Cold continuous rolling force prediction method and device based on IDBO-SA-LSTM

Abstract

The invention relates to the technical field of rolling force prediction, in particular to a cold continuous rolling force prediction method and device based on IDBO-SA-LSTM, and the method comprises the steps: optimizing a dung beetle optimization algorithm based on a preset rolling force calculation model, a Circle chaotic mapping strategy, a golden sine strategy and a dynamic weight coefficient, so as to construct an improved dung beetle optimization algorithm; performing performance test on the improved dung beetle optimization algorithm through multiple benchmark test functions to obtain the improved dung beetle optimization algorithm meeting a preset performance test requirement; and on the basis of an improved dung beetle optimization algorithm, automatic optimization is carried out on a preset adaptive long-short-term memory network, and a target SA-LSTM rolling force prediction model is constructed through the adaptive long-short-term memory network after automatic optimization, so that the rolling force is predicted in real time by using the target SA-LSTM rolling force prediction model in an online prediction stage. Therefore, the problems that an existing mechanism model contains many empirical parameters which are difficult to determine, and the generalization performance and accuracy of cold continuous rolling force prediction cannot be guaranteed are solved.

Description

Technical Field

[0001] The present application relates to the technical field of cold tandem rolling rolling force prediction, and particularly relates to a cold tandem rolling rolling force prediction method and device based on IDBO-SA-LSTM. Background Technique

[0002] In recent years, the new energy industry has developed rapidly. Aluminum foil, which is widely used as raw material for lithium battery production and manufacturing, has an increasing demand and output. Cold tandem rolling is one of the important processes for producing aluminum foil. During cold tandem rolling, rolling force is a key means to control the thickness and shape of aluminum foil products. Therefore, the rolling force calculation model has an important impact on the quality of aluminum foil products.

[0003] However, during the actual production process, the rolling force is affected by various on-site factors. Traditional mechanism models often contain many empirical parameters that are difficult to determine, making it difficult to ensure the generalization performance and accuracy of the model, which urgently needs to be solved. Summary of the Invention

[0004] The present application provides a cold tandem rolling rolling force prediction method and device based on IDBO-SA-LSTM to solve problems such as the existing mechanism model containing many empirical parameters that are difficult to determine and being unable to ensure the generalization performance and accuracy of cold tandem rolling rolling force prediction.

[0005] The first aspect embodiment of the present application provides a cold tandem rolling rolling force prediction method based on IDBO-SA-LSTM, which is applied to the offline training stage and includes the following steps: optimizing a preset dung beetle optimization algorithm based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient to construct an improved dung beetle optimization algorithm; performing performance testing on the improved dung beetle optimization algorithm through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; automatically optimizing a preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and constructing a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network to perform real-time prediction of the rolling force using the target SA-LSTM rolling force prediction model in the online prediction stage.

[0006] Optionally, in an embodiment of the present application, the performance testing of the improved dung beetle optimization algorithm by using a variety of preset benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements includes: determining the variety of benchmark test functions, and setting the population size and the maximum number of iterations corresponding to each test function in the variety of benchmark test functions, wherein the variety of benchmark test functions includes a variety of unimodal test functions and a variety of multimodal test functions; based on the population size at each iteration and the maximum number of iterations, using each test function to perform performance testing on the improved dung beetle optimization algorithm.

[0007] Optionally, in an embodiment of the present application, the automatic optimization of a preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and the construction of a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network includes: constructing the adaptive long short-term memory network based on a preset self-attention mechanism, LSTM layer, activation layer, fully connected layer, and regression layer; obtaining the original cold tandem rolling force data, and performing random shuffling and normalization operations on the original cold tandem rolling force data to construct a target training format data set, and dividing the target training format data set to generate a training data set and a test data set; initializing multiple parameters in the improved dung beetle optimization algorithm, and initializing the dung beetle population corresponding to the improved dung beetle optimization algorithm based on the Circle chaos mapping strategy to obtain multiple individual positions, and using the multiple individual positions to train the adaptive long short-term memory network to generate individual positions with a target fitness, wherein the multiple parameters include the population size, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the behavior ratios of rolling balls, reproduction, foraging, and stealing in the population; using the improved dung beetle optimization algorithm to update the individual positions with the target fitness, and iterating cyclically until the maximum number of iterations to obtain the optimal solution parameters, and training the adaptive long short-term memory network with the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model; fine-tuning the initial SA-LSTM rolling force prediction model through the test data set to obtain the target SA-LSTM rolling force prediction model.

[0008] In the second aspect of the present application, an embodiment provides a cold tandem rolling force prediction method based on IDBO-SA-LSTM, which is applied to the online prediction stage and includes: obtaining actual data during the stable rolling process of a target cold tandem rolling mill; performing normalization and fitness calculation on the actual data to obtain corresponding fitness calculation indicators; inputting the actual data into a pre-constructed target SA-LSTM rolling force prediction model, and combining the fitness calculation indicators to output a rolling force prediction result during the stable rolling process of the target cold tandem rolling mill, where the target SA-LSTM rolling force prediction model is obtained by training a preset adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm using a training data set and a test data set of preset cold tandem rolling forces.

[0009] In the third aspect of the present application, an embodiment provides a cold tandem rolling force prediction device based on IDBO-SA-LSTM, which is applied to the offline training stage and includes the following steps: an optimization module for optimizing a preset dung beetle optimization algorithm based on a preset rolling force calculation model, a Circle chaos mapping strategy, a golden sine strategy, and a dynamic weight coefficient to construct an improved dung beetle optimization algorithm; a test module for performing performance testing on the improved dung beetle optimization algorithm through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; a modeling module for automatically optimizing a preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and constructing a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network to perform real-time prediction of the rolling force using the target SA-LSTM rolling force prediction model in the online prediction stage.

