A pressure switch fatigue test system and method based on data analysis

Abstract

The present invention relates to the technical field of pressure switches, and in particular to a pressure switch fatigue test system and method based on data analysis. When the system receives abnormal data, the present invention can mark the pressure sensor that transmits the abnormal data, set a coverage range, extract the pressure sensors within the coverage range whose correlation reaches a set value, and verify whether there are other reasons for the pressure sensor that transmits the abnormal data by analyzing the changes in the collected data of the pressure sensors whose correlation reaches the set value. Thus, the abnormal data transmitted can be verified, the credibility of the abnormal data can be improved, and at the same time, the maintenance time period of the marked pressure sensor can be predicted, further determining whether the pressure sensor that transmits the abnormal data is caused by its own reasons or damaged by others, reducing the resource consumption caused by calculating unreliable data during the test process.

Description

Technical Field

[0001] The present invention relates to the technical field of pressure switches, and specifically to a pressure switch fatigue test system and method based on data analysis. Background Technique

[0002] A pressure switch includes a pressure sensor and an electrical switch. During use, it is necessary to conduct a fatigue test on the pressure switch to ensure its quality. During use, multiple environmental data of multiple pressure sensors are collected, and then the data is transmitted to a console for fatigue test analysis.

[0003] Multiple pressure sensors are set, and after uploading the data collected by each pressure sensor to the cloud platform, data for judging the status is obtained. When a certain pressure sensor uploads abnormal data, the system will automatically issue a corresponding instruction according to the settings and send an alarm message. However, the abnormal data transmitted by the pressure sensor may be abnormal data collected due to an abnormal monitored environment, and in this case, the abnormal data is correct data; but there may also be a situation where the collected data is abnormal due to problems with the pressure sensor itself, or data abnormalities caused by other reasons. At this time, the "abnormal data" is incorrect. Since the system cannot screen the authenticity of this "abnormal data", transmitting incorrect "abnormal data" will cause the system to issue incorrect instructions and make incorrect responses, resulting in a reduction in the accuracy of the test system and a great waste of the operating resources of the test system. Summary of the Invention

[0004] The purpose of the present invention is to provide a pressure switch fatigue test system and method based on data analysis to solve the problems raised in the above background technique.

[0005] To solve the above technical problems, the present invention provides the following technical solution: A pressure switch fatigue test method based on data analysis, characterized in that: the method includes the following specific steps:

[0006] S1. Obtain the monitoring data of the perception layer and extract the monitored abnormal data; the perception layer includes several pressure sensors, the pressure sensors upload the monitored data information, mark the pressure sensors that transmit abnormal data, extract the information of the marked pressure sensors, and set the time node when abnormal data is transmitted as the abnormal time point;

[0007] S2. Taking the marked pressure sensor as a reference point, set the coverage range, obtain the information of all pressure sensors within the coverage range, calculate the correlation between each pressure sensor within the coverage range and the marked pressure sensor; obtain the information of the pressure sensors whose correlation reaches the set value, and extract the data transmitted by the pressure sensors whose correlation reaches the set value at the abnormal time point, and use this data as the actual collected value;

[0008] S3. Establish a training model, use the historical collected data of the marked pressure sensor and the historical collected data of the pressure sensors whose correlation reaches the set value as the training set, train the training model, and through the training model, calculate the calculated collected data of the pressure sensors whose correlation reaches the set value at the abnormal time point, and use this data as the output calculated value; compare whether the actual collected data is the same as the output calculated value to obtain a matching result; calculate the matching rate according to the matching result, and judge the trust result of the abnormal data according to the matching rate;

[0009] S4. If the trust result output by S3 is "high credibility", the system automatically issues the system instruction corresponding to this abnormal data according to the setting; if the trust result output by S3 is "general credibility", set an expanded range, and the expanded range is different from the coverage range, and repeat S2; if the trust result output by S3 is "low credibility", extract the historical maintenance information of the marked pressure sensor, predict the maintenance time period of the marked pressure sensor, judge whether the marked pressure sensor is in the maintenance time period, and send information to the console;

[0010] S5. The test analysis module analyzes the monitoring data processed by the data processing module, and then evaluates and analyzes the test performance of the pressure switch; the optimization analysis module analyzes the process optimization direction of the tested pressure switch and obtains the time influence coefficient and the pressure influence coefficient, compares the time influence coefficient and the pressure influence coefficient, and generates a corresponding process optimization signal through the comparison result.

