CN-121578217-B - Sensor fault processing method, device and medium

CN121578217BCN 121578217 BCN121578217 BCN 121578217BCN-121578217-B

Abstract

The application discloses a sensor fault processing method, a sensor fault processing device and a sensor fault processing medium, and relates to the technical field of rail transit auxiliary converter power supply. And constructing a state estimation equation according to the sampling data and the position of the corresponding sensor in the auxiliary converter to obtain a state estimation value. And establishing a coupling relation among the sensors in the auxiliary converter according to each state estimation equation so as to determine the current reference sensor, and determining the state of the reference sensor by decoupling in combination with the operation of the coupling relation among the sensors at different positions of the multilevel converter in the auxiliary converter, so that the subsequent sequential diagnosis is facilitated. And determining fault processing results of the sensors according to the respective state estimation values of the sampling data of the current reference sensor and the adjacent sensors, and the sampling data of the other sensors except the current reference sensor and the respective state estimation values of the corresponding adjacent sensors, so as to realize accurate positioning of multiple sensors.

Inventors

ZENG MINGGAO
YI YU
LIU FULIN
LI YUN
XIA YUQUAN
HOU ZHAOWEN
ZHAO XUFENG
ZHANG YAN
YUAN FANG
DING LEILEI

Assignees

株洲中车时代电气股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260120

Claims (12)

