Tire Vulcanization Quality Prediction Method Based on Whale Optimization Algorithm Optimizing BP Neural Network

Abstract

The present invention provides a method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm, which relates to the technical field of neural network prediction. The present invention uses a neural network to predict the performance quality of tires after vulcanization, which is faster and more efficient than the traditional analysis and prediction of vulcanization experimental data, and can not only provide a model basis for subsequent process optimization, but also realize the visualization of vulcanization results. At the same time, the initial weights and thresholds of the BP neural network of the whale algorithm are utilized, so that the improved and optimized neural network has a higher-precision prediction ability, and the prediction accuracy of the tire vulcanization quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of neural network prediction, and particularly relates to a method, a system, a storage medium and an electronic device for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm. Background Art

[0002] With the increasing demand for automobiles in the market, the tire industry, as one of the main supporting industries in the automotive industry, has developed rapidly. In order not to affect the handling stability and driving safety of vehicles, while ensuring high production efficiency, the performance and quality of tires should also be emphasized simultaneously. Vulcanization, as one of the important processes in tire production and also the last process, aims to cross-link and vulcanize the unvulcanized rubber so that the tire has the required physical properties and meets the use requirements. However, due to the coupling effect of various influencing factors, the quality of vulcanized tires decreases, resulting in a low yield rate. Therefore, a large amount of work has been done on predicting the quality of vulcanized tires through experiments, theories and machine learning methods.

[0003] Compared with the traditional tire vulcanization experimental method, the machine learning method has the advantages of high efficiency, economy and accuracy, and has thus been widely applied in the industrial field. Currently, it is common to use a BP neural network to establish a network model of vulcanization process parameters and performance indicators, use the performance output predicted by the neural network as the solution method of the objective function, and use the genetic algorithm as the optimization method of vulcanization process parameters to predict and optimize vulcanization process parameters.

[0004] However, when the genetic algorithm optimizes the BP neural network, it has a certain dependence on the selection of the initial population and is prone to falling into the local optimum, resulting in unstable accuracy of its prediction results and unstable quality of tire vulcanization. Summary of the Invention

[0005] (1) Technical Problems to be Solved

[0006] Aiming at the deficiencies of the prior art, the present invention provides a method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm, and solves the technical problem that the method for predicting the vulcanization quality of tires based on a BP neural network is prone to falling into the local optimum.

[0007] (2) Technical Solutions

[0008] To achieve the above object, the present invention is realized through the following technical solutions:

[0009] In the first aspect, the present invention provides a method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm, including the following steps:

[0010] S1. Obtain historical data, where the historical data includes tire vulcanization process parameters and quality index data measured by a testing machine;

[0011] S2. Perform normalization processing on the historical data; combine with an empirical formula, train a neural network with the processed historical data to obtain the minimum neural network error value, and determine the topological structure of the BP neural network model based on the minimum neural network error value;

[0012] S3. Initialize the parameters of the BP neural network model for the tire vulcanization quality prediction scenario based on the whale algorithm;

[0013] S4. Use the whale individual position as the initial weights and thresholds of the BP neural network model, and traverse with the training error value of the BP neural network model as the fitness function;

[0014] S5. Let the whale search for prey, surround prey or drive prey and update its position. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whale. Otherwise, return to step S4 to execute again;

[0015] S6. Assign the best position to the best weights and thresholds of the BP neural network model, perform network training to obtain an optimized BP neural network model, and obtain the quality index prediction result through the optimized BP neural network model.

[0016] Preferably, initializing the parameters of the BP neural network model for the tire vulcanization quality prediction scenario based on the whale algorithm includes:

[0017] Execute parameter initialization for the model in the tire vulcanization quality prediction scenario, set the number of whales N, and the number of iterations as T max , initialize the optimization dimension as D in combination with the network topological structure, and initialize the parameters A, a, C, and improve the definition formula of the convergence factor parameter a. The calculation formula is as follows:

[0018] a = 2 - 2*(t 2 / T max 2 )

[0019] A = 2ar1 - a

[0020] C = 2r2

[0021] where a is the convergence factor, which decreases from 2 to 0 with the number of iterations, t is the current number of iterations, and T max is the maximum number of iterations; A and C are collaborative coefficient vectors, and r1 and r2 are both random numbers in (0, 1).

