Servo system reliability evaluation method based on K-S test

Abstract

The invention discloses a servo system reliability evaluation method based on K-S test. The method comprises the following steps: based on a fault data set of a servo system, performing K-S test by using an empirical distribution function and a fault probability distribution function of a Weibull distribution model; and if the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, updating the position of each particle in the state search space through the particle swarm algorithm, outputting the current parameter of the Weibull distribution model, updating the Weibull distribution model by using the current parameter, and returning to the step of performing the K-S test. According to the method, the maximum difference of individuals is concerned in the K-S test, the particle swarm algorithm takes an error sum of squares function as a target function, the error of the whole is concerned to be minimum, the two functions are combined, the parameters of the solved Weibull distribution model are more accurate, and the convergence speed is higher through the particle swarm algorithm.

Description

Technical Field

[0001] This application relates to the technical field of servo system safety assurance, and particularly to a method for evaluating the reliability of a servo system based on the K-S test. Background Art

[0002] A servomechanism, also known as a follow-up system, is a feedback control system used to accurately follow or reproduce a certain process. A servomechanism is an automatic control system that enables output controlled variables such as the position, orientation, and state of an object to follow any change in an input target (or given value).

[0003] The Weibull Distribution is a continuous probability distribution widely used in the fields of reliability engineering, survival analysis, and life data analysis. The Weibull Distribution is also a commonly used failure distribution model. By adjusting the parameters of the Weibull distribution model, different forms of failure rate changes can be simulated, including decreasing type, increasing type, and constant type, etc., which can adapt to various actual situations. The failure distribution model describes the change of the equipment failure rate over the operating time and is the basis for reliability evaluation.

[0004] However, the current Weibull distribution model used for evaluating the reliability of a servo system is not accurate enough, and there is a large deviation between the evaluation result and the actual operating condition of the servo system. An inaccurate reliability evaluation result may cause incorrect decisions and plans, thus increasing the maintenance cost. Summary of the Invention

[0005] This application provides a method for evaluating the reliability of a servo system based on the K-S test, aiming to solve the technical problems that the current Weibull distribution model used for evaluating the reliability of a servo system is not accurate enough, there is a large deviation between the evaluation result and the actual operating condition of the servo system, and an inaccurate reliability evaluation result may cause incorrect decisions and plans, thus increasing the maintenance cost.

[0006] In a first aspect, an embodiment of this application provides a method for evaluating the reliability of a servo system based on the K-S test. The method for evaluating the reliability of a servo system based on the K-S test includes:

[0007] Using the parameters of the Weibull distribution model as particles of the particle swarm algorithm and the sum of squared error function as the objective function of the particle swarm algorithm, and solving through the particle swarm algorithm based on the fault data set of the servo system, outputting the current parameters of the Weibull distribution model, and updating the Weibull distribution model with the current parameters;

[0008] Based on the fault data set of the servo system, performing a K-S test using the empirical distribution function and the failure probability distribution function of the Weibull distribution model;

[0009] If the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, then update the position of each particle in the state search space through the particle swarm algorithm, output the current parameters of the Weibull distribution model, update the Weibull distribution model using the current parameters, and return to execute the step of performing the K-S test on the fault data set of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0010] If the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, then use the current parameters output by the particle swarm algorithm as the optimal parameters of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system using the optimal parameters.

[0011] Optionally, performing the K-S test on the fault data set of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model includes:

[0012] Based on the fault data set of the servo system, calculate the K-S statistic according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0013] Determine the corresponding maximum absolute difference according to the sample size of the fault data set and the preset significance level;

[0014] If the K-S statistic is less than the maximum absolute difference, then determine that the test result is passing the K-S test;

[0015] If the K-S statistic is greater than or equal to the maximum absolute difference, then determine that the test result is not passing the K-S test.

[0016] Optionally, calculating the K-S statistic based on the fault data set of the servo system according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model includes:

[0017] Based on the fault data set of the servo system, calculate the K-S statistic through formula one according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model, and the formula one is:

[0018]

[0019] where D is the K-S statistic, F0(t i ) is the fault probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample size of the fault data set, t i$x_i$ is the $i$-th sample data of the fault data set, $\beta$ is the shape parameter of the Weibull distribution model, and $\eta$ is the scale parameter of the Weibull distribution model.

[0020] Optionally, updating the position of each particle by the particle swarm algorithm and outputting the current parameters of the Weibull distribution model includes:

[0021] Updating the position of each particle through Equation 2;

[0022] For each particle, using the updated each particle to update the Weibull distribution model, and based on the fault data set of the servo system, calculating the sum of squared errors of each particle after update through the objective function;

[0023] If the sum of squared errors of the particle after update is less than the sum of squared errors before update, then update the updated particle to the individual optimal solution of the particle;

[0024] If the sum of squared errors corresponding to the individual optimal solution of the updated particle is less than the sum of squared errors corresponding to the global optimal solution, then update the individual optimal solution of the updated particle to the global optimal solution;

[0025] Output the global optimal solution as the current parameters of the Weibull distribution model;

[0026] Among them, Equation 2 is:

[0027] x i (t + 1)=x i (t)+v i (t + 1),v i (t + 1)=ωv i (t)+c1r1(p i -x i (t))+c2r2(p g -x i (t));

[0028] The objective function is determined according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model, and the formula of the objective function is:

[0029]

[0030] Among them, x i (t + 1) is the position of particle $i$ after update, x i (t) is the position of particle $i$ before update, v i (t + 1) is the velocity of particle $i$ after update, v i (t) is the velocity of particle $i$ before update, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers between 0 and 1, p i is the individual optimal solution of particle $i$, pg is the global optimal solution of the particle, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample number of the failure data set, t i is the i-th sample data of the failure data set, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

[0031] Optionally, when the sum of squared errors corresponding to the current parameters is greater than or equal to the traditional sum of squared errors, and the number of iterations of the particle swarm optimization algorithm is less than or equal to the preset number of times, it is determined that the end condition of the particle swarm optimization algorithm is not satisfied. When the sum of squared errors corresponding to the current parameters is less than the traditional sum of squared errors, or the number of iterations of the particle swarm optimization algorithm is greater than the preset number of times, it is determined that the end condition of the particle swarm optimization algorithm is satisfied.

