Method for predicting service life of relay for operating mechanism based on grey theoretical model

Abstract

The invention discloses a method for predicting the service life of a relay for an operating mechanism based on a grey theoretical model, and relates to the technical field of a relay for a high-voltage circuit breaker operating mechanism of a power system, and the method comprises the following steps: simulating a salt mist aging test parameter test system through an electromagnetic relay, and simulating the operation environment of the relay for the operating mechanism; an alternating salt spray test is carried out, and characteristic parameters of the relay are extracted; performing dimension reduction processing on the characteristic parameters of the relay by using a principal component analysis method, and extracting digital characteristics of the characteristic parameters; on the basis of the digital characteristics of the characteristic parameters, the characteristic parameter with the maximum discriminant function value is selected by utilizing Fisher discriminant criteria; and taking the characteristic parameter with the maximum discrimination function value as a prediction variable, constructing a gray theoretical model, predicting the service life of the relay, and completing the service life prediction of the relay for the operating mechanism based on the gray theoretical model. The problem that the existing relay life prediction precision is low is solved.

Description

Technical Field

[0001] This invention relates to the field of relay technology for operating mechanisms of high-voltage circuit breakers in power systems, and in particular to a method for predicting the lifespan of relays for operating mechanisms based on a grey theory model. Background Technology

[0002] High-voltage circuit breakers, widely used switching equipment in power distribution networks, bear the crucial responsibility of controlling line switching and interrupting short-circuit currents. Their reliability has a profound impact on the safety and stable operation of the power grid. The operating mechanism, as the mechanical power source of the high-voltage circuit breaker, directly determines the opening and closing process of the circuit breaker due to its dynamic characteristics. In recent years, with the increasing scale of the power system, the number of high-voltage circuit breakers in operation has risen significantly, and the probability of mechanical failures in the circuit breaker operating mechanism has also increased markedly. Survey results show that mechanical failures such as mechanical integrity issues, refusal to open, refusal to close, and lock-up in the open and closed positions account for a large proportion of the total failure rate of circuit breakers. Among these, mechanical failures caused by secondary components of the operating mechanism, such as relays, account for more than 75% of all failure rates. Relays, located in the secondary circuit, operate in a complex environment, constantly exposed to high humidity, high salt spray, and other environmental factors, which can damage their internal structure and gradually degrade their performance. For example, the relay armature may gradually corrode, potentially causing jamming during the linkage engagement process; aging of the electrical connections in the relay's secondary circuit can lead to poor circuit contact. Therefore, it is crucial to strengthen the research on the degradation patterns of relays.

[0003] Existing technologies typically use degradation-sensitive parameters obtained from accelerated life testing to analyze relay performance changes, focusing on selecting overtravel time and pull-in time as sensitive performance parameters for research. However, relying on only one or two sensitive parameters for modeling can lead to errors in many key information, resulting in low relay prediction accuracy. Summary of the Invention

[0004] To address the aforementioned shortcomings in existing technologies, this invention provides a relay life prediction method for operating mechanisms based on a grey theory model, which solves the problem of low accuracy in existing relay life prediction methods.

[0005] To achieve the aforementioned objectives, the technical solution adopted by this invention is as follows: a method for predicting the lifespan of relays used in operating mechanisms based on a grey theory model, comprising the following steps:

[0006] S1: The electromagnetic relay is used to simulate the salt spray aging test parameter testing system, which simulates the operating environment of the relay used in the operating mechanism and conducts alternating salt spray test to extract the characteristic parameters of the relay.

[0007] S2: Principal component analysis is used to reduce the dimensionality of the relay's characteristic parameters and extract their digital features.

[0008] S3: Based on the numerical features of the feature parameters, the characteristic parameter with the largest discriminant function value is selected using the Fisher discriminant criterion;

[0009] S4: Using the characteristic parameter with the largest discriminant function value as the prediction variable, construct a grey theory model to predict the relay life, and complete the prediction of relay life for operating mechanisms based on the grey theory model.

[0010] The beneficial effects of the above solution are: the present invention effectively mines and integrates the four-dimensional characteristic parameters of the relay, and combines the principal component analysis method and the grey theory life prediction model to quickly and accurately assess the remaining life of the relay, thus solving the problem of low accuracy in existing relay life prediction.