[0022] Preferably, using the positions of whale individuals as the initial weights and thresholds of the BP neural network model, and traversing with the training error value of the BP neural network model as the fitness function, includes:

[0023] Taking the training error value error as the fitness value of the whale population, calculating the minimum fitness value of the whale population and the position of the best whale individual, using the position of the best whale individual as the initial weights and thresholds of the BP neural network model, and updating the weights and thresholds according to the gradient descent method. The update formulas are as follows:

[0024]

[0025]

[0026]

[0027]

[0028] Where: μ is the learning rate of the BP neural network, ω1 and b1 are the weights and thresholds between the input layer and the hidden layer in the BP neural network respectively, ω2 and b2 are the weights and thresholds between the hidden layer and the output layer in the BP neural network respectively, ω’1 and b’1 are the updated weights and thresholds of ω1 and b1 respectively, and ω’2 and b’2 are the updated weights and thresholds of ω2 and b2 respectively.

[0029] Preferably, the whales search for prey, surround the prey or drive the prey away and update their positions, including:

[0030] S501: Generate a random number P between 0 and 1 to determine whether the whale chooses to search for and surround the prey or use the bubble net to chase. If P < 0.5, execute step S502; otherwise, execute step S503.

[0031] S502: The whale group drives the prey away by creating a bubble net, swims around the prey in a shrinking circle, and at the same time swims along a spiral path and updates its position;

[0032] S503: When |A| > 1, the whale conducts a global search for prey and updates its position; when |A| < 1, the whale conducts a local search for prey and updates its position.

[0033] Preferably, in S502, the update method of swimming along the spiral path and updating the position includes:

[0034]

[0035]

[0036] Where, Indicates the distance between the whale individual and the prey, is the position of the prey at time t, is the position of the whale individual at time t, is the updated position, and l is a random number between [-1, 1].

[0037] Preferably, in S503, when |A| > 1, the whale conducts a global search for prey and updates its position, including:

[0038] When |A| > 1, the whale group will enter the stage of randomly searching for prey, that is, the whale group will randomly select a whale individual and update the position of the whale group towards the current random whale individual;

[0039]

[0040]

[0041] Among them, represents the distance between the currently selected random whale individual and other whale individuals at time t, is the position of a random whale individual at time t, is the position of other whale individuals at time t, is the updated position, is a random vector.

[0042] Preferably, in S503, when |A| < 1, the whale conducts a local search for prey and updates its position, including:

[0043] When |A| < 1, the whale group will enter the stage of surrounding the prey, that is, the whale group will conduct a local search for prey towards the current optimal whale individual and update its position;

[0044]

[0045]

[0046] Among them, represents the distance between the current optimal whale individual and other whale individuals at time t, is the position of the optimal whale individual at time t, is the position of other whale individuals at time t, is the updated position, is a random vector.

[0047] In a second aspect, a tire vulcanization quality prediction system based on a whale algorithm optimized BP neural network, the system includes:

[0048] A data acquisition module for acquiring historical data, where the historical data includes tire vulcanization process parameters and quality index data measured by a testing machine;

[0049] A model structure determination module for normalizing the historical data; combining empirical formulas, training a neural network with the processed historical data to obtain the minimum value of the neural network error, and determining the topological structure of the BP neural network model based on the minimum value of the neural network error;

[0050] An initialization parameter module for initializing the parameters of the BP neural network model for the tire vulcanization quality prediction scenario based on the whale algorithm;

[0051] A traversal module for using the whale individual position as the initial weights and thresholds of the BP neural network model and the training error value of the BP neural network model as the fitness function for traversal;

[0052] An optimization module for having the whale search for prey, surround prey, or drive prey and update its position. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whale. Otherwise, return to step S4 and execute again;

[0053] A prediction module for assigning the best position to the best weights and thresholds of the BP neural network model, performing network training to obtain an optimized BP neural network model, and obtaining the quality index prediction result through the optimized BP neural network model.