1. A method of sensor fault handling, comprising: Acquiring sampling data of each sensor of the auxiliary converter; constructing a state estimation equation according to the sampling data and the position of the corresponding sensor in the auxiliary converter to obtain a state estimation value; establishing a coupling relation among the sensors in the auxiliary converter according to each state estimation equation so as to determine a current reference sensor; in the coupling relation, determining fault processing results of all sensors according to respective corresponding state estimation values of the sampling data of the current reference sensor and the adjacent sensors, sampling data of other sensors except the current reference sensor and respective corresponding state estimation values of the adjacent sensors so as to complete fault positioning processing; Correspondingly, the sampling data at least comprises input voltage, input current, boost voltage, intermediate voltage, first bridge arm current, second bridge arm current, first output voltage and second output voltage, and a state estimation equation is built at the position of the auxiliary converter according to the sampling data and the corresponding sensor to obtain a state estimation value, wherein the method comprises the following steps: When the boost converter of the auxiliary converter enters a continuous state, a state estimation equation is constructed on the input voltage and the boost voltage through the duty ratio of a switching device in the auxiliary converter to obtain a corresponding first state estimation value and a corresponding second state estimation value; when the isolation converter of the auxiliary converter enters a constant gain state, a state estimation equation is constructed for the boosted voltage and the intermediate voltage through the gain of the transformer, so that a corresponding third state estimation value and a corresponding fourth state estimation value are obtained; After the three-phase inverter of the auxiliary converter is started, a state estimation equation is constructed by modulating and comparing the first output voltage, the second output voltage and the intermediate voltage to obtain a corresponding fifth state estimation value, a sixth state estimation value and a seventh state estimation value; When the output end of the auxiliary converter is not loaded, a state estimation equation of the first bridge arm current and the second bridge arm current is constructed through the first output voltage, the second output voltage and the capacitor, so that an eighth state estimation value and a ninth state estimation value which are respectively corresponding to each other are obtained; After the auxiliary converter is started, a current state estimation equation is constructed for the input voltage, the first output voltage, the second output voltage, the first bridge arm current and the second bridge arm current through calibrating the operation efficiency, so as to obtain a tenth state estimation value; Correspondingly, establishing a coupling relation between the sensors in the auxiliary converter according to the state estimation equations comprises the following steps: Establishing a first coupling relation between the input voltage and the boost voltage according to the duty ratio of a switching device in the auxiliary converter; establishing a second coupling relationship between the boost voltage and the intermediate voltage according to the transformer gain; Establishing a third coupling relationship among the first output voltage, the second output voltage and the intermediate voltage according to the modulation ratio; establishing a fourth coupling relation among the first output voltage, the second output voltage, the first bridge arm current and the second bridge arm current according to the circuit relation of the auxiliary converter; Establishing a fifth coupling relation among the input voltage, the first output voltage, the second output voltage, the first bridge arm current and the second bridge arm current according to the calibration operation efficiency; determining a final coupling relationship according to the first coupling relationship, the second coupling relationship, the third coupling relationship, the fourth coupling relationship and the fifth coupling relationship; correspondingly, determining the current reference sensor according to the coupling relation comprises: screening adjacent sensors existing before and after sampling data in the coupling relation to sample, and taking the sensor corresponding to the current sampling data sampled by the adjacent sensors existing before and after as an initial reference sensor; Screening the reference sensors corresponding to the calculation of the mutual coupling relation between the front and the rear of the sampling data of the front and the rear sensors in each initial reference sensor as a final reference sensor; one reference sensor is randomly selected from the final reference sensors as a current reference sensor, and the rest sensors except the current reference sensor are used as non-reference sensors.
2. The sensor failure processing method according to claim 1, wherein determining the failure processing result of each sensor based on the respective state estimation values of the sampling data of the current reference sensor and the neighboring sensors, the sampling data of the remaining sensors other than the current reference sensor, and the respective state estimation values of the corresponding neighboring sensors, comprises: Comparing the sampling data of the current reference sensor with the respective state estimation values of the front and rear adjacent sensors to determine a fault processing result of the current reference sensor; When the fault processing result of the current reference sensor is normal, determining a sensor to be diagnosed after the current reference sensor according to the coupling relation, taking the sensor to be diagnosed as the current sensor to be diagnosed, and comparing the sampling data of the current sensor to be diagnosed with the state estimation value corresponding to one of the adjacent sensors to determine the fault processing result of the current sensor to be diagnosed; and when the fault processing result of the current sensor to be diagnosed is normal, determining a next-stage sensor of the current sensor to be diagnosed as a new current sensor to be diagnosed according to the coupling relation, and returning to the step of comparing and determining the fault processing result of the current sensor to be diagnosed according to the sampling data of the current sensor to be diagnosed and the state estimation value corresponding to one of the corresponding adjacent sensors until the fault processing of all other sensors is completed.
3. The sensor fault handling method according to claim 2, wherein determining the fault handling result of the current reference sensor based on comparing the sampled data of the current reference sensor with the respective state estimation values of the front and rear adjacent sensors, comprises: Comparing the sampling data of the current reference sensor with the respective state estimation values of the front and rear adjacent sensors respectively; If the difference value between the sampling data of at least one current reference sensor and the state estimation values of the front and rear adjacent sensors is smaller than a first threshold value, determining that the fault processing result of the current reference sensor is normal; and if the difference value between the sampling data of the current reference sensor and the state estimation values of the front and rear adjacent sensors is larger than or equal to a first threshold value, determining that the fault processing result of the current reference sensor is a fault.
4. The sensor fault handling method according to claim 2, wherein comparing the sampled data of the current sensor to be diagnosed with the state estimation value corresponding to one of the adjacent sensors to determine the fault handling result of the current sensor to be diagnosed, comprises: if the difference value between the sampling data of the current sensor to be diagnosed and the state estimation value corresponding to one of the adjacent sensors is smaller than a second threshold value, determining that the fault processing result of the current sensor to be diagnosed is normal; and if the difference value between the sampling data of the current sensor to be diagnosed and the state estimation value corresponding to one of the adjacent sensors is larger than or equal to a second threshold value, determining that the fault processing result of the current sensor to be diagnosed is a fault.
5. A sensor fault handling method according to claim 3, wherein upon determining that the fault handling result of the current reference sensor is a fault, the method further comprises: Comparing the sampling data of the current reference sensor with the state estimation value of the front adjacent sensor to obtain a first difference value; Comparing the sampling data of the current reference sensor with the state estimation value of the next adjacent sensor to obtain a second difference value; if the first difference value and the second difference value are smaller than the first threshold value, determining that the fault processing results of the front and rear adjacent sensors are normal, and taking the state estimated values corresponding to the front and rear adjacent sensors as the replacement sampling data of the current reference sensor so as to perform redundancy work.
6. The sensor fault handling method of claim 5, wherein if the first difference and the second difference are greater than or equal to the first threshold, the method further comprises: and if the difference value of the state estimated values of the front and rear adjacent sensors is larger than a third threshold value, determining that the fault processing result of the front and rear adjacent sensors is a fault, and stopping the protection processing.
7. The sensor failure processing method according to claim 5, wherein taking the state estimation values of the front and rear adjacent sensors as the replacement sampling data of the current reference sensor, comprises: Presetting a first weight parameter and a second weight parameter of state estimation values corresponding to the front and rear adjacent sensors, wherein the second weight parameter is larger than the first weight parameter; multiplying the first weight parameter with a state estimation value corresponding to a front adjacent sensor to obtain first sampling data; multiplying the second weight parameter with a state estimation value corresponding to a rear adjacent sensor to obtain second sampling data; And adding the first sampling data and the second sampling data to obtain the replacement sampling data of the current reference sensor.
8. The sensor fault handling method of claim 4, wherein when the fault handling result of the current sensor to be diagnosed is a fault, the method further comprises: And taking the state estimation value corresponding to one of the adjacent sensors corresponding to the current sensor to be diagnosed as the replacement sampling data of the current sensor to be diagnosed so as to perform redundancy work.
9. The sensor fault handling method according to claim 5 or 8, wherein during sensor redundancy operation processing, the method further comprises: and if the other sensor of the target sensor with redundant working processing fails, stopping the auxiliary converter for protection.
10. The sensor fault handling method of claim 1, wherein the acquiring of the state estimate is performed by a processor, the method further comprising: Acquiring historical sampling data; Performing time sequence feature processing according to the historical sampling data to obtain feature parameters; The method comprises the steps of recording the calling times, calling an artificial intelligent model, inputting the characteristic parameters into the artificial intelligent model to output a current state estimation value, adding 1 to the calling times when the current state estimation value is not in a first preset range, returning to the step of calling the artificial intelligent model for training until the current state estimation value is in the first preset range or reaches the preset calling times when the calling times are not reached, and completing the training process of the artificial intelligent model; Correspondingly, performing time sequence feature processing according to the sampling data to obtain current feature parameters, and calling the artificial model which is completed with training to input the current feature parameters into the artificial model which is completed with training so as to output a final state estimation value.
11. A sensor fault handling device, comprising: A memory for storing a computer program; a processor for implementing the steps of the sensor fault handling method according to any one of claims 1 to 10 when executing the computer program.
12. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the sensor fault handling method according to any of claims 1 to 10.