[0032] Optionally, before using the parameters of the Weibull distribution model as the particles of the particle swarm optimization algorithm, it includes:

[0033] Solving the parameters of the Weibull distribution model using a traditional method, and updating the Weibull distribution model using the obtained parameters;

[0034] Based on the failure data set of the servo system, calculating the traditional sum of squared errors according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model.

[0035] Optionally, after using the current parameters output by the particle swarm optimization algorithm as the best parameters of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system using the best parameters, it includes:

[0036] Based on the failure data set of the servo system, calculating the root mean square error according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model for evaluating the accuracy of the Weibull distribution model using the root mean square error.

[0037] In a second aspect, an embodiment of the present application provides a servo system reliability evaluation device based on the K-S test. The servo system reliability evaluation device based on the K-S test includes:

[0038] The initial solution module is used to use the parameters of the Weibull distribution model as the particles of the particle swarm optimization algorithm, use the sum of squared errors function as the objective function of the particle swarm optimization algorithm, perform solution through the particle swarm optimization algorithm based on the failure data set of the servo system, output the current parameters of the Weibull distribution model, and update the Weibull distribution model using the current parameters;

[0039] A K-S test module, configured to perform a K-S test on a fault data set of a servo system by using an empirical distribution function and a fault probability distribution function of a Weibull distribution model;

[0040] An update and solution module, configured to, if the K-S test fails or the end condition of the particle swarm algorithm is not satisfied, update the position of each particle in the state search space by using the particle swarm algorithm, output the current parameters of the Weibull distribution model, update the Weibull distribution model by using the current parameters, and return to execute the step of performing a K-S test on the fault data set of the servo system by using the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0041] An output module, configured to, if the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, use the current parameters output by the particle swarm algorithm as the optimal parameters of the Weibull distribution model, so that the Weibull distribution model can evaluate the reliability of the servo system by using the optimal parameters.

[0042] In a third aspect, an embodiment of the present application provides a servo system reliability evaluation device based on K-S test. The servo system reliability evaluation device based on K-S test includes a processor, a memory, and a servo system reliability evaluation program based on K-S test stored on the memory and executable by the processor. When the servo system reliability evaluation program based on K-S test is executed by the processor, the steps of the servo system reliability evaluation method based on K-S test as described above are implemented.

[0043] In a fourth aspect, an embodiment of the present application provides a readable storage medium. A servo system reliability evaluation program based on K-S test is stored on the readable storage medium. When the servo system reliability evaluation program based on K-S test is executed by a processor, the steps of the servo system reliability evaluation method based on K-S test as described above are implemented.

[0044] The beneficial effects brought by the technical solution provided by the embodiment of the present application include:

[0045] In the embodiments of the present application, the parameters of the Weibull distribution model are used as the particles of the particle swarm algorithm, the sum of squared error function is used as the objective function of the particle swarm algorithm, and based on the fault data set of the servo system, the solution is obtained through the particle swarm algorithm, and the current parameters of the Weibull distribution model are output, and the Weibull distribution model is updated using the current parameters; based on the fault data set of the servo system, the Kolmogorov-Smirnov (K-S) test is performed using the empirical distribution function and the fault probability distribution function of the Weibull distribution model; if the K-S test is not passed, or the end condition of the particle swarm algorithm is not satisfied, then the position of each particle in the state search space is updated through the particle swarm algorithm, the current parameters of the Weibull distribution model are output, the Weibull distribution model is updated using the current parameters, and the step of performing the K-S test using the empirical distribution function and the fault probability distribution function of the Weibull distribution model based on the fault data set of the servo system is returned; if the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, then the current parameters output by the particle swarm algorithm are used as the optimal parameters of the Weibull distribution model, so that the Weibull distribution model can use the optimal parameters to evaluate the reliability of the servo system. Through the embodiments of the present application, the K-S test compares the difference between the fault probability distribution function of the Weibull distribution model to be solved and the empirical distribution function as the comparison target through a statistic, and it focuses on the maximum difference of individuals, while the particle swarm algorithm uses the sum of squared error function as the objective function, and the sum of squared error focuses on minimizing the overall error. By combining the two, when the K-S test is passed and the end condition of the particle swarm algorithm is also satisfied, the parameters of the Weibull distribution model obtained by the solution are more accurate. When the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, the parameters of the Weibull distribution model are updated through the particle swarm algorithm. Solving the parameters of the Weibull distribution model through the particle swarm algorithm has a faster convergence speed, thus improving the accuracy of the Weibull distribution model while also improving the parameter solution speed of the Weibull distribution model. Description of the Drawings

[0046] Figure 1 It is a schematic flowchart of an embodiment of the method for evaluating the reliability of a servo system based on the K-S test in the present application;

[0047] Figure 2 It is a Weibull distribution fitting diagram of an embodiment of the method for evaluating the reliability of a servo system based on the K-S test in the present application;

[0048] Figure 3 It is the fitness convergence curve of the particle swarm algorithm of an embodiment of the method for evaluating the reliability of a servo system based on the K-S test in the present application;

[0049] Figure 4 It is a comparison diagram of the reliability curves of two methods of an embodiment of the method for evaluating the reliability of a servo system based on the K-S test in the present application;

[0050] Figure 5 This is a comparison graph of the failure probability density curves of two methods in an embodiment of the servo system reliability evaluation method based on the K-S test of the present application;

[0051] Figure 6 This is a schematic diagram of the functional modules of an embodiment of the servo system reliability evaluation device based on the K-S test of the present application;

[0052] Figure 7 This is a schematic diagram of the hardware structure of the servo system reliability evaluation device based on the K-S test involved in the embodiment solution of the present application. Detailed implementation manners

[0053] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0054] To make the purpose, technical solution and advantages of the present application clearer, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0055] In a first aspect, an embodiment of the present application provides a servo system reliability evaluation method based on the K-S test.

[0056] In one embodiment, referring to Figure 1 , Figure 1 This is a schematic flowchart of an embodiment of the servo system reliability evaluation method based on the K-S test of the present application. As Figure 1 shown, the servo system reliability evaluation method based on the K-S test includes:

[0057] Step S10: Use the parameters of the Weibull distribution model as the particles of the particle swarm algorithm, use the sum of squared error function as the objective function of the particle swarm algorithm, solve through the particle swarm algorithm based on the fault data set of the servo system, output the current parameters of the Weibull distribution model, and update the Weibull distribution model with the current parameters.