[0054] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program for tire vulcanization quality prediction based on optimizing a BP neural network using the whale algorithm, wherein the computer program causes a computer to execute the method for tire vulcanization quality prediction based on optimizing a BP neural network using the whale algorithm as described above.

[0055] In a fourth aspect, the present invention provides an electronic device, including:

[0056] One or more processors;

[0057] A memory; and

[0058] One or more programs, where the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include those for executing the method for tire vulcanization quality prediction based on optimizing a BP neural network using the whale algorithm as described above.

[0059] (III) Beneficial effects

[0060] The present invention provides a method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm. Compared with the prior art, it has the following beneficial effects:

[0061] The present invention uses a neural network to predict the performance and quality of tires after vulcanization, which is faster and more efficient than the traditional analysis and prediction of vulcanization experimental data. Moreover, it can provide a model basis for subsequent process optimization and realize the visualization of vulcanization results. At the same time, the initial weights and thresholds of the BP neural network of the whale algorithm are utilized, so that the improved and optimized neural network has a higher-precision prediction ability, improving the prediction accuracy of tire vulcanization quality. Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Figure 1 It is a block diagram of the method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm in an embodiment of the present invention;

[0064] Figure 2 It is a schematic flow chart of the improved whale algorithm in an embodiment of the present invention;

[0065] Figure 3 It is a comparison chart of the fitness of the method in an embodiment of the present invention and the optimization experiments of different algorithms. Detailed Embodiments

[0066] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0067] By providing a method for predicting the vulcanization quality of tires by optimizing a BP neural network based on a whale algorithm in an embodiment of the present application, the technical problem that the method for predicting the vulcanization quality of tires based on a BP neural network is prone to falling into a local optimum is solved. The initial weights and thresholds of the BP neural network are optimized by the improved whale algorithm, improving the disadvantage that the whale algorithm falls into a local optimum during the optimization process, achieving an improvement in the accuracy of the prediction results, and thus improving the yield rate of tires.

[0068] The technical solutions in the embodiments of the present application for solving the above technical problems are generally as follows:

[0069] Compared with the traditional tire vulcanization experimental method, the machine learning method has the advantages of high efficiency, economy and accuracy, and is thus widely used in the industrial field. Due to the complex current tire vulcanization process, the existing methods have a long analysis time and cannot reflect the high efficiency. For the problem of tire vulcanization quality prediction, it is currently common to use a BP neural network to establish a network model of vulcanization process parameters and performance indicators. However, when the genetic algorithm optimizes the BP neural network, it has a certain dependence on the selection of the initial population and is prone to falling into a local optimum, resulting in unstable accuracy of its prediction results and unstable tire vulcanization quality. To solve the above problems, the embodiments of the present invention use the improved whale algorithm to optimize the initial weights and thresholds of the BP neural network, improve the shortcoming that the whale algorithm falls into a local optimum during the optimization process, realize coordinated local search during the optimization process, improve the local development ability, and make the improved and optimized neural network have a higher-precision prediction ability.

[0070] To better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments.