Description

Sensor fault processing method, device and medium Technical Field The application relates to the technical field of power supply of rail transit auxiliary converters, in particular to a sensor fault processing method, a sensor fault processing device and a sensor fault processing medium. Background The current transformer is used as an interface between a power grid and vehicle-mounted equipment, the sensor is a key component for normal and reliable operation of the current transformer, a set of auxiliary current transformer system generally comprises a plurality of even tens of sensors such as current, voltage and the like, the auxiliary current transformer generally has a multi-stage structure, the number of the sensors is large, the acquisition states are mutually coupled, the current commonly used sensor diagnosis is mostly based on a method for mutually verifying adjacent sensors, namely, the estimated value of the sensor B is calculated through the sampling value of the sensor A and is compared with the sampling value of the sensor B, whether the sensor B fails or not is judged according to the deviation grade, however, the state estimation is mutually synchronous, and the sensor A is diagnosed by the sampling value of the sensor B. Therefore, when the sensor B fails, the failures of the sensors a and B may be reported at the same time, so that it is difficult to locate the specific sensor failure, and the reliability of the auxiliary converter is reduced. Therefore, how to achieve accurate positioning of sensor faults in auxiliary converters is a need for a person skilled in the art. Disclosure of Invention The application aims to provide a sensor fault processing method, a device and a medium, which are used for solving the problems that positioning is difficult and reliability of an auxiliary converter is reduced due to the adoption of double-sensor mutual detection in a conventional scheme. In order to solve the above technical problems, the present application provides a sensor fault processing method, including: Acquiring sampling data of each sensor of the auxiliary converter; constructing a state estimation equation according to the sampling data and the position of the corresponding sensor in the auxiliary converter to obtain a state estimation value; establishing a coupling relation among the sensors in the auxiliary converter according to each state estimation equation so as to determine a current reference sensor; in the coupling relation, determining fault processing results of the sensors according to the respective state estimation values of the sampling data of the current reference sensor and the adjacent sensors, and the sampling data of the other sensors except the current reference sensor and the respective state estimation values of the corresponding adjacent sensors so as to complete fault positioning processing. On the one hand, the sampling data at least comprises an input voltage, an input current, a boost voltage, an intermediate voltage, a first bridge arm current, a second bridge arm current, a first output voltage and a second output voltage, and a state estimation equation is constructed at the position of the auxiliary converter according to each sampling data and the corresponding sensor to obtain a state estimation value, and the method comprises the following steps: When the boost converter of the auxiliary converter enters a continuous state, a state estimation equation is constructed on the input voltage and the boost voltage through the duty ratio of a switching device in the auxiliary converter to obtain a corresponding first state estimation value and a corresponding second state estimation value; when the isolation converter of the auxiliary converter enters a constant gain state, a state estimation equation is constructed for the boosted voltage and the intermediate voltage through the gain of the transformer, so that a corresponding third state estimation value and a corresponding fourth state estimation value are obtained; After the three-phase inverter of the auxiliary converter is started, a state estimation equation is constructed by modulating and comparing the first output voltage, the second output voltage and the intermediate voltage to obtain a corresponding fifth state estimation value, a sixth state estimation value and a seventh state estimation value; When the output end of the auxiliary converter is not loaded, a state estimation equation of the first bridge arm current and the second bridge arm current is constructed through the first output voltage, the second output voltage and the capacitor, so that an eighth state estimation value and a ninth state estimation value which are respectively corresponding to each other are obtained; After the auxiliary converter is started, a current state estimation equation is constructed for the input voltage, the first output voltage, the second output voltage, the first bridge arm current and the second bri