[0011] Furthermore, the electromagnetic relay simulated salt spray aging test parameter testing system in S1 includes a high-precision salt spray test chamber, a constant temperature and humidity test chamber, a signal control system, a parameter acquisition system, a lower-level computer, a higher-level computer, and a PC terminal;

[0012] The signal control system is connected to the high-precision salt spray test chamber, the constant temperature and humidity test chamber, the parameter acquisition system, and the lower-level computer, respectively. The high-precision salt spray test chamber is connected to the constant temperature and humidity test chamber and the parameter acquisition system, respectively. The constant temperature and humidity test chamber is connected to the parameter acquisition system, respectively. The upper-level computer is connected to the lower-level computer and the PC terminal, respectively.

[0013] The beneficial effect of the above-mentioned further solution is that, through the above technical solution, a test system for simulating salt spray aging test parameters is provided, which is used to simulate the operating environment of relays for operating mechanisms to carry out alternating salt spray tests.

[0014] Furthermore, the temperatures of the high-precision salt spray test chamber and the constant temperature and humidity test chamber are controlled by thermocouples;

[0015] The thermocouple includes thermocouple wire A and thermocouple wire B. One end of thermocouple wire A and one end of thermocouple wire B are connected to form a measuring working junction. The other ends of thermocouple wire A and thermocouple wire B are both connected to the measuring device through a copper wire C, forming a cold junction node. The potential E at the end of the copper wire is:

[0016] E = -e AC (t RJ )+e AB (t MJ )-e CB (t RJ )

[0017] =e AB (t MJ )-e AB (t RJ )

[0018] =E AB (t MJ ,t RJ )

[0019] Among them, e AC (·) represents the contact potential at the junction of thermocouple wire A and copper wire C, e AB (·) represents the contact potential at the measuring junction formed by thermocouple wire A and thermocouple wire B, e CB (·) represents the contact potential at the junction of thermocouple wire B and copper wire C, E AB (·) represents the output potential of thermocouple wire A and thermocouple wire B, t RJ For the reference node temperature, t MJ To measure the junction temperature.

[0020] The beneficial effects of the above-mentioned further solutions are: high-precision salt spray test chambers and constant temperature and humidity test chambers have high requirements for precise temperature control. Therefore, a thermocouple was designed to regulate the temperature. The circuit in the signal conditioning core system is related to the signal conditioning of the thermocouple to ensure that an accurate constant temperature is obtained.

[0021] Furthermore, the characteristic parameters in S1 include pull-in voltage, pull-in time, return voltage, and return time.

[0022] The beneficial effects of the above-mentioned further solution are as follows: For the relay's engagement and disengagement processes under salt spray corrosion, this solution selects four characteristic parameters representing its performance degradation, including two voltage parameters and two time parameters, which can improve the accuracy of life prediction and provide a more comprehensive analysis of the relay's characteristic parameters.

[0023] Furthermore, S2 includes the following sub-steps:

[0024] S2-1: The characteristic parameters of the relay are standardized using Z-score, and the correlation coefficient matrix of the standardized samples is calculated to obtain the dimensionless characteristic parameters. The formula is as follows:

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] in, For standardized indicator data, Xij The original dataset for the feature parameters, S is the mean of the sample data. i Let be the standard deviation of the i-th indicator. Let X be the mean of the i-th indicator. i S represents the sample dimension of the feature parameters, n represents the number of indicators, and S represents the number of indicators. j Let j be the standard deviation of the j-th variable. Let X be the mean of the j-th variable. j R is the length of the observed sample for the feature parameter. p×p Let r be the correlation coefficient matrix. ij Let X be the sample correlation coefficient, i = 1, 2, ..., p, j = 1, 2, ..., p, where p is the number of variables, and X is the coefficient of variation. ki X is the (k,i)th element of the sample deviation matrix. kj Let k be the (k,j)th element of the sample deviation matrix, where k is the selected index.

[0031] S2-2: Perform principal component analysis on the dimensionless characteristic parameters to obtain the principal component eigenvalues, and calculate the cumulative contribution rate of the principal components based on the principal component eigenvalues;

[0032] S2-3: Based on the cumulative contribution rate of the principal components, select the first principal component and the second principal component to extract mathematical features.

[0033] The beneficial effects of the above-mentioned further scheme are as follows: Since the units of actual data are usually inconsistent, Z-score standardization is applied to the original data to eliminate the influence of dimensions; finally, the first principal component and the second principal component are selected to extract mathematical features. The first principal component reflects the damage change trend of relay salt spray aging, and the second principal component illustrates the error information between the original data and the first principal component. The changing trends of the principal components can more clearly characterize the degradation process of relays under salt spray conditions.