[0010] Optionally, in an embodiment of the present application, the test module includes: a setting unit for determining the variety of benchmark test functions and setting the population size and maximum number of iterations for each test function in the variety of benchmark test functions, where the variety of benchmark test functions includes a variety of unimodal test functions and a variety of multimodal test functions; an analysis unit for performing performance testing on the improved dung beetle optimization algorithm using each test function based on the population size at each iteration and the maximum number of iterations.

[0011] Optionally, in an embodiment of the present application, the modeling module includes: a construction unit, configured to construct the adaptive long short-term memory network based on a preset self-attention mechanism, LSTM layer, activation layer, fully connected layer, and regression layer; a division unit, configured to obtain the original cold tandem rolling force data, perform random shuffling and normalization operations on the original cold tandem rolling force data to construct a target training format data set, and divide the target training format data set to generate a training data set and a test data set; an initialization unit, configured to initialize multiple parameters in the improved dung beetle optimization algorithm, and initialize the dung beetle population corresponding to the improved dung beetle optimization algorithm based on the Circle chaotic mapping strategy to obtain multiple individual positions, and use the multiple individual positions to train the adaptive long short-term memory network to generate individual positions with target fitness, where the multiple parameters include the population size, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the behavior ratios of rolling the ball, breeding, foraging, and stealing in the population; an iteration unit, configured to update the individual positions with target fitness by using the improved dung beetle optimization algorithm, and iteratively loop until the maximum number of iterations to obtain the optimal solution parameters, and train the adaptive long short-term memory network by using the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model; a fine-tuning unit, configured to fine-tune the initial SA-LSTM rolling force prediction model by using the test data set to obtain the target SA-LSTM rolling force prediction model.

[0012] An embodiment of the fourth aspect of the present application provides a cold tandem rolling force prediction device based on IDBO-SA-LSTM, which is applied to the online prediction stage, and includes: an acquisition module, configured to acquire the actual data during the stable rolling process of a target cold tandem mill; a calculation module, configured to perform normalization and fitness calculation on the actual data to obtain corresponding fitness calculation metrics; a prediction module, configured to input the actual data into a pre-constructed target SA-LSTM rolling force prediction model, and combine the fitness calculation metrics to output the rolling force prediction result during the stable rolling process of the target cold tandem mill, where the target SA-LSTM rolling force prediction model is obtained by training an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm with a preset training data set and test data set of cold tandem rolling force.

[0013] An embodiment of the fifth aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the IDBO-SA-LSTM-based cold tandem rolling force prediction method as described in the above embodiments.

[0014] In the sixth aspect of the embodiments of the present application, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and when the program is executed by a processor, it implements the above-mentioned cold tandem rolling force prediction method based on IDBO-SA-LSTM.

[0015] In the seventh aspect of the embodiments of the present application, a computer program product is provided, including a computer program, and the computer program is executed to implement the above-mentioned cold tandem rolling force prediction method based on IDBO-SA-LSTM.

[0016] Therefore, the embodiments of the present application have the following beneficial effects:

[0017] The embodiments of the present application can optimize the preset dung beetle optimization algorithm based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient to construct an improved dung beetle optimization algorithm; perform performance tests on the improved dung beetle optimization algorithm through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; based on the improved dung beetle optimization algorithm, automatically optimize the preset adaptive long short-term memory network, and construct a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network to perform real-time prediction of the rolling force using the target SA-LSTM rolling force prediction model in the online prediction stage. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of cold tandem rolling force prediction, and realizing high-precision rolling force prediction in the aluminum foil cold tandem rolling process. Thus, it solves the problems that the existing mechanism model contains many empirical parameters that are difficult to determine and cannot guarantee the generalization performance and accuracy of cold tandem rolling force prediction.

[0018] The additional aspects and advantages of the present application will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present application. Description of the Drawings

[0019] The above-mentioned and / or additional aspects and advantages of the present application will become obvious and easy to understand from the following description of the embodiments in conjunction with the drawings, where:

[0020] Figure 1 It is a flowchart of a cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the offline training stage provided according to the embodiments of the present application;

[0021] Figure 2 It is a schematic diagram of an aluminum foil cold tandem rolling production equipment provided by an embodiment of the present application;

[0022] Figure 3Schematic diagram of the distribution of a Circle chaotic map iterated 500 times provided by an embodiment of the present application;

[0023] Figure 4 In (a) is a schematic diagram of the optimization process curve of an F 1 (x) test function provided by an embodiment of the present application;

[0024] Figure 4 In (b) is a schematic diagram of the optimization process curve of an F 2 (x) test function provided by an embodiment of the present application;

[0025] Figure 4 In (c) is a schematic diagram of the optimization process curve of an F 3 (x) test function provided by an embodiment of the present application;

[0026] Figure 4 In (d) is a schematic diagram of the optimization process curve of an F 4 (x) test function provided by an embodiment of the present application;

[0027] Figure 4 In (e) is a schematic diagram of the optimization process curve of an F 5 (x) test function provided by an embodiment of the present application;

[0028] Figure 4 In (f) is a schematic diagram of the optimization process curve of an F 6 (x) test function provided by an embodiment of the present application;

[0029] Figure 5 Schematic diagram of the execution logic of cold tandem rolling force prediction based on IDBO-SA-LSTM provided by an embodiment of the present application;

[0030] Figure 6 Schematic diagram of the deep learning network structure of a rolling force prediction model provided by an embodiment of the present application;

[0031] Figure 7 Flowchart of a cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the online prediction stage provided according to an embodiment of the present application;

[0032] Figure 8 Comparison result diagram of the predicted value and the actual value of the rolling force in one pass provided by an embodiment of the present application;

[0033] Figure 9 Comparison result schematic diagram of the predicted value and the actual value of the rolling force in two passes provided by an embodiment of the present application;

[0034] Figure 10 It is an example diagram of a cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the offline training stage according to an embodiment of the present application;

[0035] Figure 11 It is an example diagram of a cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the online prediction stage according to an embodiment of the present application;

[0036] Figure 12 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

[0037] Among them, 10 - a cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the offline training stage, 20 - a cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the online prediction stage; 101 - an optimization module, 102 - a test module, 103 - a modeling module; 201 - an acquisition module, 202 - a calculation module, 203 - a prediction module; 1201 - a memory, 1202 - a processor, 1203 - a communication interface. Detailed implementation manners

[0038] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present application and should not be construed as limiting the present application.