[0011] The trust results in S300 include "high credibility", "general credibility" and "low credibility".

[0012] The S1 includes:

[0013] S101. When abnormal data appears in the perception layer, capture the abnormal data and mark the pressure sensor that transmits the abnormal data. Mark the pressure sensor that transmits the abnormal data as the M pressure sensor, and establish a three-dimensional coordinate system with the marked M pressure sensor as the origin;

[0014] S102. Obtain the type of the M pressure sensor, get the type of data collected by the M pressure sensor, and when the M pressure sensor transmits abnormal data, set the time node of transmitting the abnormal data as the abnormal time point Tx.

[0015] The said S2 includes:

[0016] S201. Extract all pressure sensors to form a set B, where B = {pressure sensor 1, pressure sensor 2, pressure sensor 3,..., pressure sensor X}; S202. Extract the names of the pressure sensor M and the pressure sensor X in the set B. Let the name of the pressure sensor M be W1, and the name of the pressure sensor X in the set B be W2. Use the following formula to calculate the similarity of the names of each pressure sensor in the set B and the pressure sensor M :

[0017]

[0018] Where, represents the similarity of the name of the pressure sensor M and the pressure sensor X in the set B; m represents the total number of characters in W1; n represents the total number of characters in W2; c is the number of matching characters between W1 and W2; L1(i) and L2(i) are the matching orders of the matching character i in W1 and W2 respectively; and are the weights of the matching degree and the matching order respectively. Usually, according to the golden ratio, they are taken as 0.6 and 0.4 respectively; t represents the position index of each character in W1; k represents the position index of each character in W2;

[0019] S203. According to the usage standard of the pressure sensor M, obtain the monitoring range of the pressure sensor M, extract all the coordinate points within this monitoring range to form a set A, where A = {(x1, y1, z1), (x2, y2, z2)......(x n , y n , z n )}; According to the usage standard of the pressure sensor X, obtain the monitoring range of the pressure sensor X, extract all the coordinate points within this monitoring range to form a set C, where C = {(a1, b1, c1), (a2, b2, c2)......(a n , b n , c n )}, extract the overlapping coordinate points between the set C and the set A to form a set D, and calculate the overlapping rate of the set A and the set C according to the following formula :

[0020]

[0021] S204. Set the correlation between pressure sensor X and pressure sensor M as R X , and calculate the correlation R between pressure sensor M and pressure sensor X in set B according to the following formula X :

[0022] R X = + ;

[0023] Arrange the pressure sensors X in set B in descending order according to the correlation R X , set the number of verification values as Z, extract Z pressure sensors X according to the descending order of the correlation, and use the data transmitted by the pressure sensor X at the abnormal time point Tx as the actual acquisition value.

[0024] The said S3 includes:

[0025] The said S301. Establish a historical time set T, extract the transmission data of pressure sensor M corresponding to the time node t in the historical time set T n , and at the time node t n , the transmission data of the pressure sensor X extracted in S204, establish a training model, use the transmission data transmitted by the pressure sensor M as the input, and use the transmission data transmitted by the pressure sensor X as the output to train the model;

[0026] The said S302. According to the training model in S301, calculate the calculated acquisition data of pressure sensor X at the abnormal time node Tx, and use this data as the output calculation value;

[0027] S303. Set the error range of the output calculation value, compare whether the actual acquisition value and the output calculation value are within the error range; if the data of the actual acquisition value and the output calculation value are within the error range, output the result of "matched"; if the data of the actual acquisition value and the output calculation value are not within the error range, output the result of "not matched";

[0028] S304. According to the matching results of the Z pressure sensors X obtained, calculate the matching rate, and output the trust result according to the matching rate. The trust result is "high credibility", "general credibility" or "low credibility"; The calculation formula of the matching rate is as follows:

[0029]

[0030] where Z represents the total number of pressure sensors X extracted in S204.

[0031] The said S4 includes:

[0032] S401. If the trust result of "high credibility" is output in S304, it indicates that the abnormal data transmitted this time is highly accurate, and the system automatically issues a pre-set instruction.

