A method for calculating the life of consumable products based on an improved wolf pack algorithm

Abstract

The present invention relates to a method for calculating the life of consumable products based on an improved wolf pack algorithm, belonging to the technical field of life calculation methods. In order to solve the problem that the existing life calculation methods for consumable products in a new environment are vulnerable to interference and can only estimate simple parameters, and cannot be used for estimating complex parameter estimates, the life calculation method includes determining that the life of the consumable product follows a Weibull distribution; determining an objective function according to the Weibull distribution; using the improved wolf pack algorithm to solve the objective function to obtain the required parameters in the old and new environments; defining a conversion coefficient and calculating the conversion coefficient according to the required parameters; and calculating the life of the consumable product in the new environment using the conversion coefficient, which consists of five steps. The improved wolf pack algorithm has a fast convergence speed and high calculation accuracy; applying the improved wolf pack algorithm to the life conversion of consumable products in a new environment can estimate complex parameters accurately and stably.

Description

Technical Field

[0001] The present invention relates to a method for calculating the life of a consumable product, specifically a method for calculating the life of a consumable product based on an improved wolf pack algorithm, belonging to the technical field of life calculation methods. Background Art

[0002] The wolf pack algorithm is an optimal decision-making algorithm generated by simulating the hunting behavior of wolf packs. By observing the hunting behavior of wolf packs, the wolves in the wolf pack are divided into three types: the alpha wolf, the scout wolf, and the striker wolf. They work according to the following rules:

[0003] (1) Alpha wolf generation: Generate n wolves in the initial space, calculate the fitness value of each wolf, and the optimal one is the alpha wolf. If there are multiple ones, randomly specify;

[0004] (2) Roaming behavior: Select the alpha wolf and use the wolves with better fitness as scout wolves. The number is taken as an integer between where is the scout wolf ratio factor, that is, the proportion of alpha wolves in the wolf pack. The scout wolves roam according to the step size . If the fitness after roaming is better than that of the alpha wolf, replace the alpha wolf; otherwise, return to the original position and continue to roam until the maximum number of roaming times is reached. After the roaming ends, update the position of the scout wolf according to the optimal fitness during the roaming process. The formula for calculating the position of the scout wolf is as follows: ,

[0005] where W' is the new position after roaming, W is the position before roaming, and randA is a uniformly distributed random number in [-1, 1];

[0006] The roaming behavior of the scout wolves is to find local optimal solutions other than the position of the alpha wolf. This behavior is related to the ability of the algorithm to jump out of local optimal solutions. The number of explorations and the number of scout wolves are parameters that can be set. However, the method in the standard wolf pack algorithm that each scout wolf has to complete a fixed number of explorations in each iteration significantly occupies a large amount of time;

[0007] (3) Raiding behavior: The striker wolves are artificial wolves closer to the alpha wolf. The striker wolves raid towards the alpha wolf according to the step size . If the fitness during the raiding process is better than that of the alpha wolf, replace the alpha wolf until the distance from the alpha wolf meets the requirement. The formula for calculating the position of the striker wolf is as follows: ,

[0008] where randB is a uniformly distributed random number in [0, 1], represents the movement direction of the striker wolf towards the alpha wolf;

[0009] The raiding behavior is to change the positions of all individuals in the whole group. On the one hand, it can jump out of the local optimal solution to a certain extent. On the other hand, it is to gather enough individuals for the siege behavior to improve the accuracy of the solution. However, for the fierce wolves in the position of inferior solutions, it is of little significance to blindly launch a raid on the leading wolf in the hope of encountering a better solution during the raid;

[0010] (5) Siege behavior: The fierce wolves and exploring wolves carry out the siege step by step If the fitness during the siege is better than the original position, the position is updated; otherwise, the position remains unchanged. The position calculation formula after the siege is as follows: ,

[0011] (6) Survival of the fittest: After the siege, the worst X wolves are eliminated, and at the same time, X wolves are randomly generated;

[0012] In the above behaviors, the three step lengths respectively correspond to the three behaviors of wandering, raiding and siege. The relationship between the three step lengths is expressed by the following formula: ,

[0013] Among them, max and min are the boundaries of the optimization solution space.