[0039] The cold tandem rolling force prediction method and device based on IDBO-SA-LSTM according to the embodiments of the present application will be described below with reference to the accompanying drawings. In view of the problems mentioned in the above background art, the present application provides a cold tandem rolling force prediction method based on IDBO-SA-LSTM. In this method, based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient, the preset dung beetle optimization algorithm is optimized to construct an improved dung beetle optimization algorithm; the performance of the improved dung beetle optimization algorithm is tested through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; based on the improved dung beetle optimization algorithm, the preset adaptive long short-term memory network is automatically optimized, and a target SA-LSTM rolling force prediction model is constructed through the automatically optimized adaptive long short-term memory network, so as to use the target SA-LSTM rolling force prediction model to perform real-time prediction on the rolling force in the online prediction stage. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of cold tandem rolling force prediction, and realizing high-precision rolling force prediction in the aluminum foil cold tandem rolling process. Thus, the problems that the existing mechanism model contains many empirical parameters that are difficult to determine and cannot guarantee the generalization performance and accuracy of cold tandem rolling force prediction are solved.

[0040] Specifically, Figure 1 The flowchart of a cold tandem rolling force prediction method based on IDBO-SA-LSTM for the offline training stage provided by the embodiments of the present application.

[0041] As Figure 1 shown, the cold tandem rolling force prediction method based on IDBO-SA-LSTM for the offline training stage includes the following steps:

[0042] In step S101, based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient, the preset dung beetle optimization algorithm is optimized to construct an improved dung beetle optimization algorithm.

[0043] In step S102, the performance of the improved dung beetle optimization algorithm is tested through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements.

[0044] In the embodiments of the present application, a rolling force calculation model can be constructed first. Those skilled in the art should understand that the commonly used cold rolling force calculation models include the Tselykov formula, the Stone formula, the Hill formula, etc., and the most commonly used is the Hill formula simplified by Bland and Ford, as shown in Equation (1):

[0045]

[0046] Among them, P is the average rolling force exerted on the rolled piece, B is the width of the rolled piece, and l c is the contact arc length between the rolled piece and the roll, Q P is the stress state coefficient related to the friction coefficient, and K T is the tension influence coefficient related to the tension of the rolled piece, is the average deformation resistance related to the chemical properties of the rolled piece.

[0047] In the embodiment of the present application, the layout of the cold tandem rolling production line for aluminum foil is as Figure 2 shown. In the actual production process, the rolling conditions will change with the change of the rolled piece process. The rolling force of each pass is not only related to process parameters such as the tension, rolling speed, and reduction of the rolled piece, but also related to boundary conditions such as the wear, cooling, and lubrication of the roll. In the theoretical calculation model of the rolling force, it is difficult to measure both the friction coefficient and the average deformation resistance, resulting in relatively low calculation accuracy and generalization performance of the theoretical model in actual production.

[0048] Furthermore, when using the traditional DBO to optimize the hyperparameters of the SA-LSTM model, it is found that the algorithm is prone to falling into local optimal solutions during the calculation process. To improve the calculation accuracy, in the embodiment of the present application, for the DBO, the distribution uniformity of the initial population is increased through the Circle chaotic mapping, the golden sine strategy is adopted to enhance the global search performance of the DBO, and a dynamic weight coefficient is introduced to balance between global search and local search, obtaining the Improved Dung Beetle Optimizer (IDBO) algorithm, which is specifically described as follows:

[0049] 1. Circle chaotic mapping:

[0050] The original Dung Beetle Optimizer (DBO) algorithm initializes the population in a random manner, which cannot ensure that the initial population is evenly distributed in the search space, resulting in a low diversity of the initial population; chaotic mapping has characteristics such as non-repetitiveness, randomness, and chaotic ergodicity, and can make the population distribution more uniform compared with the random generation relying on probability. Therefore, the embodiment of the present application can use chaotic mapping to generate the initial population to increase the diversity of potential solutions.

[0051] To enhance the stability of the DBO algorithm and the distribution uniformity of the initial population, this patent uses the Circle chaotic mapping to initialize the dung beetle population instead of random numbers. The specific calculation method of the Circle chaotic mapping is shown in Equation (2):

[0052]

[0053] Among them, x nIt represents the size of the result at the nth iteration, where both a and b are constants.

[0054] As a feasible implementation, the distribution of the Circle chaotic map iteration 500 times in the embodiments of this application is as Figure 3 shown. It can be seen from Figure 3 that the distribution of the Circle chaotic map in the space is relatively stable and uniform, with good ergodicity.

[0055] 2. Golden Sine Strategy

[0056] Golden-SA is a new type of meta-heuristic algorithm that uses the sine function for iterative optimization, which can accelerate the convergence speed and enhance the global search ability. Its update process is shown in Equation (3):

[0057]

[0058] where, and respectively represent the spatial position and the optimal position of the ith individual at the tth iteration, R 1 represents the random number of the individual's moving distance, R 2 represents the random number of the individual's moving direction, x 1 and x 2 are two constant coefficients calculated from the golden ratio.