[0034] Furthermore, the formula for calculating the cumulative contribution rate H of the principal components is:

[0035]

[0036] Where, λ l Let be the eigenvalue of the l-th group of feature parameters, o be the principal component to be selected, and l be the total number of principal components.

[0037] The beneficial effect of the above-mentioned further scheme is that the cumulative contribution rate of the principal components of the relay characteristic parameters can be obtained based on the principal component contribution rate. The larger the cumulative contribution rate of the principal components, the stronger its ability to reflect the changes in the original multidimensional characteristic parameters.

[0038] Furthermore, in S3, the Fisher criterion is used to establish a discriminant function, calculate the change value of each characteristic parameter, and select the characteristic parameter with the largest change value to reflect the relay life characteristics.

[0039] The beneficial effect of the above-mentioned further scheme is that the Fisher discriminant criterion can be used to analyze and judge the degree of correlation between and within the contact characteristic parameters, and the method of calculating the discriminant function value can quantitatively analyze the magnitude of these correlations.

[0040] Furthermore, the change value J of each characteristic parameter is calculated using the following formula:

[0041]

[0042]

[0043]

[0044] Where, n A n B and n C The number of groups for each category of data. and The average value for each category of data. Y is the average of all categories of data. Am Y Bm and Y Cm These are datasets for the feature parameters pull-in voltage, pull-in time, and return voltage, respectively, where m is the total number of data in each category.

[0045] The beneficial effect of the above further scheme is that the difference between each group of objects can be represented by the sum of squares of the separation mean difference. The change value of each feature parameter can be calculated by the above formula, and the degree of correlation of the four-dimensional feature parameters can be analyzed.

[0046] Furthermore, the characteristic parameters with the largest changes include pull-in time and return time.

[0047] The beneficial effect of the above-mentioned further scheme is that, through calculation of the discriminant function values ​​of each characteristic parameter, analysis shows that the discriminant function value of the pull-in time is the largest, and the discriminant function value of the return time is also relatively large. Therefore, the pull-in time and return time can significantly reflect the relay life characteristics. Attached Figure Description

[0048] Figure 1 This is a flowchart of a relay life prediction method for operating mechanisms based on a grey theory model.

[0049] Figure 2 This is a structural diagram of a system for testing parameters in a simulated salt spray aging test using an electromagnetic relay.

[0050] Figure 3 This is a schematic diagram of the thermocouple measurement principle.

[0051] Figure 4 This is a comparison chart of the cumulative contribution rates of principal components for characteristic parameters.

[0052] Figure 5 This is a distribution map of the average gradients of the first and second principal components under different damage levels.

[0053] Figure 6 This is a comparison chart of relay life prediction based on the grey theory model. Detailed Implementation

[0054] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0055] like Figure 1 As shown, a method for predicting the lifespan of relays for operating mechanisms based on a grey theory model includes the following steps:

[0056] S1: The electromagnetic relay is used to simulate the salt spray aging test parameter testing system, which simulates the operating environment of the relay used in the operating mechanism and conducts alternating salt spray test to extract the characteristic parameters of the relay.

[0057] S2: Principal component analysis is used to reduce the dimensionality of the relay's characteristic parameters and extract their digital features.

[0058] S3: Based on the numerical features of the feature parameters, the characteristic parameter with the largest discriminant function value is selected using the Fisher discriminant criterion;

[0059] S4: Using the characteristic parameter with the largest discriminant function value as the prediction variable, construct a grey theory model to predict the relay life, and complete the prediction of relay life for operating mechanisms based on the grey theory model.

[0060] The S1 electromagnetic relay simulated salt spray aging test parameter testing system includes a high-precision salt spray test chamber, a constant temperature and humidity test chamber, a signal control system, a parameter acquisition system, a lower-level computer, a higher-level computer, and a PC terminal, such as... Figure 2 As shown;

[0061] The signal control system is connected to the high-precision salt spray test chamber, the constant temperature and humidity test chamber, the parameter acquisition system, and the lower-level computer, respectively. The high-precision salt spray test chamber is connected to the constant temperature and humidity test chamber and the parameter acquisition system, respectively. The constant temperature and humidity test chamber is connected to the parameter acquisition system, respectively. The upper-level computer is connected to the lower-level computer and the PC terminal, respectively.

