023-12-30 2024-04-02 龙坤(无锡)智慧科技有限公司 一种基于智慧运营管理的设备故障预警方法及装置
CN117808456B (zh) * 2023-12-30 2024-06-11 龙坤(无锡)智慧科技有限公司 一种基于智慧运营管理的设备故障预警方法及装置

Also Published As

Publication number Publication date
EP4063875A1 (en) 2022-09-28
EP4063875A4 (en) 2023-08-09

Similar Documents

Publication Publication Date Title
WO2021097604A1 (zh) 一种基于多信息融合的变流器的故障预警方法及装置
CN106199305B (zh) 煤矿井下供电系统用干式变压器绝缘健康状态评估方法
KR102092185B1 (ko) 중전기기 건전성 분석 플랫폼 및 이를 이용하는 분석 방법
CN112782614A (zh) 一种基于多信息融合的变流器的故障预警方法及装置
CN105300692B (zh) 一种基于扩展卡尔曼滤波算法的轴承故障诊断及预测方法
CN105759201A (zh) 基于异常样本识别的高压断路器自诊断方法
JP5494034B2 (ja) 信頼度評価装置、信頼度評価プログラムおよび信頼度評価方法
CN103576050A (zh) 一种电容式电压互感器的运行状态评估方法
CN105003453A (zh) 一种矿井风机在线监测与故障诊断系统
CN104737029A (zh) 用于监视具有周期性行为的电性连接系统的系统及方法
CN105823960A (zh) 一种变压器绕组变形综合诊断方法及系统
CN103884943A (zh) 一种变压器绕组变形综合分析诊断方法
CN112782499B (zh) 一种基于多信息融合的变流器的状态评估方法及装置
CN106441843B (zh) 一种旋转机械故障波形识别方法
CN113902241A (zh) 一种基于综合状态评价的电网设备检修的策略系统及方法
CN110646685A (zh) 一种开关柜状态综合评估方法
CN115877205A (zh) 一种伺服电机的故障智能诊断系统及其诊断方法
AU2023210611A1 (en) Method and apparatus for operating state analysis and early warning of auxiliary device of hydroelectric station, and decision support system for hydroelectric production
CN113742993A (zh) 干式变压器寿命损失预测方法、装置、设备及存储介质
Ginart et al. Automated feature selection for embeddable prognostic and health monitoring (PHM) architectures
CN106651206A (zh) 一种继电保护测试性评估指标体系的评估方法
CN106126949A (zh) 一种汽轮发电机运行状态评估方法
CN114577470A (zh) 用于风机主轴承的故障诊断方法及系统
CN113297033A (zh) 基于云端监测数据的车辆电控系统健康评估方法和系统
CN117611392A (zh) 基于电网设备资源的数字电网综合展示系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19953096

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019953096

Country of ref document: EP

Effective date: 20220620