[0059] In the dancing behavior of the rolling ball dung beetle, the embodiments of this application can determine the moving direction of the next iteration according to the environment around the current position. However, if the current position is already a local optimal point, it will lead to extremely small changes in the dancing actions of the dung beetle, making it easy for the population to fall into the local optimal solution. Therefore, in order to enhance the global search ability of the DBO algorithm, the embodiments of this application can adopt Golden-SA to expand the search space and replace the dancing update strategy of the dung beetle when there are no obstacles; at the same time, through R 1 and R 2 the search direction and search step size of the dung beetle population can be effectively controlled, further balancing the global search ability and local search ability of the algorithm.

[0060] 3. Dynamic Weight Coefficient

[0061] In the stealing behavior of the dung beetle population, the position update strategy is greatly affected by the global optimal position; since in the later stage of iteration, the population often approaches the optimal solution interval, focusing on local search can often make the search result more accurate. In order to further improve the solution accuracy and reduce unnecessary divergence behavior, the embodiments of this application can introduce a dynamic weight coefficient to optimize the stealing behavior of the dung beetle and strengthen the local search ability of the algorithm in the later stage of iteration. The optimized stealing dung beetle update strategy is shown in Equations (4)-(6):

[0062] x i t+1 = k 1 × X b + k 2 × S × g × (|x i t - X * | + |x i t - X b |) (4)

[0063]

[0064] Wherein, k 1 and k 2 are dynamic factors determined by the number of iterations, which can enhance the local development ability of the algorithm in the later stage of iteration.

[0065] Thus, the embodiments of the present application can optimize the dung beetle optimization algorithm through the Circle chaos mapping strategy, the golden sine strategy and the dynamic weight coefficient to construct an improved dung beetle optimization algorithm; thereafter, the embodiments of the present application can perform performance testing on the improved dung beetle optimization algorithm.

[0066] Optionally, in an embodiment of the present application, the performance of the improved dung beetle optimization algorithm is tested through a variety of preset benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements, including: determining a variety of benchmark test functions, and setting the population size and the maximum number of iterations corresponding to each test function in the variety of benchmark test functions, wherein the variety of benchmark test functions includes a variety of unimodal test functions and a variety of multimodal test functions; based on the population size and the maximum number of iterations at each single iteration, the performance of the improved dung beetle optimization algorithm is tested using each test function.

[0067] In order to test the optimization performance of the IDBO algorithm, the embodiments of the present application use benchmark test functions to test the performance of the IDBO algorithm, and conduct comparative experiments with the Sparrow Search Algorithm (SSA), the Grey Wolf Optimizer (GWO) and the DBO algorithm. The basic information of the benchmark test functions is shown in Table 1:

[0068] Table 1

[0069]

[0070] It should be noted that the test functions F 1 ~ F 4 in Table 1 are unimodal functions, F 5 and F 6It is a multimodal function. When running the test function, the population size is set to 30 for each single iteration, and the maximum number of iterations is set to 1000. The optimization processes of each search algorithm on the benchmark function are as shown in (a) in Figure 4 and (f) in Figure 4 . It can be seen from (a) in Figure 4 and (f) in Figure 4 that the IDBO algorithm has a faster convergence speed and optimization accuracy than the other three algorithms.

[0071] In step S103, based on the improved dung beetle optimization algorithm, the preset adaptive long short-term memory network is automatically optimized, and the target SA-LSTM rolling force prediction model is constructed through the automatically optimized adaptive long short-term memory network, so as to use the target SA-LSTM rolling force prediction model to predict the rolling force in real time during the online prediction stage.

[0072] Furthermore, the embodiment of the present application also needs to optimize the self-attention long short-term memory network (SA-LSTM) combined with the self-attention mechanism based on the improved dung beetle optimization algorithm, so as to construct the target SA-LSTM rolling force prediction model suitable for high-precision rolling force prediction in the cold tandem rolling process of aluminum foil.

[0073] Optionally, in an embodiment of the present application, based on an improved dung beetle optimization algorithm, an automatic optimization of a preset adaptive long short-term memory network is performed, and a target SA-LSTM rolling force prediction model is constructed through the automatically optimized adaptive long short-term memory network, including: constructing an adaptive long short-term memory network based on a preset self-attention mechanism, LSTM layer, activation layer, fully connected layer, and regression layer; obtaining the original cold tandem rolling force data, and performing random shuffling and normalization operations on the original cold tandem rolling force data to construct a target training format data set, and dividing the target training format data set to generate a training data set and a test data set; initializing multiple parameters in the improved dung beetle optimization algorithm, and initializing the dung beetle population corresponding to the improved dung beetle optimization algorithm based on the Circle chaotic mapping strategy to obtain multiple individual positions, and using the multiple individual positions to train the adaptive long short-term memory network to generate individual positions with a target fitness, where the multiple parameters include the population size, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the behavior ratios of rolling balls, reproduction, foraging, and stealing in the population; using the improved dung beetle optimization algorithm to update the individual positions with the target fitness, and iterating cyclically until the maximum number of iterations to obtain the optimal solution parameters, and training the adaptive long short-term memory network with the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model; fine-tuning the initial SA-LSTM rolling force prediction model through the test data set to obtain the target SA-LSTM rolling force prediction model.