[0062] The temperature of the high-precision salt spray test chamber and the constant temperature and humidity test chamber is controlled by thermocouples;

[0063] The thermocouple includes thermocouple wire A and thermocouple wire B. One end of thermocouple wire A and one end of thermocouple wire B are connected to form a measuring working junction. The other ends of thermocouple wire A and thermocouple wire B are both connected to the measuring device through a copper wire C, forming a cold junction node. The potential E at the end of the copper wire is:

[0064] E = -e AC (t RJ )+e AB (t MJ )-e CB (t RJ )

[0065] =e AB (t MJ )-e AB (t RJ )

[0066] =E AB (t MJ ,t RJ )

[0067] Among them, e AC (·) represents the contact potential at the junction of thermocouple wire A and copper wire C, e AB (·) represents the contact potential at the measuring junction formed by thermocouple wire A and thermocouple wire B, e CB (·) represents the contact potential at the junction of thermocouple wire B and copper wire C, E AB (·) represents the output potential of thermocouple wire A and thermocouple wire B, t RJ For the reference node temperature, t MJ To measure the junction temperature.

[0068] In one embodiment of the present invention, since high-precision salt spray test chambers and constant temperature and humidity test chambers have high requirements for precise temperature control, a thermocouple was specifically designed to regulate the temperature, such as... Figure 3As shown, the circuit in the signal conditioning system is related to the signal conditioning of the thermocouple to ensure accurate and constant temperature. The thermocouple is made of two different metal wires. One end of these wires is connected to form the measuring working junction, while the other end is connected to two wires on the measuring device to form the cold junction. By measuring the potential difference between the thermocouple's working junction and the cold junction, the temperature can be calculated. Based on the thermoelectric effect, the function of the thermocouple is to convert temperature changes into a voltage signal output, thus enabling temperature detection and monitoring. The potential difference is derived from the thermoelectric potential difference between the measuring junction and the reference junction at different temperatures, not from the connection state of wire C. This potential difference is closely related to the characteristics of wires A and B in the thermocouple. This is the working principle of the thermocouple intermediate conductor law, allowing the thermocouple to be connected to the measuring instrument via wires to achieve temperature measurement. As long as the measuring junction and the reference junction are kept at the same temperature and the measuring wires are made of the same material, the temperature t of the reference junction can be used to calculate the temperature. RJ By analyzing the potential value of the measuring instrument, the thermoelectric potential corresponding to the temperature of the measuring junction is calculated, and thus the temperature t of the measuring junction is deduced. MJ Cold junction compensation compensates for the temperature of the reference junction by measuring the temperature of the connection point between the conductor C and the thermocouple. During the measurement process, the measuring instrument measures the temperature of conductor C and performs corresponding compensation calculations based on this temperature value to eliminate the influence of the cold junction, maintain the stability of the thermocouple output, and eliminate errors caused by temperature differences. This provides more accurate temperature measurement results.

[0069] The parameter acquisition system provides a three-phase voltage and current source with flexible amplitude, phase, and frequency variations. The adjustability of the magnitude and direction of these variations is sufficient to meet the requirements of various tests on different types of relays. The amplitude, phase, and frequency of the three-phase voltage and current can be manually increased or decreased in steps simultaneously, causing the relay to switch from active to inactive, measuring and recording the active value; then, it switches back to active and measures the return value. The selection, direction, and magnitude of various variations are all flexibly adjustable, and the changes in current and voltage, as well as the output power, are displayed in real time. Real-time displays include vector diagrams, waveforms, RMS values, vector sequence tables, and contact operation status, providing various real-time information. For DC testing, the module can test the active and return values ​​of DC relays, as well as their action and return times. The lower-level computer is connected to the signal control system to control the four characteristic parameters measured by the salt spray chamber, parameter acquisition system, and constant temperature and humidity test chamber; simultaneously controlling the temperature and humidity changes in the experimental environment; and transmitting data to the upper-level computer. The host computer issues a start command to the slave computer, and simultaneously displays real-time control data on the PC, completing data acquisition and monitoring. The PC displays and records the data curve distribution of characteristic parameters.