[0058] In this embodiment, the particle swarm optimization algorithm is used to solve the parameters of the optimal Weibull distribution model. The particle swarm optimization algorithm is inspired by the foraging behavior of bird flocks and is a stochastic search algorithm that mimics this behavior. Although the bird flock does not know the exact location of the food and the direction and distance of its flight path are unpredictable, through the information sharing mechanism among the group, each bird can know how far it is from the food at this moment. The bird closest to the food will send out location information to guide the group to converge towards it, thereby updating the optimal location information. Through repeated updates, the group will eventually find the exact location of the food, that is, obtain the global optimal solution of the population, and the global optimal solution of the population is the parameters of the optimal Weibull distribution model.

[0059] According to different parameters, the Weibull distribution includes two forms: two-parameter Weibull and three-parameter Weibull. When the location parameter γ≠0, the Weibull distribution is a three-parameter Weibull distribution. Conversely, when the location parameter γ = 0, the Weibull distribution is a two-parameter Weibull distribution. In practical applications, since there is a possibility of equipment failure at the beginning of operation, generally γ = 0. At this time, the expression of the failure probability density function of the Weibull distribution is: where β is the shape parameter, η is the scale parameter, and t is the time parameter. The expression of the reliability function R(t) of the Weibull distribution is: The expression of the failure rate function λ(t) is: The relationship expression between the failure probability distribution function F(t) and the reliability function R(t) of the Weibull distribution is: F(t) = 1 - R(t).

[0060] The parameters of the Weibull distribution model to be solved by the particle swarm optimization algorithm include the shape parameter β and the scale parameter η. Taking the parameters of the Weibull distribution model as the particles of the particle swarm optimization algorithm, the particle swarm optimization algorithm is used for solution. Among them, the position of the particle is the coordinate in the state search space, representing a potential solution in the state search space. The position of each particle corresponds to a set of specific <shape parameter β, scale parameter η> values. That is, the particle swarm optimization algorithm can realize the two-dimensional solution of <shape parameter β, scale parameter η> of the Weibull distribution model through the change of the particle position. The population of the particle swarm optimization algorithm is the set of all particles, that is, the set of all potential solutions. The state search space is the value range of all potential solutions. The particles move and fly in the specified state search space. Each particle has an optimal solution, which is related to the position, flight speed, individual optimal solution, and global optimal solution of the particle at the previous moment. Through continuous update and iteration, the particle swarm optimization algorithm can quickly find the optimal solution that meets the given conditions. Taking the sum of squared errors function as the objective function of the particle swarm optimization algorithm, the sum of squared errors focuses on minimizing the overall error. Solving through the particle swarm optimization algorithm has a faster convergence speed, thereby improving the accuracy of the Weibull distribution model and the parameter solution speed of the Weibull distribution model at the same time.

[0061] To solve the parameters of the Weibull distribution model by the particle swarm optimization algorithm, it is first necessary to initialize the particle swarm optimization algorithm, such as initializing the number of particles in the population, the speed range of particle movement, the position range of particle movement, the learning factor, and the inertia factor. For example, the number of particles in the population is initialized to 50, and the initial position and initial speed of each particle are randomly initialized. After initialization, the particle swarm optimization algorithm is used for solution. The particle swarm optimization algorithm takes the sum of squared errors function as the objective function. Based on the fault data set of the servo system through the objective function, the sum of squared errors of each particle is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The sum of squared errors of each particle is used as the fitness of each particle in the population. The fitness reflects the quality of the solution corresponding to each particle. The particle swarm optimization algorithm records the individual optimal solution and the corresponding fitness of each particle, as well as records the global optimal solution and the corresponding fitness of the population, as the basis for the subsequent update and iteration solution of the particle swarm optimization algorithm. The global optimal solution obtained after initialization can be output as the current parameter of the Weibull distribution model, and the Weibull distribution model is updated using the current parameter, so that the Weibull distribution model uses the latest shape parameter β and scale parameter η for the subsequent K-S test. Among them, the fault data set of the servo system can be obtained by collecting the fault data of the servo system. The specific fault data set of the servo system can be referred to Table 1 shown below.

[0062] Table 1.

[0063] Number of the fault Number of days from the initial time Number of the fault Number of days from the initial time Number of the fault Number of days from the initial time 1 22.6 9 75.3 17 92.3 2 48.5 10 76 18 95.1 3 56.3 11 79.9 19 98.5 4 58.9 12 80.8 20 102.5 5 59.2 13 82.6 21 103.3 6 64.5 14 84.3 22 110.2 7 66.8 15 85.9 23 111.1 8 67.7 16 86.3 24 123.4

[0064] Step S20: Based on the fault data set of the servo system, perform the K-S test using the empirical distribution function and the fault probability distribution function of the Weibull distribution model.

[0065] In this embodiment, the K-S test is a commonly used non-parametric statistical method, which aims to compare whether two sets of samples are from the same distribution. The K-S test compares the difference between the fault probability distribution function of the Weibull distribution model to be solved and the empirical distribution function as the comparison target through a statistic. It focuses on the maximum difference between individuals through the K-S statistic. The Weibull distribution model obtained through the K-S test is more accurate.

[0066] When performing the K-S test, for each sample data in the fault data set of the servo system shown in Table 1, the empirical distribution function F n (t i)The empirical distribution value corresponding to each sample data is calculated. For each sample data in the fault dataset of the servo system shown in Table 1, the predicted value corresponding to each sample data can also be calculated through the failure probability distribution function of the Weibull distribution model. Then, the K-S statistic is calculated based on the predicted value and the empirical distribution value corresponding to each sample data. Furthermore, the K-S test is performed based on the K-S statistic. Among them, each sample data t in the fault dataset of the servo system i The corresponding relationship and data with the empirical distribution value can be referred to Table 2 shown below.

[0067] Table 2.