[0074] In order to fully extract the deep coupling relationship between variables in the cold tandem rolling process, the embodiment of the present application can construct a deep learning model as shown in Figure 5 to predict the rolling force. In order to avoid model overfitting and improve the model operation speed, the embodiment of the present application can first use the self-attention mechanism to extract features from the original input, then use the LSTM layer to further fit the non-linear relationship between variables, and then map the output result of the LSTM layer to the dimension space where the rolling force is located through the activation layer, fully connected layer, and regression layer, so as to fully fit the coupling situation between the rolling force and the industrial field parameters.

[0075] In addition, various hyperparameters included in the LSTM layer will have a great impact on the model prediction results. The number of neurons N in the hidden layer, the initial learning rate a, and the regularization coefficient L are all important hyperparameters. Among them, the number of neurons N in the hidden layer determines the capacity of the LSTM model. If N is too small, it may cause the network to be unable to store enough state information, thus unable to effectively process the non-linear coupling relationship in the input sequence. If N is too large, it will lead to overfitting of the model, reducing the accuracy and generalization performance of the model. The initial learning rate a refers to the initial step size used by the optimization algorithm to update the model parameters at the beginning of training. If a is too small, the training speed of the model will be too slow, and it may cause the optimizer to stagnate near the local optimal solution. If a is too large, the model may not converge. The regularization coefficient L is the penalty term coefficient added to the loss function to prevent the model from being too complex or overfitting the training set data. If L is too small, the influence of the regularization term on the loss function will be weakened, easily leading to overfitting of the model. If L is too large, it will overly limit the parameter fitting ability of the model, resulting in a decline in prediction performance.

[0076] However, manually adjusting parameters requires a high labor cost. Therefore, in the embodiments of this application, the IDBO algorithm can be used to automatically optimize these hyperparameters of the LSTM model, and construct an IDBO-SA-LSTM model as shown in Figure 6 to predict the rolling force. The specific prediction process of the IDBO-SA-LSTM model is as follows:

[0077] (1) Read the original data of the cold tandem rolling force, randomly shuffle and normalize it to construct a trainable data set, and divide it into a training set and a test set according to a ratio of 8:2;

[0078] (2) Initialize the parameters of the IDBO algorithm, set the number of populations to 20, the maximum number of iterations to 50, the behavior ratios of the dung beetle populations of rolling, breeding, foraging, and stealing to 2:1:1:1, the optimization range of the number of neurons N in the hidden layer to [10, 100], the optimization range of the initial learning rate a to [1×10 -5 , 1×10 -1 , and the optimization range of the regularization coefficient L to [1×10 -8 , 1×10 -1 ;

[0079] (3) Use the Circle chaotic mapping to initialize the dung beetle population. Each individual represents a set of feasible solutions. Substitute the positions of all individuals into the SA-LSTM model for training to obtain the individual position with the optimal fitness;

[0080] (4) Optimize the DBO algorithm using the golden sine strategy and the dynamic weight coefficient to obtain the IDBO algorithm, and update the individual positions according to the IDBO algorithm;

[0081] (5) Iterate cyclically until the maximum number of iterations, substitute the obtained optimal solution parameters into the SA-LSTM model for training, and obtain the optimal SA-LSTM rolling force prediction model (i.e., the target SA-LSTM rolling force prediction model).

[0082] Furthermore, the embodiment of the present application can use the target IDBO-SA-LSTM model to predict the rolling force in the subsequent online prediction stage and perform data analysis on the prediction results, thereby improving the calculation accuracy and generalization performance of the model.

[0083] According to the cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the offline training stage proposed by the embodiment of the present application, by optimizing the preset dung beetle optimization algorithm based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient, an improved dung beetle optimization algorithm is constructed; the performance of the improved dung beetle optimization algorithm is tested through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; based on the improved dung beetle optimization algorithm, the preset adaptive long short-term memory network is automatically optimized, and a target SA-LSTM rolling force prediction model is constructed through the automatically optimized adaptive long short-term memory network to perform real-time prediction of the rolling force using the target SA-LSTM rolling force prediction model in the online prediction stage. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of cold tandem rolling force prediction and realizing high-precision rolling force prediction in the aluminum foil cold tandem rolling process.

[0084] Figure 7 It is a flowchart of a cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the online prediction stage provided by the embodiment of the present application.

[0085] As Figure 7 shown, the cold tandem rolling force prediction method based on IDBO-SA-LSTM includes the following steps:

[0086] In step S701, actual data during the stable rolling process of the target cold tandem rolling mill is obtained.

[0087] In step S702, the actual data is normalized and fitness calculation is performed to obtain corresponding fitness calculation indicators.

[0088] In step S703, the actual data is input into a pre-constructed target SA-LSTM rolling force prediction model, and combined with a fitness calculation index to output the rolling force prediction result during the stable rolling process of the target cold tandem mill. Among them, the target SA-LSTM rolling force prediction model is obtained by training a preset adaptive long short-term memory network with a training data set and a test data set of the cold tandem mill rolling force, and optimizing it through a preset improved dung beetle optimization algorithm.

[0089] The following details the process of cold tandem mill rolling force prediction based on IDBO-SA-LSTM through a specific embodiment.

[0090] 1. Hardware platform setting:

[0091] In terms of the operating system and computer hardware, the specific embodiment of this application can perform model training under the Windows 10 Professional Edition 64-bit operating system, with a processor of Intel(R) Xeon(R) Gold 5218 CPU@2.30GHz, a memory of 256GB, and a graphics card of NVIDIA TITAN RTX. In terms of software, the embodiment of this application uses MATLAB R2024b to build the network structure and conduct experiments.

[0092] 2. Data selection:

[0093] The specific embodiment of this application can monitor the 3000 series aluminum foil production line of a certain factory, and use the actual data during the stable rolling process of a 1950 mm two-stand six-high cold tandem mill as the data source. 109 stable rolling processes in different production states are selected for data collection, and 6554 groups of original data are obtained. Part of the original data is shown in Table 2.