[0014] Although the above wolf pack algorithm has the advantages of being intuitive and easy to understand, easy to implement, cooperation and competition, and dynamic update, and has been widely used in many fields, there are still some disadvantages. For complex large-scale optimization problems, the algorithm accuracy, the ability to jump out of the local optimal solution, etc. can be improved by adjusting parameters such as the number of wolf packs and the wandering step length. However, the biggest drawback is that the algorithm takes a long time. The life distribution of products and the life conversion relationship under multiple stresses cannot be directly measured by specific equipment. The common practice is to conduct life statistics on products. In engineering, the environmental parameters sensitive to products include: temperature, humidity, electrical parameters, vibration, load, etc. The values of these conditions are called stress levels.

[0015] Generally, high-reliability products will conduct corresponding accelerated stress life tests during production. For products whose life follows the Weibull distribution, its distribution parameters can be estimated by statistically analyzing their lives. The Weibull distribution is a probability distribution function proposed by the Swedish scientist Weibull in 1951. Due to its flexible performance in fitting random data, it has been widely used at home and abroad. Currently, the main Weibull parameter estimation methods include the maximum likelihood estimation method, the gray method, the approximation estimation method, etc. These methods are easily interfered and can only estimate simple parameters, and cannot be used for estimating complex parameters. Summary of the Invention

[0016] The object of the present invention is to solve the problem that the existing life calculation method of lossy products is vulnerable to interference in the new environment and can only estimate simple parameters, and cannot be used for estimating complex parameters. A life conversion method for lossy products based on an improved wolf pack algorithm is provided. The improved wolf pack algorithm has a fast convergence speed and high calculation accuracy; applying the improved wolf pack algorithm to the life conversion of lossy products in the new environment can estimate complex parameters.

[0017] To solve the above problems, the present application is implemented through the following technical solutions:

[0018] A life conversion method for lossy products based on an improved wolf pack algorithm, which is characterized in that it includes the following steps:

[0019] Step 1: Determine that the life of the lossy product follows a Weibull distribution;

[0020] Step 2: Determine the objective function according to the Weibull distribution in Step 1;

[0021] Step 3: Use the improved wolf pack algorithm to solve the objective function in Step 2 to obtain the parameters to be sought in the old and new environments;

[0022] Step 4: Define the conversion coefficient and calculate the conversion coefficient according to the parameters to be sought obtained in Step 3;

[0023] Step 5: Use the conversion coefficient in Step 4 to calculate the life of the lossy product in the new environment.

[0024] Further, the cumulative failure probability density of the lossy product in Step 1 conforms to the Weibull distribution: ;

[0025] Further, the objective function in Step 2 includes the objective function in a single environment, that is, the objective function when the environmental stress does not change, the objective function when the environmental stress changes once, and the objective function when the environmental stress changes in multiple stages;

[0026] Further, the objective function in a single environment in Step 2 is: , that is ,

[0027] In the formula: X is the observed value of the cumulative failure probability, t is the observed time point, F is the estimated value of the cumulative failure probability, m, are the shape parameter and scale parameter of the Weibull distribution to be estimated;

[0028] Further, the objective function when the environmental stress changes once in Step 2 is: , that is ,

[0029] Where: ,

[0030] is the cumulative failure probability in the new environment, for The probability of failure in the distribution curve is At the same time, As the product usage environment changes, Any time point after the product usage environment changes, is the shape parameter in the new environment, is the scale parameter in the new environment,

[0031] for - Each time point t within the time period i The cumulative failure probability of the products above;

[0032] Furthermore, the objective function when the environmental stress changes in multiple stages in step 2 is: ,

[0033] Where: : Time point The conversion time at stage k is: Calculated by the following formula: ,

[0034] The initial conditions are , reflecting the time relationship before the first stress change; is the scale parameter of the kth stage, is the shape parameter of the kth stage;

[0035] Furthermore, the step 3 specifically includes the following steps:

[0036] Step 3.1, initialize each parameter: set the parameter value range, that is, the solution space range: , ,

[0037] Is a vector, each number in the vector is the upper limit of the corresponding parameter to be estimated, Is a vector, each number in the vector is the lower limit of the corresponding parameter to be estimated;

[0038] Set the maximum number of iterations Tmax; set the step size adjustment rate ; Set the population size, initialize the coordinates of the wolf pack, and generate the wolf pack in the solution space according to a uniform distribution.