[0068] Serial number i <![CDATA[t i > <![CDATA[F n (t i )]]> Serial number i <![CDATA[t i > <![CDATA[F n (t i )]]> Serial number i <![CDATA[t i > <![CDATA[F n (t i )]]> 1 22.6 0.0417 9 75.3 0.3750 17 92.3 0.7083 2 48.5 0.0833 10 76 0.4167 18 95.1 0.7500 3 57.3 0.1250 11 79.9 0.4583 19 98.5 0.7917 4 58.9 0.1667 12 80.8 0.5000 20 102.5 0.8333 5 59.2 0.2083 13 82.6 0.5417 21 103.3 0.8750 6 64.5 0.2500 14 84.3 0.5833 22 110.2 0.9167 7 66.8 0.2917 15 85.9 0.6250 23 111.1 0.9583 8 67.7 0.3333 16 86.3 0.6667 24 123.4 1

[0069] Referring to Figure 2 , Figure 2 is the Weibull distribution fitting graph of an embodiment of the servo system reliability evaluation method based on the K-S test in this application. As Figure 2 shown, the Weibull distribution fitting graph is drawn based on the data shown in Table 2 to determine whether the fault dataset of the servo system conforms to the Weibull distribution model. It can be seen from Figure 2 that each sample data in the fault dataset of the servo system is basically on the Weibull distribution fitting line, indicating that the fault dataset of the servo system conforms to the Weibull distribution.

[0070] Step S30: If the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, update the position of each particle in the state search space through the particle swarm algorithm, output the current parameters of the Weibull distribution model, update the Weibull distribution model with the current parameters, and return to execute the step of performing the K-S test using the empirical distribution function and the failure probability distribution function of the Weibull distribution model based on the fault dataset of the servo system.

[0071] In this embodiment, if the K-S test fails, it indicates that the maximum individual difference between the solved Weibull distribution model and the empirical distribution function is still relatively large, and it is necessary to update the position of each particle in the state search space through the particle swarm optimization algorithm. The particle swarm optimization algorithm uses the sum of squared errors function as the objective function, and the sum of squared errors function focuses on minimizing the overall error. If the termination condition of the particle swarm optimization algorithm is not met, it means that the overall error between the solved Weibull distribution model and the empirical distribution function is still relatively large. Therefore, in both cases where the K-S test fails or the termination condition of the particle swarm optimization algorithm is not met, it is necessary to continue to update and iterate through the particle swarm optimization algorithm to solve for the parameters of a better Weibull distribution model. The particle swarm optimization algorithm automatically performs position optimization by updating the position of each particle in the state search space. By updating the individual optimal solution of each particle and the global optimal solution of the population, the particle swarm optimization algorithm will output the global optimal solution of the population after this update iteration as the current parameter of the Weibull distribution model, use the current parameter to update the Weibull distribution model, return to step S20, and the Weibull distribution model continues to perform the K-S test using the latest shape parameter β and scale parameter η.

[0072] Step S40, if the K-S test is passed and the termination condition of the particle swarm optimization algorithm is met, then use the current parameter output by the particle swarm optimization algorithm as the best parameter of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system using the best parameter.

[0073] In this embodiment, if the K-S test is passed and the termination condition of the particle swarm optimization algorithm is met, it means that the global optimal solution of the population output by continuously updating and iterating through the particle swarm optimization algorithm can not only meet the requirement of the maximum individual difference between the empirical distribution function and the failure probability distribution function of the Weibull distribution model, but also meet the requirement of the overall error between the empirical distribution function and the failure probability distribution function of the Weibull distribution model. By combining the K-S test and the sum of squared errors, the accuracy of the solved Weibull distribution model can be greatly improved.

[0074] In this embodiment, the particle swarm algorithm can be initialized first. After initialization, the particle swarm algorithm is used for solving. The particle swarm algorithm takes the sum of squared errors function as the objective function. Based on the fault data set of the servo system through the objective function, the sum of squared errors of each particle is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The globally optimal solution obtained after initialization and solution can be output as the current parameter of the Weibull distribution model, and the Weibull distribution model is updated using the current parameter, so that the Weibull distribution model uses the latest shape parameter β and scale parameter η for subsequent K-S tests. The K-S test compares the difference between the fault probability distribution function of the Weibull distribution model to be solved and the empirical distribution function as the comparison target through a statistic. It focuses on the maximum difference between individuals of the two through the K-S statistic. Further, if the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, it is necessary to continue to update and iteratively solve for better parameters of the Weibull distribution model through the particle swarm algorithm. The particle swarm algorithm automatically performs position optimization by updating the position of each particle in the state search space. By updating the individual optimal solution of each particle and the globally optimal solution of the population, the particle swarm algorithm will output the globally optimal solution of the population after this update iteration as the current parameter of the Weibull distribution model, update the Weibull distribution model using the current parameter, and return to execute step S20. The Weibull distribution model continues to perform the K-S test using the latest shape parameter β and scale parameter η. If the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, it indicates that the globally optimal solution of the population output by continuously updating and iterating through the particle swarm algorithm can not only meet the requirement of the maximum difference between individuals of the empirical distribution function and the fault probability distribution function of the Weibull distribution model, but also meet the requirement of the overall error between the empirical distribution function and the fault probability distribution function of the Weibull distribution model. By combining the K-S test and the sum of squared errors, the accuracy of the solved Weibull distribution model can be greatly improved, and the parameter solution speed of the Weibull distribution model solved by the particle swarm algorithm is faster, thus improving the accuracy of the Weibull distribution model while also improving the parameter solution speed of the Weibull distribution model.

[0075] Further, in one embodiment, step S20 includes:

[0076] Based on the fault data set of the servo system, the K-S statistic is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0077] According to the sample size of the fault data set and the preset significance level, the corresponding maximum absolute difference is determined;

[0078] If the K-S statistic is less than the maximum absolute difference, it is determined that the test result is passing the K-S test;

[0079] If the K-S statistic is greater than or equal to the maximum absolute difference, it is determined that the test result fails the K-S test.

[0080] In this embodiment, when performing the K-S test, for each sample data in the fault dataset of the servo system shown in Table 1, the empirical distribution value corresponding to each sample data can be calculated through the empirical distribution function F n (t i ) For each sample data in the fault dataset of the servo system shown in Table 1, the predicted value corresponding to each sample data can also be calculated through the failure probability distribution function of the Weibull distribution model. Then, based on the predicted value and the empirical distribution value corresponding to each sample data, the K-S statistic is calculated. Furthermore, according to the K-S test theory, the maximum absolute difference corresponding to different significance levels is proposed. By comparing the calculated K-S statistic with the maximum absolute difference corresponding to different significance levels, it is determined whether the K-S test passes. Among them, the corresponding relationship and data between the sample number, significance level, and maximum absolute difference of different fault datasets can be referred to Table 3.