[0033] S402. If the trust result of "medium credibility" is output in S304, a new expanded range is set again, and all pressure sensors within the expanded range are extracted to form a set F, where F = {pressure sensor 1, pressure sensor 2, pressure sensor 3,..., pressure sensor Y}. Repeat the calculation of the correlation between each pressure sensor within the coverage range and the marked pressure sensor in step S2, and output the trust result for judging the abnormal data again.

[0034] S403. If the trust result of "low credibility" is output in S304, the historical maintenance information of pressure sensor M is extracted, the maintenance time period of pressure sensor M is predicted, and it is judged whether the abnormal time point Tx is within the maintenance time period of pressure sensor M. If the abnormal time point Tx is within the maintenance time period of pressure sensor M, it is output that the natural failure probability of pressure sensor M is greater than the probability of human damage. If the abnormal time point Tx is not within the maintenance time period of pressure sensor M, it is output that the natural failure probability of pressure sensor M is less than the probability of human damage. The prediction information is sent to the console.

[0035] A pressure switch fatigue test system based on data analysis, which includes a monitoring module, a transmission module, a data processing module, a data extraction module, a data verification module, a test analysis module, and an optimization analysis module connected to each other.

[0036] The monitoring module is used to collect the environmental data of the pressure sensors in the monitored pressure switch, and transmit the collected environmental data to the data processing module through the transmission module. The data processing module screens the monitored data. When abnormal data appears, the data extraction module extracts the abnormal data, marks the pressure sensor that transmits the abnormal data, and then marks the corresponding pressure switch of the pressure sensor. The data verification module verifies the credibility of the abnormal data and outputs a corresponding instruction according to the verification result.

[0037] The monitoring module includes a pressure sensor unit for real-time monitoring of data in the environment. The transmission module includes a data transmission unit and an instruction output unit. The data transmission unit is used for real-time transmission of the monitored data, and the instruction output unit is used to send a monitoring instruction to the pressure sensor unit according to the system-set instruction.

[0038] The data processing module includes a data preprocessing unit and a classification storage unit. The data preprocessing unit is used to denoise the monitored data. After removing redundant data, the classification storage unit is used to classify and store the monitored data.

[0039] The data extraction module includes an extraction unit, a marking unit, and a modeling unit. The extraction unit is used to extract the transmitted abnormal data, and the marking unit is used to mark the pressure sensor that transmits the abnormal data. At the same time, through the modeling unit, with the marked pressure sensor as the reference point, a three-dimensional coordinate system is established, the coverage range is set, and the positions of the pressure sensors within the coverage range are marked.

[0040] The data verification module includes a correlation calculation unit, a model establishment unit, a comparison unit, and a verification information output unit. The correlation calculation unit is used to calculate the correlation between the marked pressure sensor and the pressure sensors within the coverage range, and arrange the pressure sensors within the coverage range in descending order according to the correlation, and extract the pressure sensors whose correlation reaches the set threshold. The model establishment unit is used to establish a training module, use the historical acquisition data of the marked pressure sensor as the input value, and use the historical acquisition data of the pressure sensors whose correlation reaches the threshold as the output value to train the model. Finally, calculate the output calculation value of the sensor whose correlation reaches the threshold at the abnormal time point Tx. The comparison unit is used to compare the output calculation value with the actual acquisition value, output the matching result, and calculate the matching rate, and judge the trust result of the abnormal data according to the matching rate. The verification information output unit issues the next operation corresponding to the trust result according to the output information result.

[0041] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0042] When the system receives abnormal data, the present invention can mark the pressure sensor that transmits the abnormal data, set a coverage range, extract the pressure sensors within the coverage range whose correlation reaches the set value, and verify whether there are other reasons for the pressure sensor that transmits the abnormal data by analyzing the change of the acquisition data of the pressure sensors whose correlation reaches the set value. Thus, the abnormal data transmitted can be verified, the credibility of the abnormal data can be improved, the number of incorrect instructions issued by the system can be reduced, and resources can be saved. At the same time, the maintenance time period of the marked pressure sensor can be predicted, and further judge whether the pressure sensor that transmits the abnormal data is caused by its own reason or damaged by others, reducing the resource consumption caused by calculating unreliable data during the test process. Description of the Drawings

[0043] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the accompanying drawings:

[0044] Figure 1 is a schematic flowchart of the test method in the present invention. Specific embodiments

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0046] Embodiment 1:

[0047] Please refer to Figure 1 , the present invention provides a technical solution: a pressure switch fatigue test method based on data analysis, and the method includes the following specific steps:

[0048] S1. Obtain the monitoring data of the perception layer and extract the abnormal data detected; the perception layer includes several pressure sensors, the pressure sensors upload the detected data information, mark the pressure sensors that transmit abnormal data, extract the information of the marked pressure sensors, and set the time node of transmitting abnormal data as the abnormal time point;

[0049] S1 includes:

[0050] S101. When abnormal data appears in the perception layer, capture the abnormal data and mark the pressure sensor that transmits the abnormal data. Mark the pressure sensor that transmits the abnormal data as the M pressure sensor. Taking the marked M pressure sensor as the origin, establish a three-dimensional coordinate system;

[0051] S102. Obtain the type of the pressure sensor of the M pressure sensor, obtain the data type collected by the M pressure sensor, and set the time node of transmitting abnormal data as the abnormal time point Tx when the M pressure sensor transmits abnormal data.

[0052] S2. Taking the marked pressure sensor as the reference point, set the coverage range, obtain the information of all pressure sensors within the coverage range, and calculate the correlation between each pressure sensor within the coverage range and the marked pressure sensor; obtain the information of the pressure sensors whose correlation reaches the set value, and extract the data transmitted by the pressure sensors whose correlation reaches the set value at the abnormal time point, and use this data as the actual collected value;

[0053] S2 includes:

[0054] S201. Extract all pressure sensors to form set B, where set B = {pressure sensor 1, pressure sensor 2, pressure sensor 3,..., pressure sensor X}; among them, pressure sensor M is located within the set coverage range.

[0055] S202. Extract the names of pressure sensor M and pressure sensor X in set B. Let the name of pressure sensor M be W1 and the name of pressure sensor X in set B be W2. Use the following formula to calculate the similarity of each pressure sensor in set B to pressure sensor M in terms of text names :

[0056]

[0057] Among them, represents the similarity of pressure sensor M to pressure sensor X in set B in terms of name; m represents the total number of characters in W1; n represents the total number of characters in W2; c is the number of matching characters between W1 and W2; L1(i) and L2(i) are the matching orders of the matching character i in W1 and W2 respectively; and are the weights of the matching degree and the matching order respectively. Usually, according to the golden ratio, they are taken as 0.6 and 0.4 respectively; t represents the position index of each character in W1; k represents the position index of each character in W2;

[0058] For example, for pressure sensor M; then the matching characters of W1 and W2 are "pressure", "transmission", "sensing", "device"; the matching orders in W1 and W2 are 2 (pressure), 3 (transmission), 4 (sensing), 5 (device) and 2 (pressure), 3 (transmission), 4 (sensing), 5 (device) respectively; then The calculation process of

[0059] = 2.473

[0060] S203. According to the usage standard of pressure sensor M, obtain the monitoring range of pressure sensor M, extract all coordinate points within this monitoring range to form set A, where set A = {(x1, y1, z1), (x2, y2, z2)......(x n , y n , z n )}; According to the usage standard of pressure sensor X, obtain the monitoring range of pressure sensor X, extract all coordinate points within this monitoring range to form set C, where set C = {(a1, b1, c1), (a2, b2, c2)......(a n , b n , c nExtract the overlapping coordinate points between set C and set A to form set D. Calculate the overlap rate between set A and set C according to the following formula :

[0061]

[0062] For example, set A has 10 coordinates, set C has 8 coordinates, and the overlapping coordinates of the two sets are 5. The overlapping coordinates form set D. The calculation process of the overlap rate is as follows = 0.277;

[0063] S204. Set the correlation between pressure sensor X and pressure sensor M as R X , and calculate the correlation R between pressure sensor M and pressure sensor X in set B according to the following formula X :

[0064] RX = + ;

[0065] Sort the pressure sensor X in set B in descending order according to the correlation R X . Set the number of verification values as Z. Extract Z pressure sensors X according to the descending order of the correlation, and use the data transmitted by the pressure sensor X at the abnormal time point Tx as the actual acquisition value