[0071] The embodiments of the present invention provide a tire vulcanization quality prediction method based on the whale algorithm to optimize the BP neural network, as Figure 1 shown, the method includes:

[0072] S1. Obtain historical data of tire vulcanization quality;

[0073] S2. Perform normalization processing on the historical data; combine the empirical formula, train the neural network with the processed historical data to obtain the minimum value of the neural network error, and determine the topological structure of the BP neural network model based on the minimum value of the neural network error;

[0074] S3. Based on the whale algorithm, initialize the parameters of the BP neural network model for the tire vulcanization quality prediction scenario;

[0075] S4. Use the position of the whale individual as the initial weights and thresholds of the BP neural network model, and traverse with the training error value of the BP neural network model as the fitness function;

[0076] S5. Have the whale search for prey, surround the prey or drive the prey and update the position. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whale. Otherwise, return to step S4 to execute again;

[0077] S6. Assign the optimal position to the optimal weights and thresholds of the BP neural network model, conduct network training to obtain an optimized BP neural network model, and obtain the prediction results of quality indicators through the optimized BP neural network model.

[0078] In the embodiment of the present invention, a neural network is used to predict the performance quality after tire vulcanization, which is faster and more efficient than the traditional vulcanization experimental data analysis and prediction. Moreover, it can not only provide a model basis for subsequent process optimization but also realize the visualization of vulcanization results. At the same time, the initial weights and thresholds of the whale algorithm BP neural network are utilized, enabling the improved and optimized neural network to have a higher-precision prediction ability and enhancing the prediction accuracy of tire vulcanization quality.

[0079] The following is a detailed description of each step:

[0080] In step S1, historical data of tire vulcanization quality is obtained. The specific implementation process is as follows:

[0081] Collect and obtain historical data of tire vulcanization quality to get tire vulcanization process parameters (such as vulcanization temperature t, vulcanization time T, vulcanization pressure F) and quality indicator data measured by a testing machine.

[0082] In step S2, the historical data is normalized; combined with an empirical formula, the neural network is trained with the processed historical data to obtain the minimum value of the neural network error. Based on the minimum value of the neural network error, the topological structure of the BP neural network model is determined. The specific implementation process is as follows:

[0083] S201. Normalize the historical data, and the normalization formula is as follows;

[0084]

[0085] Where: X is the variable before normalization in the historical data; max(X) and min(X) are the maximum and minimum values of the historical data respectively; Y is the variable after normalization.

[0086] It should be noted that various types of data are normalized separately.

[0087] S202. Calculate the number of hidden layer nodes of the BP neural network according to the minimum value of the neural network error, and determine the topological structure of the BP neural network model. The specific implementation process is as follows:

[0088] According to the node number interval calculated by the empirical formula and combining network training to calculate the minimum value of the neural network error, the number of hidden layer nodes of the BP neural network is determined. The empirical formula is as follows:

[0089]

[0090] Among them, n is the number of hidden nodes, n1 is the number of nodes in the input layer, n2 is the number of nodes in the output layer, and c is a constant in the interval [1, 10]. From this, it is obtained that the integer value of the hidden layer n is taken from [3, 12]. After calculation, when the number of hidden layers n = 6, the error value of the neural network is the smallest. Thus, a 3×6×4 BP neural network structure is established. In the specific implementation process, each number in [3, 12] is taken for n and substituted into the BP neural network for training respectively, and the corresponding error values are obtained. The n value corresponding to the minimum error value is taken as the number of hidden layers, which has been omitted and directly obtained. The parameter value is used later. It should be noted that during the training process, the vulcanization temperature t, vulcanization time T, and vulcanization pressure F are used as the inputs of the BP neural network model, and the quality index data is used as the output of the BP neural network model.

[0091] In step S3, based on the whale algorithm, the parameters of the BP neural network model are initialized for the tire vulcanization quality prediction scenario. The specific implementation process is as follows:

[0092] For the initialization of the model execution parameters in the tire vulcanization quality prediction scenario, set the number of whales N, the number of iterations as Tmax, initialize the optimization dimension as D in combination with the network topology structure, and initialize the parameters A, a, C, and improve the definition formula of the convergence factor parameter a. The calculation formula is as follows:

[0093] a = 2 - 2*(t 2 / Tmax 2 )

[0094] A = 2ar1 - a

[0095] C = 2r2

[0096] Among them, a is the convergence factor, which decreases from 2 to 0 with the number of iterations, t is the current number of iterations, Tmax is the maximum number of iterations; A and C are the cooperative coefficient vectors, and r1 and r2 are both random numbers in (0, 1).