[0081] Table 3.

[0082]

[0083] Furthermore, in one embodiment, calculating the K-S statistic based on the fault dataset of the servo system according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model includes:

[0084] Based on the fault dataset of the servo system, the K-S statistic is calculated through Formula 1 according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model. Formula 1 is:

[0085]

[0086] Among them, D is the K-S statistic, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample number of the fault dataset, t i is the i-th sample data of the fault dataset, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

[0087] In this embodiment, for each sample data in the fault dataset of the servo system shown in Table 1, the empirical distribution function F n (t i)Calculate the empirical distribution value corresponding to each sample data. For each sample data in the fault dataset of the servo system shown in Table 1, the fault probability distribution function F0(t) of the Weibull distribution model can also be used. i )Calculate the predicted value corresponding to each sample data. By taking the maximum value of the difference between the predicted value and the empirical distribution value of all sample data through the above formula (1), the K-S statistic D is calculated.

[0088] Furthermore, in one embodiment, updating the position of each particle through the particle swarm algorithm and outputting the current parameters of the Weibull distribution model includes:

[0089] Update the position of each particle through formula (2);

[0090] For each particle, use the updated particle to update the Weibull distribution model. Based on the fault dataset of the servo system, calculate the sum of squared errors of each updated particle through the objective function;

[0091] If the sum of squared errors of the updated particle is less than the sum of squared errors before the update, update the updated particle to the individual optimal solution of the particle;

[0092] If the sum of squared errors corresponding to the individual optimal solution of the updated particle is less than the sum of squared errors corresponding to the global optimal solution, update the individual optimal solution of the updated particle to the global optimal solution;

[0093] Output the global optimal solution as the current parameters of the Weibull distribution model;

[0094] Among them, formula (2) is:

[0095] x i (t + 1) = x i (t) + v i (t + 1), v i (t + 1) = ωv i (t) + c1r1(p i -x i (t)) + c2r2(p g -x i (t));

[0096] The objective function is determined according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The formula of the objective function is:

[0097]

[0098] Among them, x i (t + 1) is the updated position of particle i, x i (t) is the position of particle i before the update, vi (t + 1) is the updated velocity of particle i, v i (t) is the velocity of particle i before update, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers between 0 and 1, p i is the individual optimal solution of particle i, p g is the global optimal solution of the particle, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample size of the failure data set, t i is the i-th sample data of the failure data set, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

[0099] In this embodiment, it can be seen from Formula 2 that in one iteration of the particle swarm algorithm, the position of each particle in the state search space will be updated. The updated position of each particle i is related to the position before update, the velocity before update, the velocity after update, the individual optimal solution, and the global optimal solution. By updating the position of the particle in the state search space, the global optimal solution of the population is solved. The position of each particle in the state search space represents a set of values of <shape parameter β, scale parameter η> of the Weibull distribution model. For each particle, the updated Weibull distribution model is used with each updated particle. Based on the failure data set of the servo system, the sum of squared errors after the update of each particle is calculated through the objective function. It can be seen from the formula of the objective function that the objective function is determined according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model. The sum of squared errors of each particle is used as the fitness of each particle in the population. The fitness reflects the quality of the solution corresponding to each particle. If the sum of squared errors after the update of the particle is less than the sum of squared errors before the update, the updated particle is updated to the individual optimal solution of the particle. If the sum of squared errors corresponding to the individual optimal solution of the updated particle is less than the sum of squared errors corresponding to the global optimal solution, the individual optimal solution of the updated particle is updated to the global optimal solution. Thus, it can be seen that the update iteration objective of the particle swarm algorithm is to make the sum of squared errors corresponding to the individual optimal solution smaller and to make the sum of squared errors corresponding to the global optimal solution smaller. Finally, the global optimal solution of the population is output as the current parameter of the Weibull distribution model, that is, the particle swarm algorithm outputs the global optimal solution after this update iteration, and the sum of squared errors corresponding to the global optimal solution is the smallest. Refer to Figure 3 , Figure 3 is the fitness convergence curve of the particle swarm algorithm for an embodiment of the servo system reliability evaluation method based on the K - S test in this application, as Figure 3As shown, when using the particle swarm optimization algorithm to solve, the global optimal solution can be obtained at the 70th iteration. At this time, the shape parameter β = 4.08771, the scale parameter η = 89.662, and the corresponding fitness value is 0.019677, that is, the minimum squared error is 0.019677.

[0100] Further, in one embodiment, when the sum of squared errors corresponding to the current parameters is greater than or equal to the traditional sum of squared errors, and the number of iterations of the particle swarm optimization algorithm is less than or equal to the preset number of times, it is determined that the end condition of the particle swarm optimization algorithm is not satisfied. When the sum of squared errors corresponding to the current parameters is less than the traditional sum of squared errors, or the number of iterations of the particle swarm optimization algorithm is greater than the preset number of times, it is determined that the end condition of the particle swarm optimization algorithm is satisfied.

[0101] In this embodiment, the end conditions of the particle swarm optimization algorithm mainly include two aspects. One is the sum of squared errors corresponding to the current parameters, and the other is the number of iterations. When the sum of squared errors corresponding to the current parameters is less than the traditional sum of squared errors, it indicates that the parameters of the Weibull distribution model obtained by this method are better than those obtained by the traditional method. Or when the number of iterations of the particle swarm optimization algorithm is greater than the preset number of times, since the global optimal solution output by each iteration of the particle swarm optimization algorithm is not worse than the global optimal solution output by the previous iteration, therefore, after the number of iterations of the particle swarm optimization algorithm reaches the preset number of times, it can be determined that the end condition of the particle swarm optimization algorithm is satisfied.

[0102] Further, in one embodiment, before using the parameters of the Weibull distribution model as the particles of the particle swarm optimization algorithm, it includes:

[0103] Solving the parameters of the Weibull distribution model using the traditional method, and updating the Weibull distribution model with the obtained parameters;

[0104] Based on the fault data set of the servo system, the traditional sum of squared errors is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model.