[0094] Table 2

[0095]

[0096] As can be seen from Table 2, the embodiment of this application can use variables such as exit speed, air system pressure, work roll bending force, backup roll bending force, shifting position, entrance thickness difference, entrance tension, intermediate tension, exit tension, absolute position, and thickness error as input variables, and the rolling force as the output variable.

[0097] 3. Normalization method and fitness calculation method:

[0098] The specific embodiment of this application can use the maximum-minimum normalization method to map the original data to the [0,1] interval, and the specific calculation method is shown in Equation (7):

[0099]

[0100] Among them, x represents the original data, and x max is the maximum value in the original data, and x min is the minimum value in the original data, and x norm represents the value after the original data is normalized.

[0101] In the process of using IDBO for optimization, the embodiments of the present application can use the Root Mean Squared Error (RMSE) as the fitness calculation index, and its specific calculation method is shown in Equation (8):

[0102]

[0103] Among them, represents the predicted value of the rolling force, and x i represents the actual value of the rolling force.

[0104] After experiments, the embodiments of the present application can obtain the rolling force prediction of the IDBO-SA-LSTM model for two passes as Figure 8 and Figure 9 shown. From Figure 8 and Figure 9 it can be seen that the probability that the rolling force prediction error of the model within the 4% error band for two passes is greater than 99%.

[0105] According to the cold tandem rolling force prediction method based on IDBO-SA-LSTM proposed by the embodiments of the present application for the online prediction stage, by obtaining the actual data during the stable rolling process of the target cold tandem rolling mill; normalizing and calculating the fitness of the actual data to obtain the corresponding fitness calculation index; inputting the actual data into the pre-constructed target SA-LSTM rolling force prediction model, and combining the fitness calculation index to output the rolling force prediction result during the stable rolling process of the target cold tandem rolling mill, where the target SA-LSTM rolling force prediction model is obtained by training a preset training data set and test data set of cold tandem rolling force through an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of cold tandem rolling force prediction, and realizing high-precision rolling force prediction during the aluminum foil cold tandem rolling process.

[0106] Secondly, a cold tandem rolling force prediction device based on IDBO-SA-LSTM proposed by the embodiments of the present application is described with reference to the accompanying drawings.

[0107] Figure 10 is a block diagram of the cold tandem rolling force prediction device based on IDBO-SA-LSTM for the offline training stage of the embodiments of the present application.

[0108] As Figure 10 shown, the cold tandem rolling force prediction device 10 based on IDBO-SA-LSTM applied to the offline training stage includes: an optimization module 101, a test module 102, and a modeling module 103.

[0109] Among them, the optimization module 101 is used to optimize the preset dung beetle optimization algorithm based on a preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy, and dynamic weight coefficient, so as to construct an improved dung beetle optimization algorithm.

[0110] The test module 102 is used to perform performance tests on the improved dung beetle optimization algorithm through a preset variety of benchmark test functions, so as to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements.

[0111] The modeling module 103 is used to automatically optimize the preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and construct a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network, so as to use the target SA-LSTM rolling force prediction model to perform real-time prediction of the rolling force in the online prediction stage.

[0112] Optionally, in an embodiment of the present application, the test module 200 includes: a setting unit and an analysis unit.

[0113] Among them, the setting unit is used to determine a variety of benchmark test functions, and set the population size and maximum number of iterations corresponding to each test function in the variety of benchmark test functions, where the variety of benchmark test functions includes a variety of unimodal test functions and a variety of multimodal test functions.

[0114] The analysis unit is used to perform performance tests on the improved dung beetle optimization algorithm using each test function based on the population size and maximum number of iterations during a single iteration.

[0115] Optionally, in an embodiment of the present application, the modeling module 300 includes: a construction unit, a division unit, an initialization unit, an iteration unit, and a fine-tuning unit.

[0116] Among them, the construction unit is used to construct an adaptive long short-term memory network based on a preset self-attention mechanism, LSTM layer, activation layer, fully connected layer, and regression layer.

[0117] The division unit is used to obtain the original cold tandem rolling force data, and perform random shuffling and normalization operations on the original cold tandem rolling force data to construct a target training format dataset, and divide the target training format dataset to generate a training dataset and a test dataset.

[0118] An initialization unit is used to initialize multiple parameters in the improved dung beetle optimization algorithm, and based on the Circle chaotic mapping strategy, initialize the dung beetle population corresponding to the improved dung beetle optimization algorithm to obtain multiple individual positions, and use the multiple individual positions to train an adaptive long short-term memory network to generate individual positions with the target fitness. Among them, the multiple parameters include the population size, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the behavior ratios of rolling the ball, reproduction, foraging, and stealing in the population.

[0119] An iteration unit is used to update the individual positions with the target fitness using the improved dung beetle optimization algorithm, and iterate cyclically until the maximum number of iterations to obtain the optimal solution parameters, and train the adaptive long short-term memory network with the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model.

[0120] A fine-tuning unit is used to fine-tune the initial SA-LSTM rolling force prediction model through the test data set to obtain the target SA-LSTM rolling force prediction model.

[0121] It should be noted that the foregoing explanation of the embodiment of the cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the offline training stage also applies to the cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the offline training stage of this embodiment, and will not be elaborated here.

