Sensor abnormality detection method

Abstract

The application provides an anomaly detection method for a sensor used for measuring an electrical parameter, the anomaly detection method comprising the following steps. At a detection time, an output value of the sensor and a true value of the electrical parameter are obtained, the output value is compared with the true value, and a comparison result is output. A first preset number of comparison results are obtained. It is judged whether the first preset number of comparison results meet a first condition, and a first result is output. The following N-1 layer procedures are executed, N>=2, wherein each layer procedure comprises the following steps. A first N-1 result of an Nth preset number is obtained. It is judged whether the first N-1 result of the Nth preset number meets an Nth condition, and an Nth result is output. The working state of the sensor is judged according to the Nth result.

Description

Technical Field

[0001] This case involves sensors, and more particularly a method for detecting anomalies in a sensor. Background Technology

[0002] During product design, testing personnel frequently utilize sensors to detect various data. However, due to limitations in performance or resources of traditional sensors, the data detected may exhibit low refresh rates and data lag in practical applications. Furthermore, the amount of data detected by the sensor itself may fluctuate, making it difficult to automatically detect abnormal states when the sensor itself malfunctions. Traditional sensors address these issues by filtering the detected data before taking the values; however, this method has significant limitations and a relatively high false detection rate.

[0003] Therefore, it is necessary to develop an anomaly detection method for sensors to solve the problems faced by previous technologies. Summary of the Invention

[0004] The purpose of this invention is to provide an anomaly detection method for a sensor. The method utilizes a detection module to acquire a first preset number of comparison results, determines whether these results meet a first condition, and outputs a first result. Then, it acquires an Nth preset number of (N-1)th results, determines whether these Nth preset number of (N-1)th results meet a Nth condition, and outputs the Nth result. Finally, it determines the sensor's operating state based on the Nth result. This anomaly detection method employs a multi-layer detection approach, reducing the impact of feedback lag and instability, and improving the sensor's measurement accuracy. Furthermore, this anomaly detection method allows for flexible combination of detection parameters, increasing its applicability and reducing the detection error rate.

[0005] To achieve the above objectives, this invention provides an anomaly detection method for a sensor used to measure electrical parameters. The anomaly detection method includes the following steps: At the detection time, acquire the sensor's output value and the true value of the electrical parameter, compare the output value with the true value, and output the comparison result. Acquire a first preset number of comparison results. Determine whether the first preset number of comparison results meet a first condition and output the first result. Execute the following N-1 layer program, where N≥2, where each layer program includes the following steps: acquire the Nth preset number of N-1th results. Determine whether the Nth preset number of N-1th results meet the Nth condition and output the Nth result. Determine the sensor's operating state based on the Nth result.

[0006] According to one embodiment of the present invention, the anomaly detection method further includes the step of outputting a comparison result, which includes the following steps: obtaining the absolute value of the difference between the output value and the true value; comparing the absolute value of the difference with a first threshold; if the absolute value of the difference is less than or equal to the first threshold, the comparison result is "true"; if the absolute value of the difference is greater than the first threshold, the comparison result is "false".

[0007] According to one embodiment of the present invention, the anomaly detection method further includes: a first condition being that the proportion of the number of comparison results being "true" in a first preset number of comparisons is greater than or equal to a second threshold; or sorting the first preset number of comparison results such that the number of consecutive comparison results being "true" is greater than or equal to a third threshold.

[0008] According to one embodiment of the present invention, the anomaly detection method further includes a second threshold of 50% and a third threshold of 50% of the first preset number of times.

[0009] According to one embodiment of the present invention, the anomaly detection method further includes the step of outputting a first result, which includes the following steps: If a first preset number of comparison results meet a first condition, the first result is "true"; if a first preset number of comparison results do not meet the first condition, the first result is "false".

[0010] According to one embodiment of the present invention, the anomaly detection method further includes the following: the Nth condition is that the proportion of the number of N-1 results that are "true" in the Nth preset number of times is greater than or equal to the N+2th threshold; or the Nth preset number of N-1 results are sorted, and the number of consecutive N-1 results that are "true" is greater than or equal to the N+3th threshold.

[0011] According to one embodiment of the present invention, the anomaly detection method further includes the step of outputting the Nth result, which includes the following steps: if the (N-1)th preset number of comparison results meet the Nth condition, then the Nth result is "true"; and if the (N-1)th preset number of comparison results do not meet the Nth condition, then the Nth result is "false".