[0097] In step S4, the whale individual positions are used as the initial weights and thresholds of the BP neural network model, and the training error value error of the BP neural network model is used as the fitness function value fitness for traversal. The specific implementation process is as follows:

[0098] Take the training error value error as the fitness value of the whale population, calculate the minimum fitness value of the whale population and the position of the best whale individual, use the position of the best whale individual as the initial weights and thresholds of the BP neural network model, and update the weights and thresholds according to the gradient descent method. The update formula is as follows:

[0099]

[0100]

[0101]

[0102]

[0103] Where: μ is the learning rate of the BP neural network, ω1 and b1 are the weights and thresholds between the input layer and the hidden layer in the BP neural network respectively, ω2 and b2 are the weights and thresholds between the hidden layer and the output layer in the BP neural network respectively, ω'1 and b'1 are the updated weights and thresholds of ω1 and b1 respectively, and ω'2 and b'2 are the updated weights and thresholds of ω2 and b2 respectively.

[0104] The process of the improved whale algorithm is as Figure 2 shown.

[0105] In step S5, the whales search for prey, surround the prey or drive the prey and update their positions. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whales. Otherwise, return to step S4 and execute again.

[0106] It should be noted that the whales in the embodiments of the present invention are whales with improved parameter definition formulas.

[0107] S501. Generate a random number P between 0 and 1 to determine whether the whale chooses to search for and surround the prey or use the bubble net to chase. If P < 0.5, execute step S502; otherwise, execute step S503.

[0108] S502. The whale group drives the prey by making a bubble net, swims around the prey in a continuously shrinking circle, and at the same time swims along a spiral path and updates its position.

[0109]

[0110]

[0111] Where, [[ID=ID=38]]represents the distance between the whale individual and the prey (the current optimal solution), is the position of the prey at time t, is the position of the whale individual at time t, and l is a random number between [-1, 1].

[0112] S503. When |A| > 1, the whale conducts a global search for prey and updates its position; when |A| < 1, the whale conducts a local search for prey and updates its position. Specifically:

[0113] When |A| > 1, the whale group will enter the stage of randomly searching for prey, that is, the whale group will randomly select a whale individual and update the position of the whale group towards the current randomly selected whale individual;

[0114]

[0115]

[0116] Among them, represents the distance between the currently selected random whale individual and other whale individuals at time t, is the position of a randomly selected whale individual at time t, is the position of other whale individuals at time t, is a random vector.

[0117] When |A| < 1, the whale group will enter the stage of surrounding the prey, that is, the whale group will conduct local search for prey towards the current optimal whale individual and update the position;

[0118]

[0119]

[0120] Among them, represents the distance between the current optimal whale individual and other whale individuals at time t, is the position of the optimal whale individual at time t, is the position of other whale individuals at time t, is a random vector.

[0121] In step S6, the best position is assigned to the best weights and thresholds of the BP neural network model, and network training is carried out to obtain an optimized BP neural network model. The quality index prediction result is obtained through the optimized BP neural network model. The specific implementation process is as follows:

[0122] The best position obtained in S5 is assigned to the best weights and thresholds of the BP neural network, and network training is carried out to obtain an optimized BP neural network model. The vulcanization temperature t, vulcanization time T, and vulcanization pressure F within a reasonable range are input into the optimized BP neural network model, and the quality index prediction result is output.

[0123] Figure 3 Show the fitness comparison graph of the improved whale algorithm and different algorithm optimization experiments in the embodiments of the present invention. Among them, WOA - BP refers to the existing whale optimization algorithm for optimizing the BP neural network, IWOA - BP refers to the method of the embodiments of the present invention, and GWO - BP refers to the gray wolf optimization algorithm for optimizing the BP neural network. It can be seen from Figure 3 that the optimization effect of the method of the embodiments of the present invention is the best.