[0122] According to the cold tandem rolling force prediction device based on IDBO-SA-LSTM proposed in the embodiment of the present application, it includes an optimization module 101, which is used to optimize the preset dung beetle optimization algorithm based on a preset rolling force calculation model, Circle chaotic mapping strategy, golden sine strategy, and dynamic weight coefficient to construct an improved dung beetle optimization algorithm; a test module 102, which is used to perform performance tests on the improved dung beetle optimization algorithm through a preset variety of benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; a modeling module 103, which is used to automatically optimize the preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and construct a target SA-LSTM rolling force prediction model through the automatically optimized adaptive long short-term memory network to perform real-time prediction of the rolling force using the target SA-LSTM rolling force prediction model in the online prediction stage. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of the cold tandem rolling force prediction, and realizing high-precision rolling force prediction in the aluminum foil cold tandem rolling process.

[0123] Figure 11 It is a block diagram of the cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the online prediction stage of the embodiment of the present application.

[0124] As Figure 11 shown, the cold tandem rolling force prediction device 20 based on IDBO-SA-LSTM applied to the online prediction stage includes: an acquisition module 201, a calculation module 202, and a prediction module 203.

[0125] Among them, the acquisition module 201 is used to acquire the actual data during the stable rolling process of the target cold tandem rolling mill.

[0126] The calculation module 202 is used to perform normalization and fitness calculation on the actual data to obtain the corresponding fitness calculation index.

[0127] The prediction module 203 is used to input the actual data into the pre-constructed target SA-LSTM rolling force prediction model, and combine the fitness calculation index to output the rolling force prediction result during the stable rolling process of the target cold tandem rolling mill. Among them, the target SA-LSTM rolling force prediction model is obtained by training the preset training data set and test data set of the cold tandem rolling force through an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm.

[0128] It should be noted that the foregoing explanation of the embodiment of the cold tandem rolling force prediction method based on IDBO-SA-LSTM applied to the online prediction stage also applies to the cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the online prediction stage of this embodiment, and will not be elaborated here.

[0129] The cold tandem rolling force prediction device based on IDBO-SA-LSTM applied to the online prediction stage proposed according to the embodiment of the present application includes an acquisition module 201 for acquiring the actual data during the stable rolling process of the target cold tandem rolling mill; a calculation module 202 for performing normalization and fitness calculation on the actual data to obtain the corresponding fitness calculation index; a prediction module 203 for inputting the actual data into the pre-constructed target SA-LSTM rolling force prediction model, and combining the fitness calculation index to output the rolling force prediction result during the stable rolling process of the target cold tandem rolling mill. Among them, the target SA-LSTM rolling force prediction model is obtained by training the preset training data set and test data set of the cold tandem rolling force through an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm. The present application optimizes the long short-term memory network combined with the self-attention mechanism through the improved dung beetle optimization algorithm, effectively improving the accuracy and generalization performance of the cold tandem rolling force prediction, and realizing the high-precision rolling force prediction in the aluminum foil cold tandem rolling process.

[0130] Figure 12 It is a schematic structural diagram of the electronic device provided by the embodiment of the present application. The electronic device may include:

[0131] A memory 1201, a processor 1202, and a computer program stored on the memory 1201 and executable on the processor 1202.

[0132] When the processor 1202 executes the program, it implements the cold tandem rolling force prediction method based on IDBO - SA - LSTM provided in the above embodiments.

[0133] Furthermore, the electronic device further includes:

[0134] A communication interface 1203 for communication between the memory 1201 and the processor 1202.

[0135] The memory 1201 is used to store a computer program executable on the processor 1202.

[0136] The memory 1201 may include a high - speed RAM memory, and may also include non - volatile memory, such as at least one disk memory.

[0137] If the memory 1201, the processor 1202, and the communication interface 1203 are implemented independently, the communication interface 1203, the memory 1201, and the processor 1202 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 12 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0138] Optionally, in a specific implementation, if the memory 1201, the processor 1202, and the communication interface 1203 are integrated on a chip, the memory 1201, the processor 1202, and the communication interface 1203 can communicate with each other through an internal interface.

[0139] The processor 1202 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0140] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned cold tandem rolling force prediction method based on IDBO-SA-LSTM is implemented.

[0141] An embodiment of the present application also provides a computer program product, including a computer program, and when the computer program is executed, it is used to implement the above-mentioned cold tandem rolling force prediction method based on IDBO-SA-LSTM.

[0142] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0143] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0144] Any process or method description in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiment of the present application includes additional implementations, where the functions can be executed in a manner that is not shown or discussed in the order, including in a substantially simultaneous manner or in the reverse order according to the involved functions, which should be understood by those skilled in the art of the embodiments of the present application.

[0145] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definable list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in conjunction with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection part (electronic device) having one or N wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0146] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or combinations thereof. In the above-described embodiments, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0147] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0148] In addition, each functional unit in various embodiments of the present application may be integrated into one processing module, may exist separately as individual physical units, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0149] The above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A cold rolling force prediction method based on IDBO-SA-LSTM, applied in the offline training stage, characterized in that: The following steps are involved: Based on the preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy and dynamic weight coefficient, the preset dung beetle optimization algorithm is optimized to construct an improved dung beetle optimization algorithm; Performing a performance test on the improved dung beetle optimization algorithm through a plurality of preset benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; Based on the improved dung beetle optimization algorithm, the preset adaptive long short-term memory network is automatically optimized, and the target SA-LSTM rolling force prediction model is constructed by the adaptive long short-term memory network after automatic optimization, so that the rolling force can be predicted in real time using the target SA-LSTM rolling force prediction model in the online prediction stage.

2. The cold rolling force prediction method based on IDBO-SA-LSTM according to claim 1 is characterized in that: The improved dung beetle optimization algorithm is subjected to performance testing by using a plurality of preset benchmark test functions to obtain an improved dung beetle optimization algorithm that meets preset performance test requirements, including: Determine the multiple benchmark test functions, and set the population size and the maximum number of iterations corresponding to each test function in the multiple benchmark test functions in a single iteration, wherein the multiple benchmark test functions include multiple unimodal test functions and multiple multimodal test functions; Based on the population size in the single iteration and the maximum number of iterations, each test function is used to perform a performance test on the improved dung beetle optimization algorithm.