[0012] According to one embodiment of the present invention, the anomaly detection method further includes the step of determining the operating state of the sensor based on the Nth result, which includes the following steps: if the Nth result is "true", then the operating state of the sensor is determined to be normal; and if the Nth result is "false", then the operating state of the sensor is determined to be abnormal.

[0013] According to one embodiment of the present invention, the anomaly detection method further includes that the N-1 layer program runs synchronously. Attached Figure Description

[0014] Figure 1 The circuit topology diagram of the detection system;

[0015] Figure 2 for Figure 1 The flowchart illustrates the steps of the anomaly detection method for the sensors in the detection system shown; and

[0016] Figure 3 for Figure 1 The diagram shows the results corresponding to each layer of the sensor program in the detection system.

[0017] The reference numerals in the attached figures are explained as follows:

[0018] 1: Detection System

[0019] 2: Exchange Source

[0020] 3: Inverter

[0021] 4: Sensors

[0022] 5: Detector

[0023] 51: Detector

[0024] 52: Comparator

[0025] 53: Analyzer

[0026] 6: DC source

[0027] 7: Load

[0028] S1-S4: Steps Detailed Implementation

[0029] Some typical embodiments that embody the features and advantages of this invention will be described in detail in the following description. It should be understood that this invention can have various variations in different forms, all of which do not depart from the scope of this invention, and the descriptions and drawings therein are essentially for illustrative purposes and not for limiting this invention.

[0030] Please see Figure 1 and Figure 2 ,in Figure 1 The circuit topology diagram of the detection system is shown below. Figure 2 This is a flowchart of the steps in the sensor anomaly detection method. For ease of description of the technical content of this case, it is combined with... Figure 1The detection system shown is for illustrative purposes only, but the sensor anomaly detection method provided in this application is not limited to this detection system. Detection system 1 includes an AC power source 2, an inverter 3, a meter, a detector 5, a DC power source 6, and a load 7, wherein the meter can be considered as sensor 4 in this application. The inverter 3 is connected to the AC power source 2 to convert the electrical energy supplied by the AC power source 2. Sensor 4 measures the electrical parameters of the inverter 3 based on currents I1 and I2 and voltages V1 and V2, and provides an output value based on the measured electrical parameters, where the electrical parameters correspond to the internal characteristics of the inverter 3, and the output value is the measured value of the electrical parameter by sensor 4. DC power source 6 is connected to inverter 3, and load 7 is connected to AC power source 2.

[0031] The detection module 5 is connected to the sensor 4 and the inverter 3, and includes a detector 51, a comparator 52, and a judge 53. The detector 51 executes the first-level program, acquiring the output value of the sensor 4 at the detection moment and directly acquiring the true value of the electrical parameters corresponding to the inverter 3. In other words, the true value acquired by the detector 5 is the actual value of the electrical parameters of the inverter 3. The comparator 52 compares the output value provided by the sensor 4 with the true value acquired by the detector 51 multiple times, and provides the comparison results for the corresponding number of comparisons between the output value and the true value. For example, the comparator 52 performs five comparisons between the output value and the true value and provides five comparison results. The judge 53 acquires the multiple comparison results provided by the comparator 52 and further determines whether the multiple comparison results provided by the comparator 52 meet a first condition, and outputs a first result. For example, the judge 53 determines whether the five comparison results provided by the comparator 52 meet the first condition, and outputs the first result. In this embodiment, after the judge 53 outputs the first result, it clears the internal count of five comparisons to prepare for subsequent judgments. When different results are output later, the internal count will also be cleared. This will not be described again.

[0032] Furthermore, the judge 53 executes additional multi-layer procedures. For example, the judge 53 executes the following layer of procedures: The judge 53 repeatedly acquires the first result three times and determines whether the three first results meet the second condition, thereby outputting a second result. The judge 53 determines the operating state of the sensor 4 based on the second result. In summary, the judge 53 of the detection unit 5 executes a total of two layers of procedures and finally determines the operating state of the sensor 4 based on the second result obtained from the second layer of procedures. Of course, in one embodiment, the judge 53 may execute more layers of procedures and determine the operating state of the sensor 4 based on the last layer of procedures. In this embodiment, the time interval between the time when the judge 53 outputs the first result and the time when it outputs the second result can be adjusted as needed.