[0124] An embodiment of the present invention further provides a tire vulcanization quality prediction system based on optimizing a BP neural network by a whale algorithm. The system includes:

[0125] A data acquisition module, configured to acquire historical data, where the historical data includes tire vulcanization process parameters and quality index data measured by a testing machine;

[0126] A model structure determination module, configured to perform normalization processing on the historical data; combine empirical formulas, train a neural network through the processed historical data to obtain the minimum value of the neural network error, and determine the topological structure of the BP neural network model based on the minimum value of the neural network error;

[0127] An initialization parameter module, configured to initialize the parameters of the BP neural network model for the tire vulcanization quality prediction scenario based on the whale algorithm;

[0128] A traversal module, configured to use the whale individual position as the initial weights and thresholds of the BP neural network model, and use the training error value of the BP neural network model as a fitness function to perform traversal;

[0129] An optimization module, configured to search for prey, surround prey, or drive prey by the whale and update the position. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whale, otherwise return to step S4 to execute again;

[0130] A prediction module, configured to assign the best position to the best weights and thresholds of the BP neural network model, perform network training to obtain an optimized BP neural network model, and obtain a quality index prediction result through the optimized BP neural network model.

[0131] It can be understood that the tire vulcanization quality prediction system based on optimizing a BP neural network by a whale algorithm provided by the embodiment of the present invention corresponds to the above-mentioned tire vulcanization quality prediction method based on optimizing a BP neural network by a whale algorithm. For the explanations, examples, beneficial effects, etc. of the relevant content, reference can be made to the corresponding content in the tire vulcanization quality prediction method based on optimizing a BP neural network by a whale algorithm, which will not be elaborated here.

[0132] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program for predicting tire vulcanization quality based on optimizing a BP neural network by a whale algorithm. Wherein, the computer program enables a computer to execute the tire vulcanization quality prediction method based on optimizing a BP neural network by a whale algorithm as described above.

[0133] An embodiment of the present invention further provides an electronic device, including:

[0134] One or more processors;

[0135] a memory; and

[0136] one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs include a method for predicting the quality of tire vulcanization by optimizing a BP neural network based on the whale algorithm as described above.

[0137] In summary, compared with the prior art, the following beneficial effects are achieved:

[0138] 1. In the embodiment of the present invention, a neural network is used to predict the performance quality after tire vulcanization, which is faster and more efficient than the traditional analysis and prediction of vulcanization experimental data. Moreover, it can provide a model basis for subsequent process optimization and realize the visualization of vulcanization results. At the same time, the initial weights and thresholds of the whale algorithm BP neural network are utilized, enabling the improved and optimized neural network to have a higher-precision prediction ability and enhancing the prediction accuracy of tire vulcanization quality.

[0139] 2. The calculation formula of the convergence factor a in the whale algorithm is improved, and the initial weights and thresholds of the BP neural network are optimized by using the improved whale algorithm, which improves the drawback of the whale algorithm being trapped in local optimum during the optimization process, realizes coordinated local search during the optimization process, improves the local development ability, and further enhances the prediction accuracy of tire vulcanization quality.

[0140] 1. In the embodiment of the present invention, the initial weights and thresholds of the BP neural network are optimized by using the improved whale algorithm, which improves the drawback of the whale algorithm being trapped in local optimum during the optimization process, realizes coordinated local search during the optimization process, improves the local development ability, and enables the improved and optimized neural network to have a higher-precision prediction ability. At the same time, in the embodiment of the present invention, a neural network is used to predict the performance quality after tire vulcanization, which is faster and more efficient than the traditional analysis and prediction of vulcanization experimental data, and provides a model basis for subsequent optimization of process parameters and can realize the visualization of vulcanization results.

[0141] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the said element.