[0066] S3. Establish a training model. Use the historical acquisition data of the labeled pressure sensors and the historical acquisition data of the pressure sensors with the correlation reaching the set value as the training set to train the training model. Through the training model, calculate the calculated acquisition data of the pressure sensors with the correlation reaching the set value at the abnormal time point, and use this data as the output calculation value; Compare whether the actual acquisition data is the same as the output calculation value to obtain a matching result; Calculate the matching rate according to the matching result, and judge the trust result of the abnormal data according to the matching rate. The trust result in S300 includes "high credibility", "general credibility" and "low credibility";

[0067] S3 includes

[0068] S301. Establish a historical time set T. Extract the transmission data of pressure sensor M corresponding to the time node t n in the historical time set T, and the transmission data of pressure sensor X extracted in S204 at the time node t n . Establish a training model. Use the transmission data transmitted by pressure sensor M as the input and the transmission data transmitted by pressure sensor X as the output to train the model

[0069] S302. Calculate the calculated acquisition data of pressure sensor X at the abnormal time node Tx according to the training model in S301, and use this data as the output calculation value.

[0070] S303. Set the error range of the output calculation value, and compare whether the actual acquisition value and the output calculation value are within the error range. If the data of the actual acquisition value and the output calculation value are within the error range, output the result of "matched"; if the data of the actual acquisition value and the output calculation value are not within the error range, output the result of "not matched".

[0071] S304. Calculate the matching rate according to the obtained matching results of Z pressure sensors X, and set the matching rate as P. If P≥70%, output the trust result as "high credibility"; if 70%>P≥30%, output the trust result as "general credibility"; if P<30%, output the trust result as "low credibility". The calculation formula of the matching rate is as follows:

[0072]

[0073] Where Z represents the total number of pressure sensors X extracted in S204.

[0074] S4. If the trust result of "high credibility" is output in S3, the system automatically issues the system instruction corresponding to the abnormal data according to the setting; if the trust result of "general credibility" is output in S3, set an enlarged range, and the enlarged range is different from the coverage range, and repeat S2; if the trust result of "low credibility" is output in S3, extract the historical maintenance information of the marked pressure sensor, predict the maintenance time period of the marked pressure sensor, judge whether the marked pressure sensor is in the maintenance time period, and send information to the console.

[0075] S4 includes:

[0076] S401. When the trust result of "high credibility" is output in S304, it means that the accuracy of the abnormal data transmitted this time is high, and the system automatically issues the set instruction.

[0077] S402. When the trust result of "general credibility" is output in S304, set a new enlarged range again, extract all pressure sensors within the enlarged range to form a set F, and the set F = {pressure sensor 1, pressure sensor 2, pressure sensor 3,..., pressure sensor Y}, repeat the step of calculating the correlation between each pressure sensor within the coverage range and the marked pressure sensor in S2, and output the trust result of judging the abnormal data again.

[0078] S403. If the trust result of "low credibility" is output in S304, extract the historical maintenance information of the pressure sensor M, predict the maintenance time period of the pressure sensor M, and determine whether the abnormal time point Tx is within the maintenance time period of the pressure sensor M. If the abnormal time point Tx is within the maintenance time period of the pressure sensor M, output that the natural failure probability of the pressure sensor M is greater than the probability of human damage. If the abnormal time point Tx is not within the maintenance time period of the pressure sensor M, output that the natural failure probability of the pressure sensor M is less than the probability of human damage. Send the prediction information to the console.

[0079] S5. The test analysis module analyzes the monitored data processed by the data processing module, and then evaluates and analyzes the test performance of the pressure switch. The optimization analysis module analyzes the process optimization direction of the tested pressure switch and obtains the time influence coefficient and the pressure influence coefficient, compares the time influence coefficient and the pressure influence coefficient, and generates corresponding process optimization signals based on the comparison results.