[0033] like Figure 2As shown, the steps of the anomaly detection method in this case are as follows. First, step S1 is executed: at the detection time, the output value of sensor 4 and the true value of the electrical parameters corresponding to inverter 3 are acquired, the output value and the true value are compared, and the comparison result is output. Next, step S2 is executed: the first preset number of comparison results are acquired, and the Nth preset number of N-1th results are acquired. Next, step S3 is executed: whether the first preset number of comparison results meet the first condition is determined, and the first result is output; whether the Nth preset number of N-1th results meet the Nth condition is determined, and the Nth result is output. Next, step S4 is executed: the working state of sensor 4 is determined based on the Nth result.

[0034] As can be seen from the above, in Figure 1 In the detection system 1 shown, the method for anomaly detection of sensor 4 utilizes detection module 5 to obtain a first preset number of comparison results, determine whether the first preset number of comparison results meet a first condition, and output a first result. Then, it obtains the Nth preset number of (N-1)th results, determines whether the Nth preset number of (N-1)th results meet the Nth condition, and outputs the Nth result. Furthermore, it determines the working state of sensor 4 based on the Nth result. The anomaly detection method in this case employs a multi-layer detection approach, reducing the impact of feedback lag and instability, and improving the measurement accuracy of sensor 4. This anomaly detection method can flexibly combine detection parameters, increasing its applicability and reducing the detection error rate.

[0035] In this embodiment, in the step of comparing the output value provided by the sensor 4 with the real value obtained by the detector 51 to output a comparison result, the comparator 52 can further subtract the output value from the real value and take the absolute value to obtain the absolute value of the difference, and compare the magnitude of the absolute value of the difference with the preset first threshold. If the absolute value of the difference is less than or equal to the first threshold, the comparison result provided by the comparator 52 is displayed as "true", and if the absolute value of the difference is greater than the first threshold, the comparison result provided by the comparator 52 is displayed as "false".

[0036] In this embodiment, the first condition may be whether the proportion of the number of comparison results displayed as "true" in the comparison results provided by comparator 52 is greater than or equal to a second threshold, where the second threshold may be, for example, 50%. Specifically, when the proportion of the number of comparison results displayed as "true" in all comparison results provided by comparator 52 is greater than or equal to the second threshold, the first result output by the judge 53 is "true"; when the proportion of the number of comparison results displayed as "true" is less than the second threshold, the first result output by the judge 53 is "false". Alternatively, the first condition may be that the multiple comparison results provided by comparator 52 are sorted according to the output order, and whether the number of consecutive "true" occurrences is greater than or equal to a third threshold, where the third threshold may be, for example, set to 50% of the number of comparison results provided by comparator 52. Specifically, when the multiple comparison results provided by comparator 52 are sorted according to the output order, if the number of consecutive "true" occurrences is greater than or equal to the third threshold, the first result output by the judge 53 is "true"; if the number of consecutive "true" occurrences is less than the third threshold, the first result output by the judge 53 is "false".

[0037] Similarly, the second condition could be whether the proportion of "true" results in all first results is greater than or equal to a fourth threshold. Specifically, when the proportion of "true" results in all first results is greater than or equal to the fourth threshold, the second result output by the judge 53 is "true," and when the proportion of "true" results in all first results is less than the fourth threshold, the second result output by the judge 53 is "false." Alternatively, the second condition could be whether the number of consecutive "true" results is greater than or equal to a fifth threshold when all first results are sorted in output order. Specifically, when the number of consecutive "true" results is greater than or equal to the fifth threshold when all first results are sorted in output order, the second result output by the judge 53 is "true," and when the number of consecutive "true" results is less than the fifth threshold, the second result output by the judge 53 is "false."

[0038] The judgment unit 53 of the detection unit 5 further determines that the working state of the sensor 4 is normal when the second result is displayed as "true", and determines that the working state of the sensor 4 is abnormal when the second result is displayed as "false".

[0039] In the above embodiment, the anomaly detection method contains only two layers of programs. When the number of program layers executed is N, the detection unit 5 determines the working state of the sensor 4 based on the Nth result. In the Nth layer of the program (N≥2), the Nth preset number of N-1th results are obtained. The judge 53 then determines whether the Nth preset number of N-1th results meet an Nth condition and outputs the Nth result. When the Nth result is "true", the working state of the sensor 4 is considered normal; when the Nth result is "false", the working state of the sensor 4 is considered abnormal. In the specific execution process, furthermore, the Nth layer or the (N-1)th layer of programs other than the first layer can run synchronously.