[0039] Step 3.2, Wolf pack recombination: Calculate the objective function value of each individual in the wolf pack , select the best one as the wolf king. If is a fierce wolf, otherwise it is a scout wolf;

[0040] In the formula: is the exploration position of the wolf king, is the position before exploration, is the wandering step size;

[0041] Step 3.3, Scout wolf wandering: The scout wolf performs the wandering behavior and calculates its objective function value . If the scout wolf is better than the objective function value of the wolf king, that is , eliminate the leading wolf and the fierce wolves, generate new wolves at the new position, and go to Step 3.2 until the maximum number of iterations is reached;

[0042] The scout wolf wanders according to ;

[0043] In the formula: is the new position after exploration, is the position before exploration, is a random number uniformly distributed in [-1, 1];

[0044] Step 3.4, Fierce wolf siege: The fierce wolves perform the siege behavior and calculate their objective function values . If the fierce wolves are better than the objective function value of the wolf king, that is , go to Step 3.2 until the maximum number of iterations is reached,

[0045] The fierce wolves perform the siege behavior according to ;

[0046] Step 3.5, Reach the maximum number of iterations, and output the coordinates of the final leading wolf;

[0047] Furthermore, in Step 3.2, the wandering step size : ,

[0048] In the formula: and are the boundaries of the solution space, is the step size adjustment rate;

[0049] Furthermore, the conversion coefficient in Step 4 is the ratio of the time used when the reliability in the new and old environments is equal: ,

[0050] where: R is the reliability equal in two environments. Inside the parentheses, i represents the old environment when the environmental stress undergoes 1 mutation or the i-th stage when the environmental stress undergoes K mutations, j represents the new environment when the environmental stress undergoes 1 mutation or the j-th stage when the environmental stress undergoes K mutations, and k≥2;

[0051] Calculate the conversion coefficient according to the shape parameter m and scale parameter μ obtained in step 3 ;

[0052] Further, the specific method of step 5 is: X （j） =Z* X （i） ,

[0053] where: X （j） is the life of the wear-out product in the new environment, and X （i） is the life of the wear-out product in the old environment.

[0054] The improved wolf pack calculation method of this application improves the wandering behavior of the exploring wolves as follows: change the behavior of the exploring wolves to explore only once in a single iteration loop instead of multiple times, the result rule of each round of iteration remains unchanged, and the calculation formula for the position of the exploring wolves remains unchanged; combine the raiding behavior and the siege behavior. Those within the radius near the alpha wolf are fierce wolves, and the rest are exploring wolves. The fierce wolves search for better solutions within the radius sphere according to the calculation formula for the position of the fierce wolves. Such improvement can improve the speed and accuracy of the algorithm;

[0055] The improved wolf pack algorithm can meet the calculation of the life of wear-out products when the environment changes: propose an accumulative failure model for multiple environmental stress change scenarios; at this time, the objective function is simplified to the superposition of the multi-stage Weibull model, maintaining compatibility with the single stress change model; this modeling method realizes the continuous prediction of the failure probability under complex environmental stress histories, providing theoretical support for the design of multi-stage accelerated life tests; the improved wolf pack algorithm has a better convergence speed and higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is the flow chart of the present invention;

[0057] Figure 2 is the flow chart of step 3;

[0058] Figure 3 is the speed comparison chart between the conventional wolf pack algorithm and this application;

[0059] Figure 4 is the cumulative failure probability data of the wear-out product in the old and new environments in Example 3;

[0060] Figure 5It is the prediction diagram of the cumulative failure probability of the wear-out product in Example 3 under the new environment. Specific Embodiment

[0061] The following refers to the accompanying drawings to give the specific embodiments of the present invention to further illustrate the composition of the present invention.