[0080] The specific implementation method is as follows: When transmitting abnormal data to the system, mark the pressure sensor M that transmits the abnormal data at this time. Taking the pressure sensor M as the origin, establish a three-dimensional coordinate system and set a coverage range. For example, set the coverage range as a circle with the pressure sensor M as the center and a radius of 10m. Extract the coordinates of all pressure sensors X within the coverage range, calculate the correlation between the pressure sensor X and the pressure sensor M, extract the pressure sensor X whose correlation reaches the set value, and compare whether the data collected by the pressure sensor X matches the abnormal data at the abnormal time node Tx. That is to say, when the pressure sensor M collects abnormal data, the collected data of the pressure sensor X should also change. Use the training model to calculate the output calculation value that the pressure sensor X should change to, and compare it with the actual collected value actually collected by the pressure sensor X to verify the credibility of the abnormal data collected by the pressure sensor M. According to the matching results of the extracted multiple pressure sensors X, finally output the trust result of the abnormal data. The system can perform corresponding next operations according to the trust result. By verifying the abnormal data, the number of incorrect instructions issued by the system can be reduced, saving resources.

[0081] Embodiment 2:

[0082] A pressure switch fatigue test system based on data analysis, the system includes a monitoring module, a transmission module, a data processing module, a data extraction module, a data verification module, a test analysis module and an optimization analysis module connected to each other;

[0083] The monitoring module is used to collect the environmental data of the pressure sensor in the monitored pressure switch, and transmit the collected environmental data to the data processing module through the transmission module. The data processing module screens the monitored data. When abnormal data appears, the data extraction module extracts the abnormal data, marks the pressure sensor that transmits the abnormal data, and then marks the pressure switch corresponding to the pressure sensor. The data verification module verifies the credibility of the abnormal data and outputs corresponding instructions according to the verification result;

[0084] The monitoring module includes a pressure sensor unit for real-time monitoring of data in the environment; the transmission module includes a data transmission unit and an instruction output unit. The data transmission unit is used for real-time transmission of the monitored data, and the instruction output unit is used to send a monitoring instruction to the pressure sensor unit according to the instruction set by the system.

[0085] The data processing module includes a data preprocessing unit and a classification storage unit. The data preprocessing unit is used to denoise the monitored data. After removing redundant data, the classification storage unit is used to classify and store the monitored data.

[0086] The data extraction module includes an extraction unit, a marking unit and a modeling unit; the extraction unit is used to extract the transmitted abnormal data, and use the marking unit to mark the pressure sensor that transmits the abnormal data. At the same time, through the modeling unit, with the marked pressure sensor as the reference point, a three-dimensional coordinate system is established, the coverage range is set, and the positions of the pressure sensors within the coverage range are marked.

[0087] The data verification module includes a correlation calculation unit, a model establishment unit, a comparison unit, and a verification information output unit. The correlation calculation unit is used to calculate the correlation between the marked pressure sensor and the pressure sensors within the coverage range, and sort the pressure sensors within the coverage range in descending order according to the correlation, and extract the pressure sensors whose correlation reaches the set threshold; the model establishment unit is used to establish a training module, use the historical acquisition data of the marked pressure sensor as the input value, and use the historical acquisition data of the pressure sensors whose correlation reaches the threshold as the output value to train the model. Finally, calculate the output calculation value of the sensor whose correlation reaches the threshold at the abnormal time point Tx; use the comparison unit to compare the output calculation value with the actual acquisition value, output the matching result, and calculate the matching rate, and judge the trust result of the abnormal data according to the matching rate; the verification information output unit issues the next operation corresponding to the trust result according to the output information result.

[0088] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A pressure switch fatigue testing method based on data analysis, characterized in that: The method comprises the following specific steps: S1. Acquire monitoring data of the perception layer and extract abnormal data monitored; the perception layer includes a plurality of pressure sensors, which upload the monitored data information, mark the pressure sensor that transmits abnormal data, extract the marked pressure sensor information, and set the time node of the abnormal data transmission as the abnormal time point; S2. Taking the marked pressure sensor as a reference point, set a coverage range, obtain information of all pressure sensors within the coverage range, calculate the correlation between each pressure sensor within the coverage range and the marked pressure sensor; obtain information of pressure sensors whose correlation reaches a set value, and extract data transmitted by the pressure sensor whose correlation reaches the set value at the abnormal time point, and use the data as the actual collected value; S3, establish a training model, use the historical data of the marked pressure sensor and the historical data of the pressure sensor whose correlation reaches the set value as the training set, train the training model, calculate the calculated data of the pressure sensor whose correlation reaches the set value at the abnormal time point through the training model, and use the data as the output calculation value; Compare the actual collected data with the output calculated value to see if they are the same, and obtain a matching result; calculate a matching rate based on the matching result, and determine the trust result of the abnormal data based on the matching rate; S4. If S3 outputs a trust result of "high credibility", the system automatically issues a system instruction corresponding to the abnormal data according to the settings; if S3 outputs a trust result of "average credibility", an expansion range is set, and the expansion range is different from the coverage range, and S2 is repeated; if S3 outputs a trust result of "low credibility", the historical maintenance information of the marked pressure sensor is extracted, the maintenance time period of the marked pressure sensor is predicted, and it is determined whether the marked pressure sensor is in the maintenance time period, and the information is sent to the console; S5. The test analysis module analyzes the monitoring data processed by the data processing module, and then evaluates and analyzes the test performance of the pressure switch; the optimization analysis module analyzes the process optimization direction of the tested pressure switch and obtains the time influence coefficient and the pressure influence coefficient, compares the time influence coefficient with the pressure influence coefficient and generates a corresponding process optimization signal through the comparison result.