[0105] In this embodiment, the traditional method such as the least squares method. To facilitate the comparison of the accuracy of the Weibull distribution model solved by this method and the Weibull distribution model solved by the traditional method, the parameters of the Weibull distribution model can be solved in advance using the traditional method, and the obtained parameters can be used to update the Weibull distribution model. Based on the fault data set of the servo system, the traditional sum of squared errors is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The sum of squared errors obtained by the traditional method can be compared with the sum of squared errors obtained by this method, and the sum of squared errors obtained by the traditional method can be used as the iteration end condition of the particle swarm optimization algorithm.

[0106] Further, in one embodiment, after using the current parameters output by the particle swarm algorithm as the optimal parameters of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system, the following steps are included:

[0107] Based on the fault data set of the servo system, the root mean square error is calculated according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model for evaluating the accuracy of the Weibull distribution model using the root mean square error.

[0108] In this embodiment, the root mean square error, also known as the standard error, is used to measure the deviation between the observed value and the true value, and can well reflect the deviation between the fault probability distribution function of the Weibull distribution model and the empirical distribution function. Among them, the calculation formula of the root mean square error is: Among them, RMSE is the root mean square error, F n (t i ) is the empirical distribution function, F0(t i ) is the fault probability distribution function of the Weibull distribution model, n is the sample size of the fault data set, and t i is the i-th sample data of the fault data set.

[0109] For traditional methods such as the least squares method, the parameters of the Weibull distribution model obtained by the traditional method and this method and the corresponding root mean square error can be referred to Table 4.

[0110] Table 4.

[0111] Estimation method Parameter values of the Weibull distribution model Root mean square error Traditional least squares method β = 3.2374, η = 90.7905 0.003 This method β = 40.0877, η = 89.662 0.0008

[0112] It can be seen from Table 4 that the root mean square error of the Weibull distribution parameters obtained by this method is smaller, which indicates that the parameters of the Weibull distribution model calculated by this method are more accurate.

[0113] Substitute the shape parameter β and scale parameter η of the Weibull distribution model obtained by the traditional least squares method and this method into the calculation formula of the reliability function respectively, and the comparison diagram of the reliability curves of the two methods can be obtained. Refer to Figure 4 , Figure 4 is the comparison diagram of the reliability curves of the two methods in one embodiment of the servo system reliability evaluation method based on the K-S test of the present application. As Figure 4 shown, the reliability curve obtained by this method is closer to the reliability curve of the source data. Therefore, it can also be explained that the parameters of the Weibull distribution model calculated by this method are more accurate.

[0114] Furthermore, as the running time increases, the reliability of the servo system gradually decreases, and the decreasing speed first increases and then decreases, which conforms to the reliability change trend of mechanical equipment. Substituting the shape parameter β and scale parameter η of the Weibull distribution model obtained by the traditional least squares method and this method into the calculation formula of the failure probability density function respectively, a comparison diagram of the failure probability density curves of the two methods can be obtained. Referring to Figure 5 , Figure 5 is a comparison diagram of the failure probability density curves of two methods in an embodiment of the servo system reliability evaluation method based on the K-S test of the present application. As shown in Figure 5 , the failure probability density curves of the servo system obtained by the two methods are extremely close in shape, both showing a trend of rising first and then falling, and the peaks of both appear at about 63 days. Further observing these two curves, it can be found that the peak of the failure probability density curve obtained by this method is lower, and the failure events are distributed in a wider range. To reduce the number of failures of the servo system and improve its availability, inspections should be strengthened near the peak of the failure probability density to ensure that possible faults can be detected and processed in time.

[0115] In a second aspect, an embodiment of the present application further provides a servo system reliability evaluation device based on the K-S test.

[0116] In one embodiment, referring to Figure 6 , Figure 6 is a schematic diagram of the functional modules of an embodiment of the servo system reliability evaluation device based on the K-S test of the present application. As shown in Figure 6 , the servo system reliability evaluation device based on the K-S test includes:

[0117] The initial solution module 10 is configured to use the parameters of the Weibull distribution model as the particles of the particle swarm algorithm, use the sum of squared error function as the objective function of the particle swarm algorithm, and perform a solution through the particle swarm algorithm based on the fault data set of the servo system, output the current parameters of the Weibull distribution model, and update the Weibull distribution model with the current parameters;

[0118] The K-S test module 20 is configured to perform a K-S test based on the fault data set of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0119] The update and solution module 30 is configured to, if the K-S test fails or the end condition of the particle swarm algorithm is not satisfied, update the position of each particle in the state search space through the particle swarm algorithm, output the current parameters of the Weibull distribution model, update the Weibull distribution model with the current parameters, and return to execute the step of performing a K-S test based on the fault data set of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0120] An output module 40, configured to, if the K-S test is passed and the end condition of the particle swarm optimization algorithm is satisfied, use the current parameters output by the particle swarm optimization algorithm as the optimal parameters of the Weibull distribution model, so that the Weibull distribution model can use the optimal parameters to evaluate the reliability of the servo system.

[0121] Further, in an embodiment, the K-S test module 20 includes:

[0122] A K-S statistic calculation unit, configured to calculate a K-S statistic based on the fault data set of the servo system according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model;

[0123] A first determination unit, configured to determine a corresponding maximum absolute difference according to the sample number of the fault data set and a preset significance level;

[0124] A second determination unit, configured to determine that the test result passes the K-S test if the K-S statistic is less than the maximum absolute difference;

[0125] A third determination unit, configured to determine that the test result fails the K-S test if the K-S statistic is greater than or equal to the maximum absolute difference.