3. The cold rolling force prediction method based on IDBO-SA-LSTM according to claim 1 is characterized in that: Based on the improved dung beetle optimization algorithm, the preset adaptive long short-term memory network is automatically optimized, and the target SA-LSTM rolling force prediction model is constructed through the adaptive long short-term memory network after automatic optimization, including: Based on the preset self-attention mechanism, LSTM layer, activation layer, fully connected layer and regression layer, the adaptive long short-term memory network is constructed; Acquiring raw data of rolling force of tandem cold rolling, and performing random scattering and normalization operations on the raw data of rolling force of tandem cold rolling to construct a target training format data set, and dividing the target training format data set to generate a training data set and a test data set; Initializing multiple parameters in the improved dung beetle optimization algorithm, and based on the Circle chaos mapping strategy, initializing the dung beetle population corresponding to the improved dung beetle optimization algorithm to obtain multiple individual positions, and using the multiple individual positions to train the adaptive long short-term memory network to generate individual positions of target fitness, wherein the multiple parameters include the population number, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the proportion of rolling, reproduction, foraging, and stealing behaviors in the population; Using the improved dung beetle optimization algorithm to update the individual position of the target fitness, and iterating cyclically to the maximum number of iterations to obtain the optimal solution parameters, and training the adaptive long short-term memory network by the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model; The initial SA-LSTM rolling force prediction model is fine-tuned using the test data set to obtain the target SA-LSTM rolling force prediction model.

4. A cold rolling force prediction method based on IDBO-SA-LSTM, applied to the online prediction stage, characterized in that: The following steps are involved: Acquire actual data of the target cold rolling mill during steady rolling; Normalizing and calculating the fitness of the actual data to obtain a corresponding fitness calculation index; The actual data is input into a pre-constructed target SA-LSTM rolling force prediction model, and combined with the fitness calculation index to output the rolling force prediction result of the target cold rolling mill during the stable rolling process, wherein the target SA-LSTM rolling force prediction model is obtained by training a preset cold rolling rolling force training data set and a test data set through an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm.

5. A cold rolling force prediction device based on IDBO-SA-LSTM, applied in the offline training stage, characterized in that: include: An optimization module is used to optimize the preset dung beetle optimization algorithm based on the preset rolling force calculation model, Circle chaos mapping strategy, golden sine strategy and dynamic weight coefficient to construct an improved dung beetle optimization algorithm; A testing module, used to perform a performance test on the improved dung beetle optimization algorithm through a plurality of preset benchmark test functions to obtain an improved dung beetle optimization algorithm that meets the preset performance test requirements; A modeling module is used to automatically optimize a preset adaptive long short-term memory network based on the improved dung beetle optimization algorithm, and to construct a target SA-LSTM rolling force prediction model through the adaptive long short-term memory network after automatic optimization, so as to use the target SA-LSTM rolling force prediction model to perform real-time prediction of rolling force in the online prediction stage.

6. The cold rolling force prediction device based on IDBO-SA-LSTM according to claim 5, characterized in that: The modeling module includes: A construction unit, used to construct the adaptive long short-term memory network based on a preset self-attention mechanism, an LSTM layer, an activation layer, a fully connected layer and a regression layer; A partitioning unit is used to obtain raw data of rolling force of tandem cold rolling, and to perform random scattering and normalization operations on the raw data of rolling force of tandem cold rolling to construct a target training format data set, and to partition the target training format data set to generate a training data set and a test data set; An initialization unit, used to initialize multiple parameters in the improved dung beetle optimization algorithm, and based on the Circle chaos mapping strategy, initialize the dung beetle population corresponding to the improved dung beetle optimization algorithm to obtain multiple individual positions, and use the multiple individual positions to train the adaptive long short-term memory network to generate individual positions of target fitness, wherein the multiple parameters include the number of populations, the maximum number of iterations, the number of neurons in the hidden layer, the initial learning rate, the regularization coefficient, and the proportion of rolling, reproduction, foraging, and stealing behaviors in the population; An iteration unit, used for updating the individual position of the target fitness by using the improved dung beetle optimization algorithm, and cyclically iterating to the maximum number of iterations to obtain the optimal solution parameters, and training the adaptive long short-term memory network by using the optimal solution parameters and the training data set to obtain an initial SA-LSTM rolling force prediction model; A fine-tuning unit is used to fine-tune the initial SA-LSTM rolling force prediction model through the test data set to obtain the target SA-LSTM rolling force prediction model.

7. A cold rolling force prediction device based on IDBO-SA-LSTM, applied to the online prediction stage, characterized in that: include: An acquisition module, used to acquire actual data during the stable rolling process of the target cold rolling mill; A calculation module, used for normalizing and calculating the fitness of the actual data to obtain a corresponding fitness calculation index; A prediction module is used to input the actual data into a pre-constructed target SA-LSTM rolling force prediction model, and combine it with the fitness calculation index to output the rolling force prediction result of the target cold rolling mill during the stable rolling process, wherein the target SA-LSTM rolling force prediction model is obtained by training a preset cold rolling rolling force training data set and a test data set through an adaptive long short-term memory network optimized by a preset improved dung beetle optimization algorithm.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the cold rolling force prediction method based on IDBO-SA-LSTM as described in any one of claims 1 to 3 or claim 4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the cold rolling force prediction method based on IDBO-SA-LSTM as described in any one of claims 1 to 3 or claim 4.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the cold rolling force prediction method based on IDBO-SA-LSTM as described in any one of claims 1 to 3 or claim 4.

Images