[0062] Example 1. A method for converting the life of a wear-out product based on an improved wolf pack algorithm, the specific process is as Figure 1 shown, including the following steps:

[0063] Step 1: Determine that the life of the wear-out product follows a Weibull distribution;

[0064] Step 2: Determine the objective function according to the Weibull distribution in Step 1;

[0065] Step 3: Use the improved wolf pack algorithm to solve the objective function in Step 2 to obtain the required parameters in the old and new environments;

[0066] Step 4: Define the conversion coefficient and calculate the conversion coefficient according to the required parameters obtained in Step 3;

[0067] Step 5: Calculate the life of the wear-out product in the new environment using the conversion coefficient in Step 4.

[0068] Among them, in Step 1, when the failure mechanism of the wear-out product remains unchanged, it is determined that the life X of the wear-out product follows a Weibull distribution with a shape parameter m and a scale parameter μ;

[0069] Cumulative probability density of the wear-out product: ;

[0070] Among them, the objective function in Step 2 includes the objective function in a single environment, that is, the objective function when the environmental stress does not change, the objective function when the environmental stress changes once, and the objective function when the environmental stress changes in multiple stages;

[0071] Among them, the objective function in a single environment is: , that is ,

[0072] In the formula: X is the observed value of the cumulative failure probability, t is the observed time point, F is the estimated value of the cumulative failure probability, m, are the shape parameter and scale parameter of the Weibull distribution to be estimated;

[0073] The objective function when the environmental stress changes once is: , that is ,

[0074] In the formula: ,

[0075] is the cumulative failure probability in the new environment, is the failure probability and at the same time point in the distribution curve, is the moment when the product usage environment changes, is any time point after the product usage environment changes, is the shape parameter in the new environment, is the scale parameter in the new environment,

[0076] is - the cumulative failure probability of the product at each time point t within the time period i ;

[0077] The objective function when the environmental stress changes in multiple stages is: ,

[0078] Assume that the environmental stress undergoes K mutations, forming K + 1 stages, and the parameters of each stage are and , and the objective function is defined as the sum of the absolute deviations between the observed failure probability and the model prediction value;

[0079] In the formula: : time point is the converted time at the k-th stage, is calculated by the following formula: ,

[0080] The initial condition is , reflecting the time relationship before the first stress change; is the scale parameter of the k-th stage, is the shape parameter of the k-th stage;

[0081] Among them, as Figure 2 shown, step 3 specifically includes the following steps:

[0082] Step 3.1. Initialize each parameter: Set the value range of the parameter, that is, the solution space range: , ,

[0083] is a vector, and each number in the vector is the upper limit of the corresponding parameter to be estimated, is a vector, and each number in the vector is the lower limit of the corresponding parameter to be estimated;

[0084] Set the maximum number of iterations Tmax; set the step size adjustment rate ; set the population size, initialize the coordinates of the wolf pack, and generate the wolf pack in the solution space according to a uniform distribution;

[0085] Step 3.2, Wolf pack reorganization: Calculate the objective function value of each individual in the wolf pack , select the best one as the wolf king, if is a fierce wolf, otherwise it is a scout wolf;

[0086] In the formula: is the exploration position of the wolf king, is the position before exploration, is the wandering step size;

[0087] Step 3.3, Scout wolf wandering: The scout wolf performs the wandering behavior and calculates its objective function value , if the scout wolf is better than the objective function value of the wolf king, that is , eliminate the leading wolf and fierce wolves, generate new wolves at the new position, and go to step 3.2 until the maximum number of iterations is reached;

[0088] The described scout wolf wanders according to Execute the wandering,

[0089] In the formula: is the new position after exploration, is the position before exploration, is a random number uniformly distributed in [-1, 1];

[0090] The wandering behavior of the scout wolf is to find the local optimal solution other than the position of the leading wolf. This behavior is related to the ability of the algorithm to jump out of the local optimal solution. The number of its explorations and the number of scout wolves are parameters that can be set;

[0091] Step 3.4, Fierce wolf siege: The fierce wolf performs the siege behavior and calculates its objective function value , if the fierce wolf is better than the objective function value of the wolf king, that is , go to step 3.2 until the maximum number of iterations is reached,

[0092] The described fierce wolf performs the siege behavior according to Execute the siege behavior;