[0070] Before the experiment, a weakly acidic sodium chloride solution was prepared, maintaining its pH between 6.5 and 7.2, while ensuring the salt concentration was (5±0.5)%. After meeting these conditions, the solution was poured into a high-precision salt spray chamber. The chamber temperature was set to 45℃, and the pressure in the pressure tank was kept constant. Once set, the salt spray was applied to the relay. After 4 hours of spraying, the relay was removed and placed in a constant temperature and humidity chamber for 4 hours of humidified storage (45℃, 95% humidity). A 4-hour spraying followed by 4 hours of constant temperature and humidity storage is defined as one aging cycle. To clarify the impact of salt spray aging on the relay's degradation performance, four characteristic parameters were measured using an AT-900 integrated automated testing system after each test cycle. Each characteristic parameter was tested 25 times, and the data was recorded as a dataset until the relay completely deteriorated and failed.

[0071] The characteristic parameters in S1 include pull-in voltage, pull-in time, return voltage, and return time.

[0072] S2 includes the following steps:

[0073] S2-1: The characteristic parameters of the relay are standardized using Z-score, and the correlation coefficient matrix of the standardized samples is calculated to obtain the dimensionless characteristic parameters. The formula is as follows:

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] in, For standardized indicator data, X ij The original dataset for the feature parameters, S is the mean of the sample data. i Let be the standard deviation of the i-th indicator. Let X be the mean of the i-th indicator. i S represents the sample dimension of the feature parameters, n represents the number of indicators, and S represents the number of indicators. j Let j be the standard deviation of the j-th variable. Let X be the mean of the j-th variable. j R is the length of the observed sample for the feature parameter. p×p Let r be the correlation coefficient matrix. ij Let X be the sample correlation coefficient, i = 1, 2, ..., p, j = 1, 2, ..., p, where p is the number of variables, and X is the coefficient of variation. kiX is the (k,i)th element of the sample deviation matrix. kj Let k be the (k,j)th element of the sample deviation matrix, where k is the selected index.

[0080] S2-2: Perform principal component analysis on the dimensionless characteristic parameters to obtain the principal component eigenvalues, and calculate the cumulative contribution rate of the principal components based on the principal component eigenvalues;

[0081] S2-3: Based on the cumulative contribution rate of the principal components, select the first principal component and the second principal component to extract mathematical features.

[0082] As the salt spray aging time increases, the relay's electrical parameters and time parameters exhibit unique changing trends, reflecting changes in the relay's operating condition and degree of damage. There are certain connections and distinctions among these multidimensional characteristic parameters. To eliminate redundancy and interference information and improve computational capabilities, this scheme employs principal component analysis to perform dimensionality reduction processing on the four-dimensional characteristic parameter data.

[0083] Principal Component Analysis (PCA) is a statistical process for feature extraction. Its core idea is to use orthogonal transformations to convert the observations of a set of interrelated variables into a set of linearly independent variable values, thus reflecting most of the information in the original data through dimensionality reduction. Assuming the original dataset has n samples, each containing p variables, then an n×p type original data matrix is ​​defined as follows:

[0084]

[0085] In matrix X n×p In the diagram, column vector X i (i = 1, 2, 3, ..., n) is the indicator of the feature parameter, i.e., the sample dimension, X j (j = 1, 2, 3, ..., p) represents the observation sample length for each parameter. Since the units of actual data are often inconsistent, Z-score standardization is applied to the original data to eliminate the influence of unit of measurement. The correlation coefficient matrix R of the standardized samples is then calculated. p×p .

[0086] The formula for calculating the cumulative contribution rate H of the principal components is:

[0087]

[0088] Where, λ l Let be the eigenvalue of the l-th group of feature parameters, o be the principal component to be selected, and l be the total number of principal components.

[0089] Principal component analysis was performed on 10 sets of standardized four-dimensional feature parameters to calculate matrix R. p×pThe eigenvalues ​​λ1, λ2, λ3, λ4 (λ1≥λ2≥λ3≥λ4≥0) and the corresponding eigenvectors α. i The contribution rate of the o-th principal component is The cumulative contribution rate of the first o principal components is The larger the contribution rate of the principal components, the stronger their ability to reflect the changes in the comprehensive original multidimensional characteristic parameters. Table 1 lists the eigenvalues ​​of the four principal components related to the cumulative salt spray aging time in 10 sets of data.