[0040] To provide a clearer picture of the beneficial effects of this case on reducing the probability of sensor false detections, please refer to [link / reference]. Figure 3 , it is Figure 1 The diagram illustrates the results corresponding to each layer of the sensor program in the detection system. For the first layer, n1 comparison results are acquired. The first condition is that when the number of "true" results among the n1 comparison results is greater than or equal to m1, the first result output by the first layer is "true". For the second layer, n2 first results are acquired. The second condition is that when the number of "true" results among the n2 first results is greater than or equal to m2, the second result output by the second layer is "true". And so on. For the Nth layer, n(N) (N-1)th results are acquired. The Nth condition is that when the number of "true" results among the n(N) (N-1)th results is greater than or equal to m(N), the Nth result output by the Nth layer is "true". In this embodiment, assuming the probability of each comparison result being "true" is a, the error detection probability of the Nth result is (1-a). n1m1*n2m2*…n(N)m(N) In other words, the error detection probability is negatively correlated with the number of execution layers of the program, the number of results obtained from the previous program in each layer, and the judgment conditions set in each layer. The more execution layers, the greater the number of results obtained from the previous program in each layer, and the more stringent the judgment conditions, the lower the error detection probability. Therefore, the sensor anomaly detection method provided in this case can effectively reduce the error detection probability of the sensor and improve the utilization rate of the sensor by executing multi-layer detection programs on the sensor.

[0041] In summary, the detection system in this case utilizes a detection module to acquire a first preset number of comparison results, determines whether these results meet a first condition, and outputs a first result. Then, it acquires the Nth preset number of (N-1)th results, determines whether these results meet the Nth condition, and outputs the Nth result. Finally, it determines the sensor's operating status based on the Nth result. This anomaly detection method employs a multi-layer detection approach, reducing the impact of feedback lag and instability, and improving the sensor's measurement accuracy. Furthermore, this anomaly detection method allows for flexible combination of detection parameters, increasing its applicability and reducing the detection error rate.

Claims

1. An anomaly detection method for a sensor, said sensor being used to measure an electrical parameter, characterized in that, include: At a detection moment, an output value of the sensor and a true value of the electrical parameter are acquired, the output value and the true value are compared, and a comparison result is output. Obtain the first preset number of comparison results; Determine whether the first preset number of comparison results meet a first condition, and output a first result; Execute the following N-1 level program, where N≥2, where each level of the program includes: Get the Nth preset number of results; Determine whether the Nth preset number of N-1th results meets an Nth condition, and output an Nth result; and The sensor's operating state is determined based on the Nth result.

2. The anomaly detection method according to claim 1, characterized in that, The steps for outputting the comparison results include: Obtain the absolute value of the difference between the output value and the true value; Compare the absolute value of the difference with the magnitude of a first threshold; If the absolute value of the difference is less than or equal to the first threshold, then the comparison result is "true"; and If the absolute value of the difference is greater than the first threshold, then the comparison result is "false".

3. The anomaly detection method according to claim 2, characterized in that, The first condition is: The proportion of the number of comparison results that are "true" in the first preset number of times is greater than or equal to a second threshold; or The comparison results are sorted according to a first preset number of times, and the number of consecutive comparison results of "true" is greater than or equal to a third threshold.

4. The anomaly detection method according to claim 3, characterized in that, The second threshold is 50%; the third threshold is 50% of the first preset number of times.

5. The anomaly detection method according to claim 1, characterized in that, The steps for outputting the first result include: If the first preset number of comparison results meet the first condition, then the first result is "true"; and If the first preset number of comparison results do not meet the first condition, then the first result is "false".

6. The anomaly detection method according to claim 5, characterized in that, The Nth condition is: The proportion of the number of "true" results in the N-1th preset number of times is greater than or equal to a threshold number (N+2th). or Sort the Nth preset number of N-1th results, and the number of consecutive N-1th results that are "true" is greater than or equal to an N+3th threshold.

7. The anomaly detection method according to claim 1, characterized in that, The steps for outputting the Nth result include: If the (N-1)th preset number of comparison results meets the Nth condition, then the Nth result is "true"; and If the (N-1)th preset number of comparison results does not meet the Nth condition, then the Nth result is "false".

8. The anomaly detection method according to claim 7, characterized in that, The step of determining the working state of the sensor based on the Nth result includes: If the Nth result is "true", then the sensor is considered to be in normal working condition; and If the Nth result is "false", then the working state of the sensor is considered abnormal.

9. The anomaly detection method according to claim 1, characterized in that, The N-1 layer program runs synchronously.

Images