[0093] The raiding behavior is to change the positions of all individuals in the whole. On the one hand, it can jump out of the local optimal solution to a certain extent. On the other hand, it is to gather enough individuals for the siege behavior to improve the accuracy of the solution. In order to improve the algorithm speed and retain the ability of the algorithm to improve the solution accuracy, the improvement is as follows: Combine the raiding behavior and the siege behavior. Those within the radius near the leading wolf are fierce wolves, and the rest are scout wolves;

[0094] Step 3.5: Reach the maximum number of iterations and output the final coordinates of the alpha wolf;

[0095] Among them, in the said Step 3.2, the walking step size : ,

[0096] In the formula: and are the boundaries of the solution space, is the step size factor; among them, the conversion coefficient in the said Step 4 is the ratio of the time used when the reliability in the new and old environments is equal: ,

[0097] In the formula: R is the equal reliability in the two environments. Inside the parentheses, i represents the old environment when the environmental stress has 1 mutation or the i-th stage when the environmental stress has K mutations,

[0098] j represents the new environment when the environmental stress has 1 mutation or the j-th stage when the environmental stress has K mutations, k≥2; the conversion coefficient is calculated according to the shape parameter m and the scale parameter μ obtained in Step 3;

[0099] Among them, the specific method of the said Step 5 is: X （j） =Z* X （i） ,

[0100] In the formula: X （j） is the life of the wear-out product in the new environment, and X （i） is the life of the wear-out product in the old environment.

[0101] Example 2. To verify the effectiveness of the improved wolf pack algorithm in Example 1, a set of cumulative probability densities conforming to the Weibull distribution are simulated and generated: , in the formula and are the parameters to be obtained, and are respectively set as =1.9, =30. Take 100 data points with a uniform distribution to sample and record the cumulative failure probability function value to obtain the simulated measured value ,

[0102] Use the conventional wolf pack algorithm and the improved wolf pack algorithm to solve its and respectively. The parameter settings are as follows:

[0103] A. Set the objective function value , among which represents according to The cumulative probability density of the Weibull distribution of the parameter to be solved is shown as follows. is the parameter to be solved;

[0104] B. Set the parameter value range, that is, the solution space range is 0 ;

[0105] C. Set the maximum number of iterations T max = 50;

[0106] D. Set the step factor = 20;

[0107] E. Set the population size to 100.

[0108] The final output result of the conventional wolf pack algorithm = 1.89201, = 30.1373, and the final output result of the improved wolf pack algorithm = 1.90307, = 29.8877. Both are very close to the true value.

[0109] As Figure 3 shown, the speed comparison between the conventional wolf pack algorithm and the improved wolf pack algorithm: To compare the algorithm running time, for the conventional wolf pack algorithm, Figure 3 the horizontal axis in represents the number of iterations and the wandering times of the exploring wolves. It can be seen that compared with the conventional wolf pack algorithm, the improved wolf pack algorithm has a better convergence speed and higher accuracy.

[0110] Example 3. Conduct a simulation experiment verification when the environmental stress of the method in Example 1 undergoes 1 mutation.

[0111] Use a set of 59 simulation data points as Figure 4 shown for verification. At the moment, the product usage environment changes.

[0112] The specific steps are as follows:

[0113] Step 2. Set the objective function value , and the other parameters are the same as those in Example 3; Use the improved wolf pack calculation method to calculate and obtain = 128.9330.

[0114] Set the objective function value , and the other parameters are the same as those in Example 2 ( (1) = 1.9, (1) = 30); Use the improved wolf pack algorithm to calculate and obtain = 63.7480.

[0115] Step 4: Calculate the product conversion factor in the new environment .

[0116] Step 5: Since , the service life of the product under the new platform should be converted by half, and the subsequent usage of the product can be inferred based on the calculation results as Figure 5 shown.

[0117] Example 4. Conduct a simulation experiment verification when the environmental stress of the method in Example 1 undergoes k (k≥2) mutations. In this example, k is taken as 4.