[0142] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the vulcanization quality of tires by optimizing a BP neural network based on the whale algorithm, characterized in that, It includes the following steps: S1. Obtain historical data, where the historical data includes tire vulcanization process parameters and quality index data measured by a testing machine; S2. Perform normalization processing on the historical data; Combine an empirical formula, train a neural network with the processed historical data to obtain the minimum neural network error value, and determine the topological structure of the BP neural network model based on the minimum neural network error value; S3. Based on the whale algorithm, initialize the parameters of the BP neural network model for the tire vulcanization quality prediction scenario, including: For the initialization model execution parameters in the scenario of tire vulcanization quality prediction, set the number of whales N and the number of iterations T max , initialize the optimization dimension as D in combination with the network topology structure, and initialize the parameters A, a, and C. And improve the definition formula of the convergence factor parameter a. The calculation formula is as follows: a = 2 - 2*(t 2 / T max 2 ) A = 2ar1 - a C＝2r2 where a is the convergence factor, which decreases from 2 to 0 with the number of iterations, t is the current iteration number, and T max is the maximum number of iterations; A and C are the cooperation coefficient vectors, and both r1 and r2 are random numbers in the range (0, 1); S4. Use the whale individual position as the initial weights and thresholds of the BP neural network model, and traverse with the training error value of the BP neural network model as the fitness function, including: Take the training error value error as the fitness value of the whale population, calculate the minimum fitness value of the whale population and the best whale individual position, use the best whale individual position as the initial weights and thresholds of the BP neural network model, and update the weights and thresholds according to the gradient descent method. The update formula is as follows: Where: μ is the learning rate of the BP neural network, ω1 and b1 are the weights and thresholds between the input layer and the hidden layer in the BP neural network respectively, ω2 and b2 are the weights and thresholds between the hidden layer and the output layer in the BP neural network respectively, ω'1 and b'1 are the updated weights and thresholds of ω1 and b1 respectively, and ω'2 and b'2 are the updated weights and thresholds of ω2 and b2 respectively; S5. Have the whales search for prey, surround prey, or drive away prey and update their positions. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whales. Otherwise, return to step S4 and execute again; S6. Assign the best position to the best weights and thresholds of the BP neural network model, perform network training to obtain an optimized BP neural network model, and obtain the quality index prediction result through the optimized BP neural network model; Among them, having the whales search for prey, surround prey, or drive away prey and update their positions includes: S501. Generate a random number P between 0 and 1 to determine whether the whales choose to search for and surround prey or use the bubble net to chase. If P < 0.5, execute step S502; otherwise, execute step S503; S502. The whale group drives away prey by creating a bubble net, swims around the prey in a continuously shrinking circle, and at the same time swims along a spiral path and updates its position; S503. When |A| > 1, the whales perform global search for prey and update their positions; when |A| < 1, the whales perform local search for prey and update their positions.

2. The tire vulcanization quality prediction method based on the whale algorithm optimized BP neural network according to claim 1, characterized in that, In S502, the update method of swimming along the spiral path and updating the position includes: Among them, represents the distance between the whale individual and the prey, is the position of the prey at time t, is the position of the whale individual at time t, is the updated position, and l is a random number between [-1, 1].

3. The method for predicting the vulcanization quality of tires based on optimizing the BP neural network by the whale algorithm according to claim 1, wherein In S503, when |A| > 1, the whales perform global search for prey and update their positions, including: When |A| > 1, the whale group will enter a random prey search stage, that is, the whale group will randomly select a whale individual and update the position of the whale group towards the current random whale individual; Among them, represents the distance between the currently selected random whale individual at time t and other whale individuals, is the position of a random whale individual at time t, is the position of other whale individuals at time t, is the updated position, is a random vector.

4. The tire vulcanization quality prediction method based on optimizing the BP neural network by the whale algorithm according to claim 1, characterized in that, In S503, when |A| < 1, the whales perform local search for prey and update their positions, including: When |A| < 1, the whale group will enter the stage of surrounding the prey, that is, the whale group will conduct local search for the prey towards the current optimal whale individual and update the position. Among them, represents the distance between the current optimal whale individual and other whale individuals at time t, is the position of the optimal whale individual at time t, is the position of other whale individuals at time t, is the updated position, is a random vector.