[0126] Further, in an embodiment, the K-S statistic calculation unit is configured to:

[0127] Calculate a K-S statistic based on the fault data set of the servo system according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model through Formula 1, and the Formula 1 is:

[0128]

[0129] where D is the K-S statistic, F0(t i ) is the fault probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample number of the fault data set, t i is the i-th sample data of the fault data set, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

[0130] Further, in an embodiment, the servo system reliability evaluation device based on the K-S test further includes a particle swarm optimization algorithm update module, configured to:

[0131] Update the position of each particle through Formula 2;

[0132] For each particle, update the Weibull distribution model using the updated particle. Based on the fault data set of the servo system, calculate the sum of squared errors of each updated particle through the objective function;

[0133] If the sum of squared errors of the updated particle is less than the sum of squared errors before the update, update the updated particle to the individual optimal solution of the particle;

[0134] If the sum of squared errors corresponding to the individual optimal solution of the updated particle is less than the sum of squared errors corresponding to the global optimal solution, update the individual optimal solution of the updated particle to the global optimal solution;

[0135] Output the global optimal solution as the current parameter of the Weibull distribution model;

[0136] Among them, the formula two is:

[0137] x i (t + 1) = x i (t) + v i (t + 1), v i (t + 1) = ωv i (t) + c1r1(p i -x i (t)) + c2r2(p g -x i (t));

[0138] The objective function is determined according to the empirical distribution function and the failure probability distribution function of the Weibull distribution model. The formula of the objective function is:

[0139]

[0140] Among them, x i (t + 1) is the updated position of particle i, x i (t) is the position of particle i before the update, v i (t + 1) is the updated velocity of particle i, v i (t) is the velocity of particle i before the update, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers between 0 and 1, p i is the individual optimal solution of particle i, p g is the global optimal solution of the particle, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample size of the fault data set, t i$x_i$ is the $i$-th sample data of the fault data set, $\beta$ is the shape parameter of the Weibull distribution model, and $\eta$ is the scale parameter of the Weibull distribution model.

[0141] Further, in one embodiment, when the sum of squared errors corresponding to the current parameters is greater than or equal to the traditional sum of squared errors, and the number of iterations of the particle swarm optimization algorithm is less than or equal to the preset number, it is determined that the end condition of the particle swarm optimization algorithm is not satisfied. When the sum of squared errors corresponding to the current parameters is less than the traditional sum of squared errors, or the number of iterations of the particle swarm optimization algorithm is greater than the preset number, it is determined that the end condition of the particle swarm optimization algorithm is satisfied.

[0142] Further, in one embodiment, the servo system reliability evaluation device based on the K-S test further includes a traditional solution module for:

[0143] Solving the parameters of the Weibull distribution model using a traditional method, and updating the Weibull distribution model with the solved parameters;

[0144] Based on the fault data set of the servo system, calculating the traditional sum of squared errors according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model.

[0145] Further, in one embodiment, the servo system reliability evaluation device based on the K-S test further includes a root mean square error evaluation module for:

[0146] Based on the fault data set of the servo system, calculating the root mean square error according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model, so as to evaluate the accuracy of the Weibull distribution model using the root mean square error.

[0147] Wherein, the function implementation of each module in the above servo system reliability evaluation device based on the K-S test corresponds to each step in the above embodiment of the servo system reliability evaluation method based on the K-S test, and its function and implementation process will not be elaborated here one by one.

[0148] In a third aspect, an embodiment of the present application provides a servo system reliability evaluation device based on the K-S test.

[0149] Refer to Figure 7 , Figure 7 is a schematic hardware structure diagram of the servo system reliability evaluation device based on the K-S test involved in the solution of the embodiment of the present application. In the embodiment of the present application, the servo system reliability evaluation device based on the K-S test may include a processor, a memory, a communication interface, and a communication bus.

[0150] Among them, the communication bus can be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0151] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces, etc., which are used to implement the interconnection of components inside the servo system reliability evaluation device based on the K-S test, as well as interfaces for implementing the interconnection between the servo system reliability evaluation device based on the K-S test and other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, a fiber optic interface, an ATM interface, etc.; the user device can be a display, a keyboard, etc.

[0152] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical memory, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0153] The processor can be a general-purpose processor, which can call the servo system reliability evaluation program stored in the memory and execute the servo system reliability evaluation method provided in the embodiments of the present application. For example, the general-purpose processor can be a central processing unit (CPU). Among them, the method executed when the servo system reliability evaluation program based on the K-S test is called can refer to the various embodiments of the servo system reliability evaluation method based on the K-S test in the present application, which will not be elaborated here.

[0154] Those skilled in the art can understand that Figure 7 the hardware structure shown in

[0155] does not constitute a limitation to the present application, and may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0156] The readable storage medium of the present application stores a servo system reliability evaluation program based on the K-S test. When the servo system reliability evaluation program based on the K-S test is executed by a processor, the steps of the servo system reliability evaluation method as described above are implemented.

[0157] Among them, the method implemented when the servo system reliability assessment program based on the K-S test is executed can refer to the various embodiments of the servo system reliability assessment method based on the K-S test in this application, which will not be elaborated here.

[0158] It should be noted that the serial numbers of the embodiments of this application above are only for description and do not represent the superiority or inferiority of the embodiments.

[0159] The terms "including" and "having" and any variations thereof in the specification, claims and above-mentioned drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices. The descriptions with terms such as "first", "second" and "third" are used to distinguish different objects, etc., which do not represent a sequence, nor do they limit that "first", "second" and "third" are different types.

[0160] In the description of the embodiments of this application, "exemplary", "for example" or "for instance" are used to mean for example, illustration or explanation. Any embodiment or design solution described as "exemplary", "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary", "for example" or "for instance" is intended to present relevant concepts in a specific manner.

[0161] In the description of the embodiments of this application, unless otherwise specified, " / " means "or". For example, A / B can mean A or B; "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "a plurality of" means two or more than two.

[0162] In some processes described in the embodiments of this application, a plurality of operations or steps appear in a specific order, but it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of this application or may be executed in parallel. The serial numbers of the operations are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0163] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium as described above (such as ROM / RAM, magnetic disk, optical disc), and includes several instructions for causing a terminal device to execute the methods described in various embodiments of the present application.