[0118] I. List of Observation Data on Stage-by-Stage Failures

[0119] Implementation scenario: A batch of products (N = 220 is the total sample size) undergoes 4 stages of environmental stress changes (land-based storage → sea-based storage → sea-based cyclic load → sea-based storage), with a total observation period of 21 months, and the cumulative failure quantity is recorded once a month.

[0120] The observation data is shown in Table 1:

[0121] Table 1 Observation Data Table on Stage-by-Stage Failures

[0122] .

[0123] II. Setting of Parameters for Wolf Pack Algorithm

[0124] The core parameters of the algorithm are defined as follows:

[0125] The size of the wolf pack is 100, the maximum number of iterations is 100, the search range of the scale parameter μ is [20, 150], and the search range of the shape parameter m is [1, 3].[[]]

[0126] III. Setting of Fitness Function for Wolf Pack Algorithm

[0127] The single-point deviation is the absolute difference between the two: .

[0128] Global fitness: The sum of the deviations at all observation time points is used as the fitness value of the wolf pack individuals: .

[0129] IV. Calculation of Equivalent Time

[0130] For each wolf (candidate solution):

[0131] 1. Input: The parameters carried by the wolf (μ1, m1, μ2, m2,...);

[0132] 2. Calculate the equivalent time term:

[0133] Calculate ΔT̃ k(using the μ and m parameters of the current wolf); calculate for each stage ;

[0134] 3. Calculate the predicted failure probability ;

[0135] 4. Calculate the deviation: the absolute difference from the observed failure probability ;

[0136] 5. Fitness value: the sum of the deviations at all time points.

[0137] For example, assume the parameter estimation value of a certain wolf , and now calculate the equivalent time term for the 16th month (stage 4). The stage division is shown in Table 2:

[0138] Table 2 Stage Division Table

[0139] .

[0140] (1) Stage 1 (k = 1, months 1 - 6) ,

[0141] Set , then: , , ,

[0142] (2) Stage 2 (k = 2, months 7 - 12)

[0143] , , ,

[0144] (3) Stage 3 (k = 3, months 13 - 15)

[0145] , , ,

[0146] (4) Stage 4 (k = 4, months 16 - 21)

[0147] , , ,

[0148] , ,

[0149] Single-point deviation , this deviation will be included in the total fitness value. The wolf pack algorithm continuously adjusts the parameters to minimize the total deviation at all time points. Table 3 shows the list of deviations for each point of this wolf.

[0150] Table 3 Parameter Estimation Values are Wolf Deviation Table for Each Point

[0151] Month Stage Predicted failure number Actual observed number Absolute deviation 1 1 ≈0.03 0 ≈0.03 2 1 ≈0.14 0 ≈0.14 3 1 ≈0.31 0 ≈0.31 4 1 ≈0.55 0 ≈0.55 5 1 ≈0.86 1 ≈0.14 6 1 ≈1.23 1 ≈0.23 7 2 ≈1.27 1 ≈0.27 8 2 ≈1.37 1 ≈0.37 9 2 ≈1.54 1 ≈0.54 10 2 ≈1.78 2 ≈0.22 11 2 ≈2.09 2 ≈0.09 12 2 ≈2.46 2 ≈0.46 13 3 ≈2.57 2 ≈0.57 14 3 ≈3.26 3 ≈0.26 15 3 ≈3.73 3 ≈0.73 16 4 ≈3.82 4 ≈0.18 17 4 ≈4.28 5 ≈0.72 18 4 ≈5.35 7 ≈1.65 19 4 ≈7.24 7 ≈0.24 20 4 ≈9.20 7 ≈2.20 21 4 ≈10.93 7 ≈3.93 ;

[0152] The current wolf fitness value is the sum of total absolute deviations ;

[0153] V. Output the fitting parameters and life conversion for each stage, as shown in Table 4

[0154] Table 4 Fitting Parameters and Life Conversion Table for Each Stage

[0155] .

[0156] If the land-based storage life of this product is 60 months, then at this time, while ensuring the reliability of the product remains unchanged, the remaining life is .