[0090] Table 1. Eigenvalues ​​of principal components for each group

[0091]

[0092]

[0093] As can be clearly seen from the table above, among the 10 sample data, the first principal component has the largest eigenvalue, the second principal component has a relatively smaller eigenvalue, and the third and fourth principal components have progressively smaller eigenvalues, accounting for a relatively low proportion. The first principal component is a linear combination direction vector obtained after standardizing the original variables. It contains the largest interpretable variance among the original variables and is the main source of data variability, reflecting the principle that the larger the principal component eigenvalue, the more severe the relay damage. The second principal component is a direction vector orthogonal to the first principal component and with the second largest variance. Both can effectively integrate the variation information of the original multidimensional feature parameters. Based on the eigenvalues ​​of the principal components of the characteristic parameters of each group in Table 1, and using the previously mentioned formula to calculate the cumulative contribution rate of the principal components, the cumulative contribution rate of the principal components of the relay's characteristic parameters can be obtained as follows: Figure 4 As shown.

[0094] Degradation parameters such as pull-in voltage, release voltage, pull-in time, and release time are all closely related to the internal structure of the relay affected by salt spray corrosion. In other words, the four-dimensional feature parameters encompass information related to relay damage. Stacked bar charts can more intuitively display the contribution rate of the principal components. The data is divided into 10 columns, each representing 10 different groups of salt spray aging time data. Each column, arranged from bottom to top, presents the cumulative effect of the contribution rates of the four principal components, from the first to the fourth. By stacking the contribution percentages of each principal component, the cumulative contribution rate of the principal components is formed. The vertical axis in the chart clearly highlights the proportion of each column of principal components. It can be seen that the contribution rate of the first principal component is approximately 40%, and since it has the largest contribution rate, it integrates the information of most of the original variables, reflecting the degree of damage to the relay. The contribution rate of the second principal component is slightly lower, but still around 30%. The cumulative contribution rates of the first two principal components in the 10 samples generally reach approximately 70%, indicating that the first two principal components can typically represent the information of all feature parameters.

[0095] The second principal component reveals the error information between the original data and the first principal component. The smaller the absolute value of the second principal component, the smaller the error between the original data and the first principal component, indicating that the characteristic parameters change significantly under salt spray erosion, the correlation is strong, and the analysis results are more accurate. The first two principal components are sufficient to retain most of the correlation between the characteristic parameters, satisfying the conditions for principal component selection. Therefore, this scheme studies the first two principal components to analyze the damage degree of the relay.

[0096] To quantitatively describe the correlation characteristics between the first and second principal components, the average gradient h of the principal components is defined here. i Its physical meaning is the vector difference between the average value of the principal components in the second half of the salt spray life test and the average value in the first half:

[0097]

[0098] In the formula, h i h is the average gradient of the i-th principal component. i The absolute value represents the magnitude of the change, T i Let be the data corresponding to the i-th principal component, and n be the total number of data points. Based on the above formula, the relationship between the average gradient values ​​of the first two principal components and different damage levels of the relay is calculated, as shown in the figure. Figure 5 As shown in the figure, the graph is divided into four quadrants using the mean gradient values ​​of the first and second principal components as the x and y coordinates, respectively. It is evident that the discrete points under different damage levels exhibit different clustering phenomena: in the virgin and failed states, the absolute value of the mean gradient of the first principal component is larger, but in opposite directions (positive for the virgin state and negative for the failed state), and the second principal component value is larger in the virgin state; the coordinate points of mild damage are distributed in the first quadrant, above the x-axis, and all are less than 0.5 from the x-axis; the points of the warning state are distributed in all three quadrants, but mainly centered around the origin, with the largest span along the y-axis. The coordinate points of severe damage are clustered to the lower left of the origin, i.e., in the third quadrant, and the absolute values ​​of the mean gradients of the first and second principal components are both less than 0.5. Therefore, it is feasible to use the first two principal components instead of the four feature parameters when assessing the severity of relay damage under salt spray conditions.

[0099] The multidimensional relay fusion method based on principal component analysis transforms multiple linearly correlated characteristic parameters into linearly independent principal components for discriminant classification. This approach integrates the characteristics of each characteristic parameter and simplifies data processing by selecting the first few principal components. To analyze the correlation between the principal components and the original characteristic parameters in performance changes under different salt spray durations, and to prepare for the selection of predictive variables, Fisher's discriminant criterion is used to establish a discriminant function to calculate the degree of separation of each characteristic parameter, selecting parameters with larger separation ratios as the research objects. Then, the change values ​​of the research objects are calculated, and parameter analysis is performed under different degrees of damage. The difference between each group of objects can be represented by the sum of squares of the mean separation differences, analyzing the degree of correlation of the four-dimensional characteristic parameters.

[0100] In S3, the Fisher criterion is used to establish a discriminant function, calculate the change value of each characteristic parameter, and select the characteristic parameter with the largest change value to reflect the relay life characteristics.