2. A pressure switch fatigue testing method based on data analysis according to claim 1, characterized in that: The S1 includes: S101, when abnormal data is detected in the perception layer, the abnormal data is captured, and the pressure sensor that transmits the abnormal data is marked, and the pressure sensor that transmits the abnormal data is marked as M pressure sensor, and a three-dimensional coordinate system is established with the marked M pressure sensor as the origin; S102 , obtaining the pressure sensor type of the M pressure sensor, obtaining the data type collected by the M pressure sensor, and when the M pressure sensor transmits abnormal data, setting the time node of transmitting the abnormal data as the abnormal time point Tx.

3. The method for fatigue testing a pressure switch based on data analysis according to claim 2, characterized in that: The S2 includes: S201, extract all pressure sensors to form a set B, where set B = {pressure sensor 1, pressure sensor 2, pressure sensor 3, pressure sensor X}; S202, extract the names of pressure sensor M and pressure sensor X in set B, assume that the name of pressure sensor M is W1, and the name of pressure sensor X in set B is W2, and use the following formula to calculate the similarity between the text names of each pressure sensor in set B and pressure sensor M : in, represents the similarity in name between pressure sensor M and pressure sensor X in set B; m represents the total number of characters in W1; n represents the total number of characters in W2; c represents the number of matching characters in W1 and W2; L1(i) and L2(i) are the matching sequences of matching character i in W1 and W2 respectively; and are the weights of the matching degree and matching order, which are 0.6 and 0.4 respectively according to the golden ratio; t represents the position index of each character in W1; k represents the position index of each character in W2; S203, according to the use standard of the pressure sensor M, obtain the monitoring range of the pressure sensor M, extract all the coordinate points within the monitoring range, and form a set A, where the set A={(x1, y1, z1), (x2, y2, z2)...(x n ,y n , z n )}; According to the use standard of pressure sensor X, the monitoring range of pressure sensor X is obtained, and all coordinate points within the monitoring range are extracted to form a set C, where set C={(a1, b1, c1), (a2, b2, c2)......(a n , b n , c n )}, extract the coincident coordinate points between set C and set A to form set D, and calculate the coincidence rate between set A and set C according to the following formula : S204: Set the correlation between the pressure sensor X and the pressure sensor M to R X , calculate the correlation R between pressure sensor M and pressure sensor X in set B according to the following formula X : R X = + ； The pressure sensors X in set B are sorted according to the correlation R X Arrange in descending order, set the number of verification values ​​to Z, extract Z pressure sensors X according to the descending order of correlation, and use the data transmitted by the pressure sensor X at the abnormal time point Tx as the actual collected value.

4. The method for fatigue testing a pressure switch based on data analysis according to claim 3, characterized in that: The S3 includes: S301: Establish a historical time set T, extract time node t in the historical time set T n The corresponding transmission data of the pressure sensor M, and the n When the transmission data of the pressure sensor X extracted in S204 is used to establish a training model, the transmission data transmitted by the pressure sensor M is used as input, and the transmission data transmitted by the pressure sensor X is used as output to train the model; S302: Calculate the calculated collected data of the pressure sensor X at the abnormal time node Tx according to the training model in S301, and use the data as the output calculated value; S303, setting the error range of the output calculation value, comparing the actual collected value with the output calculation value to see if they are within the error range; if the data of both the actual collected value and the output calculation value are within the error range, outputting a "match" result; if the data of both the actual collected value and the output calculation value are not within the error range, outputting a "mismatch" result; S304: Calculate the matching rate based on the matching results of the Z pressure sensors X, and output a trust result based on the matching rate, wherein the trust result is "high credibility", "average credibility" or "low credibility". The matching rate is calculated as follows: Wherein Z represents the total number of pressure sensors X extracted in S204.