[0164] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A servo system reliability evaluation method based on the K-S test, characterized in that, The servo system reliability evaluation method based on the K-S test includes: Using the parameters of the Weibull distribution model as the particles of the particle swarm algorithm, and using the sum of squared error function as the objective function of the particle swarm algorithm. Based on the fault data set of the servo system, solve through the particle swarm algorithm, output the current parameters of the Weibull distribution model, and update the Weibull distribution model with the current parameters; Based on the fault data set of the servo system, conduct the K-S test using the empirical distribution function and the fault probability distribution function of the Weibull distribution model; If the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, update the position of each particle in the state search space through the particle swarm algorithm, output the current parameters of the Weibull distribution model, update the Weibull distribution model with the current parameters, and return to execute the step of conducting the K-S test using the empirical distribution function and the fault probability distribution function of the Weibull distribution model based on the fault data set of the servo system; If the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, use the current parameters output by the particle swarm algorithm as the optimal parameters of the Weibull distribution model, so that the Weibull distribution model can use the optimal parameters to evaluate the reliability of the servo system.

2. The reliability evaluation method of the servo system based on the K-S test according to claim 1, wherein The conducting of the K-S test using the empirical distribution function and the fault probability distribution function of the Weibull distribution model based on the fault data set of the servo system includes: Based on the fault data set of the servo system, calculate the K-S statistic according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model; Determine the corresponding maximum absolute difference according to the sample number of the fault data set and the preset significance level; If the K-S statistic is less than the maximum absolute difference, determine that the test result is passing the K-S test; If the K-S statistic is greater than or equal to the maximum absolute difference, determine that the test result is not passing the K-S test.

3. The reliability evaluation method of the servo system based on the K-S test according to claim 2, wherein The calculating of the K-S statistic according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model based on the fault data set of the servo system includes: Based on the fault data set of the servo system, calculate the K-S statistic through formula one according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The formula one is: where D is the K-S statistic, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample size of the failure data set, t i is the i-th sample data of the failure data set, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

4. The reliability evaluation method of the servo system based on the K-S test according to claim 1, wherein The updating of the position of each particle through the particle swarm algorithm and outputting the current parameters of the Weibull distribution model includes: Update the position of each particle through formula two; For each particle, update the Weibull distribution model using the updated each particle, and calculate the sum of squared errors after update for each particle based on the fault data set of the servo system through the objective function; If the sum of squared errors after particle update is less than the sum of squared errors before update, update the updated particle to the individual optimal solution of the particle; If the sum of squared errors corresponding to the individual optimal solution of the updated particle is less than the sum of squared errors corresponding to the global optimal solution, update the individual optimal solution of the updated particle to the global optimal solution; Output the global optimal solution as the current parameters of the Weibull distribution model; Among them, the formula two is: x i (t + 1) = x i (t) + v i (t + 1), v i (t + 1) = ωv i (t) + c1r1(p i -x i (t)) + c2r2(p g -x i (t)); The objective function is determined according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model. The formula of the objective function is: where x i (t + 1) is the updated position of particle i, x i (t) is the position of particle i before update, v i (t + 1) is the updated velocity of particle i, v i (t) is the velocity of particle i before update, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers between 0 and 1, p i is the individual optimal solution of particle i, p g is the global optimal solution of the particle, F0(t i ) is the failure probability distribution function of the Weibull distribution model, F n (t i ) is the empirical distribution function, R0(t i ) is the reliability function of the Weibull distribution model, n is the sample size of the failure data set, t i is the i-th sample data of the failure data set, β is the shape parameter of the Weibull distribution model, and η is the scale parameter of the Weibull distribution model.

5. The reliability evaluation method of the servo system based on the K-S test according to claim 4, characterized in that When the sum of squared errors corresponding to the current parameter is greater than or equal to the traditional sum of squared errors, and the number of iterations of the particle swarm algorithm is less than or equal to the preset number of times, it is determined that the end condition of the particle swarm algorithm is not satisfied. When the sum of squared errors corresponding to the current parameter is less than the traditional sum of squared errors, or the number of iterations of the particle swarm algorithm is greater than the preset number of times, it is determined that the end condition of the particle swarm algorithm is satisfied.

6. The servo system reliability evaluation method based on the K-S test according to claim 5, characterized in that, Before using the parameters of the Weibull distribution model as the particles of the particle swarm algorithm, it includes: Solving the parameters of the Weibull distribution model using traditional methods, and updating the Weibull distribution model using the obtained parameters; Based on the fault dataset of the servo system, calculating the traditional sum of squared errors according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model.

7. The reliability evaluation method of the servo system based on the K-S test according to claim 1, characterized in that, After using the current parameter output by the particle swarm algorithm as the best parameter of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system using the best parameter, it includes: Based on the fault dataset of the servo system, calculating the root mean square error according to the empirical distribution function and the fault probability distribution function of the Weibull distribution model for evaluating the accuracy of the Weibull distribution model using the root mean square error.

8. A servo system reliability evaluation device based on the K-S test, characterized in that The servo system reliability evaluation device based on the K-S test includes: The initial solution module is used to use the parameters of the Weibull distribution model as the particles of the particle swarm algorithm, use the sum of squared errors function as the objective function of the particle swarm algorithm, solve through the particle swarm algorithm based on the fault dataset of the servo system, output the current parameter of the Weibull distribution model, and update the Weibull distribution model using the current parameter; The K-S test module is used to perform the K-S test based on the fault dataset of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model; The update solution module is used to, if the K-S test is not passed or the end condition of the particle swarm algorithm is not satisfied, update the position of each particle in the state search space through the particle swarm algorithm, output the current parameter of the Weibull distribution model, update the Weibull distribution model using the current parameter, and return to execute the step of performing the K-S test based on the fault dataset of the servo system using the empirical distribution function and the fault probability distribution function of the Weibull distribution model; The output module is used to, if the K-S test is passed and the end condition of the particle swarm algorithm is satisfied, use the current parameter output by the particle swarm algorithm as the best parameter of the Weibull distribution model for the Weibull distribution model to evaluate the reliability of the servo system using the best parameter.

9. A servo system reliability evaluation device based on the K-S test, characterized in that, The servo system reliability evaluation device based on the K-S test includes a processor, a memory, and a servo system reliability evaluation program based on the K-S test stored on the memory and executable by the processor. When the servo system reliability evaluation program based on the K-S test is executed by the processor, it implements the steps of the servo system reliability evaluation method based on the K-S test as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that, The readable storage medium stores a servo system reliability evaluation program based on the K-S test. When the servo system reliability evaluation program based on the K-S test is executed by a processor, the steps of the servo system reliability evaluation method based on the K-S test according to any one of claims 1 to 7 are implemented.

Images