[0101] The formula for calculating the change value J of each characteristic parameter is:

[0102]

[0103]

[0104]

[0105] Where, n A n B and n C The number of groups for each category of data. and Y is the average of the data for each category, and Y is the average of the data for all categories. Am Y Bm and Y Cm These are datasets for the feature parameters pull-in voltage, pull-in time, and return voltage, respectively, where m is the total number of data in each category.

[0106] The characteristic parameters with the largest changes include pull-in time and return time.

[0107] The discriminant function values ​​of each characteristic parameter were calculated and are shown in Table 2. It can be seen that the discriminant function value of the pull-in time is the largest, and the discriminant function value of the return time is also relatively large. Therefore, the pull-in time and return time can significantly reflect the relay life characteristics, so these two parameters are selected for modeling.

[0108] Table 2. Discrimination function values ​​for characteristic parameters

[0109]

[0110] The training and test sets used in this scheme were obtained through previous experiments, ensuring data reliability. To fully assess the health status of the relays, the existing datasets were classified into five damage levels based on the inherent degradation performance of the relays, as shown in Table 3.

[0111] Table 3 Classification and Labeling of Relay Damage Levels

[0112]

[0113] The grey theory model treats discrete data scattered along the time axis as a continuously changing sequence. By employing accumulation and subtraction, it weakens the unknown factors in the grey system while strengthening the influence of known factors. Finally, it constructs a continuous differential equation with time as the variable, and uses mathematical methods to determine the parameters in the equation, thereby achieving the prediction objective. This scheme applies the grey prediction model to the prediction of relay lifespan based on multivariate electrical parameters, and analyzes and compares its prediction effects.

[0114] When modeling, pull-in time and return time are selected. However, the pull-in voltage and return voltage are also related to the time parameter during the operation of the relay. In order to integrate and bind the four parameters, the original feature sequence and the relevant factor sequence are input through Matlab. A gray theory lifetime prediction model based on multi-parameter fusion is adopted.

[0115] like Figure 6 The figure shows the relay life prediction results based on the grey theory model. In this embodiment, the failure threshold is selected as 350ms. Figure 5 It can be seen that the engagement time reaches the failure threshold when the salt spray aging cycle is close to the fourth cycle. The actual lifespan of the relay is 30.896 hours, the predicted lifespan is 31.408 hours, and the relative error is 1.657%. The lifespan prediction accuracy is relatively high. The grey model established by segmenting the average data from the first five salt spray aging cycles can well reflect the actual trend of overtravel time changing with the number of operations. The fitting curve and the predicted curve have a high degree of agreement, and the error between the predicted value and the actual value can be kept within a small range.

[0116] This invention considers the selection criteria for relay characteristic parameters and integrates the weights of four-dimensional characteristic parameters. Studying the relay failure mechanism helps to understand the changing trends of characteristic parameters under various failure modes. Furthermore, there are many characteristic parameters reflecting the lifespan of relays used in operating mechanisms; selecting parameters with a significant impact on relay lifespan as the research object allows for more effective failure mechanism identification and lifespan prediction. To improve the accuracy of lifespan prediction, four relay characteristic parameters are selected as the research object: pull-in voltage, pull-in time, return voltage, and return time, allowing for a more comprehensive analysis of the relay's characteristic parameters. After noise reduction processing of the characteristic parameters, principal component analysis is used to reduce the dimensionality of the characteristic parameter dataset and extract the corresponding digital features; Fisher discriminant analysis is used to analyze the characteristic parameters, selecting the characteristic parameter with the largest discriminant function value as the prediction variable; given limited data, a grey theory model is selected to predict the relay lifespan, achieving relay lifespan assessment. The model proposed in this invention has high lifespan prediction accuracy and good feasibility. This method has practical value for improving the reliability of relays used in operating mechanisms, enhancing their operation management, and facilitating maintenance.

[0117] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of the invention.

Claims

1. A method for predicting the lifespan of relays used in operating mechanisms based on a grey theory model, characterized in that, Includes the following steps: S1: The electromagnetic relay is used to simulate the salt spray aging test parameter testing system, which simulates the operating environment of the relay used in the operating mechanism and conducts alternating salt spray test to extract the characteristic parameters of the relay. S2: Principal component analysis is used to reduce the dimensionality of the relay's characteristic parameters and extract their digital features. S3: Based on the numerical features of the feature parameters, the characteristic parameter with the largest discriminant function value is selected using the Fisher discriminant criterion; S4: Using the characteristic parameter with the largest discriminant function value as the prediction variable, construct a grey theory model to predict the relay life, and complete the prediction of relay life for operating mechanisms based on the grey theory model.

2. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 1, characterized in that, The electromagnetic relay simulated salt spray aging test parameter testing system in S1 includes a high-precision salt spray test chamber, a constant temperature and humidity test chamber, a signal control system, a parameter acquisition system, a lower computer, a upper computer, and a PC terminal; The signal control system is connected to the high-precision salt spray test chamber, the constant temperature and humidity test chamber, the parameter acquisition system, and the lower-level computer, respectively. The high-precision salt spray test chamber is connected to the constant temperature and humidity test chamber and the parameter acquisition system, respectively. The constant temperature and humidity test chamber is connected to the parameter acquisition system, respectively. The upper-level computer is connected to the lower-level computer and the PC terminal, respectively.

3. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 2, characterized in that, The temperatures of the high-precision salt spray test chamber and the constant temperature and humidity test chamber are controlled by thermocouples. The thermocouple includes thermocouple wire A and thermocouple wire B. One end of thermocouple wire A and one end of thermocouple wire B are connected to form a measuring working junction. The other ends of thermocouple wire A and thermocouple wire B are both connected to the measuring device through a copper wire C, forming a cold junction node. The potential E at the end of the copper wire is: E=-e AC (t RJ )+e AB (t MJ )-e CB (t RJ ) =e AB (t MJ )-e AB (t RJ ) =E AB (t MJ ,t RJ ) Among them, e AC (·) represents the contact potential at the junction of thermocouple wire A and copper wire C, e AB (·) represents the contact potential at the measuring junction formed by thermocouple wire A and thermocouple wire B, e CB (·) represents the contact potential at the junction of thermocouple wire B and copper wire C, E AB (·) represents the output potential of thermocouple wire A and thermocouple wire B, t RJ For the reference node temperature, t MJ To measure the junction temperature.

4. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 1, characterized in that, The characteristic parameters in S1 include pull-in voltage, pull-in time, return voltage, and return time.

5. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 1, characterized in that, S2 includes the following sub-steps: S2-1: The characteristic parameters of the relay are standardized using Z-score, and the correlation coefficient matrix of the standardized samples is calculated to obtain the dimensionless characteristic parameters. The formula is as follows: in, For standardized indicator data, X ij The original dataset for the feature parameters, S is the mean of the sample data. i Let be the standard deviation of the i-th indicator. Let X be the mean of the i-th indicator. i S represents the sample dimension of the feature parameters, n represents the number of indicators, and S represents the number of indicators. j Let j be the standard deviation of the j-th variable. Let X be the mean of the j-th variable. j R is the length of the observed sample for the feature parameter. p×p Let r be the correlation coefficient matrix. ij Let X be the sample correlation coefficient, i = 1, 2, ..., p, j = 1, 2, ..., p, where p is the number of variables, and X is the coefficient of variation. ki X is the (k,i)th element of the sample deviation matrix. kj Let k be the (k,j)th element of the sample deviation matrix, where k is the selected index. S2-2: Perform principal component analysis on the dimensionless characteristic parameters to obtain the principal component eigenvalues, and calculate the cumulative contribution rate of the principal components based on the principal component eigenvalues; S2-3: Based on the cumulative contribution rate of the principal components, select the first principal component and the second principal component to extract mathematical features.

6. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 5, characterized in that, The formula for calculating the cumulative contribution rate H of the principal components is as follows: Where, λ l Let be the eigenvalue of the l-th group of feature parameters, o be the principal component to be selected, and l be the total number of principal components.

7. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 1, characterized in that, In step S3, a discriminant function is established using the Fisher discrimination criterion, and the change values ​​of each characteristic parameter are calculated. The characteristic parameter with the largest change value is selected to reflect the relay life characteristics.

8. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 7, characterized in that, The formula for calculating the change value J of each characteristic parameter is as follows: Where, n A n B and n C The number of groups for each category of data. and The average value for each category of data. Y is the average of all categories of data. Am Y Bm and Y Cm These are datasets for the feature parameters pull-in voltage, pull-in time, and return voltage, respectively, where m is the total number of data in each category.

9. The method for predicting the lifespan of relays for operating mechanisms based on a grey theory model according to claim 7, characterized in that, The characteristic parameters with the largest changes include the pull-in time and the return time.

Images