5. A pressure switch fatigue testing method based on data analysis according to claim 4, characterized in that: The S4 includes: S401. If S304 outputs a trust result of "high credibility", it means that the abnormal data transmitted this time is highly accurate, and the system automatically issues the set instructions; S402, if S304 outputs a trust result of "general credibility", then set a new expanded range again, extract all pressure sensors within the expanded range to form a set F, where set F = {pressure sensor 1, pressure sensor 2, pressure sensor 3, pressure sensor Y}, repeat step S2 to calculate the correlation between each pressure sensor in the coverage range and the marked pressure sensor, and output the trust result of judging the abnormal data again; S403. If S304 outputs a trust result of "low credibility", the historical maintenance information of the pressure sensor M is extracted, the maintenance time period of the pressure sensor M is predicted, and it is determined whether the abnormal time point Tx is within the maintenance time period of the pressure sensor M; if the abnormal time point Tx is within the maintenance time period of the pressure sensor M, the output probability of natural failure of the pressure sensor M is greater than the probability of man-made damage; if the abnormal time point Tx is not within the maintenance time period of the pressure sensor M, the output probability of natural failure of the pressure sensor M is less than the probability of man-made damage; the prediction information is sent to the console.

6. A test system for the pressure switch fatigue test method based on data analysis according to any one of claims 1 to 5, characterized in that: The system includes a monitoring module, a transmission module, a data processing module, a data extraction module, a data verification module, a test analysis module and an optimization analysis module which are connected to each other; The monitoring module is used to collect environmental data of the pressure sensor in the monitored pressure switch, and transmit the collected environmental data to the data processing module through the transmission module. The data processing module screens the monitored data. When abnormal data appears, the data extraction module is used to extract the abnormal data, mark the pressure sensor that transmits the abnormal data, and then mark the pressure switch corresponding to the pressure sensor. The data verification module verifies the credibility of the abnormal data and outputs the corresponding instruction according to the verification result. The monitoring module includes a pressure sensor unit, which is used to monitor the data in the environment in real time; the transmission module includes a data transmission unit and an instruction output unit, the data transmission unit is used to transmit the monitored data in real time, and the instruction output unit is used to send monitoring instructions to the pressure sensor unit according to the instructions set by the system.

7. The test system according to claim 6, characterized in that: The data processing module includes a data preprocessing unit and a classification storage unit. The data preprocessing unit is used to perform denoising on the monitored data, remove redundant data, and then use the classification storage unit to classify and store the monitored data.

8. The test system according to claim 6, characterized in that: The data extraction module includes an extraction unit, a marking unit and a modeling unit; the extraction unit is used to extract the transmitted abnormal data, and use the marking unit to mark the pressure sensor that transmits the abnormal data. At the same time, through the modeling unit, a three-dimensional coordinate system is established with the marked pressure sensor as a reference point, a coverage range is set, and the position of the pressure sensor within the coverage range is marked.

9. The test system according to claim 6, characterized in that: The data verification module includes a correlation calculation unit, a model building unit, a comparison unit, and a verification information output unit. The correlation calculation unit is used to calculate the correlation between the marked pressure sensor and the pressure sensors within the coverage range, and to arrange the pressure sensors within the coverage range in descending order according to the correlation, and to extract the pressure sensors whose correlation reaches a set threshold; the model building unit is used to establish a training module, and use the historical collection data of the marked pressure sensor as an input value, and the historical collection data of the pressure sensor whose correlation reaches the threshold as an output value, to train the model, and finally calculate the output calculation value of the sensor whose correlation reaches the threshold at the abnormal time point Tx; The comparison unit is used to compare the output calculated value with the actual collected value, output a matching result, and calculate a matching rate, and determine the trust result of the abnormal data according to the matching rate; The verification information output unit issues a next step operation corresponding to the trust result according to the output information result.

Images