[0157] The specific embodiments described in this article are merely illustrative of the spirit of the present invention. Those skilled in the art of the present invention can make various modifications or supplements to the described specific embodiments or use similar methods for substitution, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A life calculation method for consumable products based on an improved wolf pack algorithm, characterized by: The following steps are involved: Step 1: Determine that the life of consumable products follows the Weibull distribution; Step 2: Determine the objective function based on the Weibull distribution in step 1; The objective function includes an objective function under a single environment, i.e., an objective function when the environmental stress does not change, an objective function when the environmental stress changes once, and an objective function when the environmental stress changes in multiple stages; The objective function in the single environment is: , Right now , Where: X is the observed value of cumulative failure probability, t is the observation time point, F is the estimated value of cumulative failure probability, m, are the shape and scale parameters of the Weibull distribution to be estimated; The objective function when the environmental stress undergoes a single change is: ， Right now: , Where: , is the cumulative failure probability in the new environment, for The probability of failure in the distribution curve is At the same time, As the product usage environment changes, Any time point after the product usage environment changes, is the shape parameter in the new environment, is the scale parameter in the new environment, for - Each time point t within the time period i The cumulative failure probability of the products above; The objective function when the environmental stress changes in multiple stages is: ， Where: For time point The conversion time at stage k is: Calculated by the following formula: ， The initial conditions are , reflecting the time relationship before the first stress change; is the scale parameter of the kth stage, is the shape parameter of the kth stage; Step 3: Use the improved wolf pack algorithm to solve the objective function in step 2 to obtain the required parameters under the new and old environments. Specifically, the following steps are included: Step 3.1, initialize each parameter: set the parameter value range, that is, the solution space range: ， ， Is a vector, each number in the vector is the upper limit of the corresponding parameter to be estimated, Is a vector, each number in the vector is the lower limit of the corresponding parameter to be estimated; Set the maximum number of iterations Tmax; set the step size adjustment rate ; Set the population size, initialize the wolf pack coordinates, and generate the wolf pack in a uniform distribution in the solution space; Step 3.2, wolf pack reorganization: calculate the objective function value of each individual wolf pack , select the best one as the wolf king, if It is a fierce wolf, and vice versa it is a scouting wolf; Where: Scouting for the Wolf King's location, To explore the front position, is the walking stride length; Step 3.3: The wolf walks: The wolf performs the walk behavior and calculates its objective function value. , if the scout wolf is better than the wolf king objective function value, that is , eliminate the alpha wolf and the fierce wolf, generate a new wolf in the new position, and go to step 3.2 until the maximum number of iterations is reached; The wolf-hunting Execute walk, Where: To explore the new location, To explore the front position, is a uniformly distributed random number between [-1, 1]; Step 3.4, Wolf Siege: The wolf performs the siege behavior and calculates its objective function value , if the fierce wolf is better than the wolf king objective function value, that is , go to step 3.2 until the maximum number of iterations is reached, The fierce wolf press Performing acts of mobbing; Step 3.5: When the maximum number of iterations is reached, output the final coordinates of the wolf leader; Step 4: Define the conversion coefficient and calculate the conversion coefficient based on the required parameters obtained in step 3. The ratio of the time taken when the reliability in the new and old environments is equal: , Where: R is the reliability that is equal in the two environments, and the i in the brackets represents the old environment when the environmental stress undergoes one mutation or the i-th stage when the environmental stress undergoes K mutations. j represents the new environment when the environmental stress undergoes one mutation or the jth stage when the environmental stress undergoes K mutations, k ≥ 2; The conversion coefficient is calculated based on the shape parameter m and scale parameter μ obtained in step 3. ; Step 5: Use the conversion factor from step 4 to calculate the life of the consumable product under the new environment.

2. The life conversion method for consumable products based on the improved wolf pack algorithm according to claim 1, characterized in that: The cumulative failure probability density of the consumable products in step 1 conforms to the Weibull distribution: 。 3. The life conversion method for consumable products based on the improved wolf pack algorithm according to claim 1 or 2, characterized in that: In step 3.2, the walking step length : ， Where: Adjust the rate for the step size.

4. The life conversion method for consumable products based on the improved wolf pack algorithm according to claim 3 is characterized by: The specific method of step 5 is: （j） =Z*X （i） , Where: X （j） For the life of consumable products in new environments, X （i） It is the life span of consumable products in the old environment.

Images