5. A tire vulcanization quality prediction system based on optimizing the BP neural network by the whale algorithm, characterized in that, The system includes: A data acquisition module, used to execute S1, acquire historical data, where the historical data includes tire vulcanization process parameters and quality index data measured by a testing machine. A model structure determination module, used to execute S2, perform normalization processing on the historical data; combine empirical formulas, train a neural network with the processed historical data to obtain the minimum neural network error value, and determine the topological structure of the BP neural network model based on the minimum neural network error value. An initialization parameter module, used to execute S3, initialize the parameters of the BP neural network model for the tire vulcanization quality prediction scenario based on the whale algorithm, including: For the execution parameters of the initialization model in the tire vulcanization quality prediction scenario, set the number of whales N and the number of iterations as T max , initialize the optimization dimension as D in combination with the network topology structure, and initialize the parameters A, a, and C. And improve the definition formula of the convergence factor parameter a. The calculation formula is as follows: a = 2 - 2*(t 2 / T max 2 ) A = 2ar1 - a C＝2r2 where a is the convergence factor, which decreases from 2 to 0 with the number of iterations, t is the current number of iterations, and T max is the maximum number of iterations; A and C are the collaborative coefficient vectors, and both r1 and r2 are random numbers in the range (0, 1); A traversal module, used to execute S4, use the whale individual position as the initial weights and thresholds of the BP neural network model, and the training error value of the BP neural network model as the fitness function to perform traversal, including: Use the training error value error as the fitness value of the whale population, calculate the minimum fitness value of the whale population and the position of the best whale individual, use the position of the best whale individual as the initial weights and thresholds of the BP neural network model, and update the weights and thresholds according to the gradient descent method. The update formula is as follows: Where: μ is the learning rate of the BP neural network, ω1 and b1 are the weights and thresholds between the input layer and the hidden layer in the BP neural network respectively, ω2 and b2 are the weights and thresholds between the hidden layer and the output layer in the BP neural network respectively, ω'1 and b'1 are the updated weights and thresholds of ω1 and b1 respectively, and ω'2 and b'2 are the updated weights and thresholds of ω2 and b2 respectively. An optimization module, used to execute S5, make the whales search for prey, surround the prey or drive the prey and update the position. When the given accuracy requirement is met or the maximum number of iterations is reached, stop the iterative optimization and output the best position of the whales. Otherwise, return to step S4 to execute again. A prediction module, used to execute S6, assign the best position to the best weights and thresholds of the BP neural network model, conduct network training to obtain an optimized BP neural network model, and obtain the quality index prediction result through the optimized BP neural network model. Among them, making the whales search for prey, surround the prey or drive the prey and update the position includes: S501: Generate a random number P between 0 and 1 to determine whether the whale chooses to search for and surround the prey or use the bubble net to chase. If P < 0.5, execute step S502; otherwise, execute step S503. S502: The whale group drives the prey by creating a bubble net, swims around the prey in a continuously shrinking circle, and at the same time swims along a spiral path and updates the position. S503: When |A| > 1, the whales conduct global search for prey and update the position; when |A| < 1, the whales conduct local search for prey and update the position.

6. A computer-readable storage medium, characterized in that, It stores a computer program for predicting the vulcanization quality of tires based on optimizing a BP neural network by a whale algorithm, wherein the computer program causes a computer to execute the method for predicting the vulcanization quality of tires based on optimizing a BP neural network by a whale algorithm as described in any one of claims 1 to 4.

7. An electronic device, characterized in that, It includes: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include a method for executing the prediction of the vulcanization quality of tires based on optimizing a BP neural network by a whale algorithm as described in any one of claims 1 to 4.

Images