Cigarette paper tape splicing prediction method and device, equipment and storage medium

Abstract

The invention discloses a cigarette paper tape splicing prediction method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: determining each parameter value corresponding to each target process parameter of a cigarette paper tape splicing process; determining parameter value groups based on the parameter values, and performing an orthogonal test based on the parameter value groups to obtain target cigarette paper tape splicing success rates corresponding to the parameter value groups; constructing a first target cigarette paper tape splicing prediction model based on a preset sparrow search algorithm, a preset back propagation neural network model, the parameter value group and the target cigarette paper tape splicing success rate; constructing a second target cigarette paper tape splicing prediction model based on the optimized preset sparrow search algorithm, the preset back propagation neural network model, the parameter value group and the target cigarette paper tape splicing success rate; and predicting the splicing success rate of the cigarette paper tape by using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model. The method can accurately predict the splicing success rate of the cigarette paper tape.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a cigarette paper strip splicing prediction method, device, equipment and storage medium. Background Art

[0002] At present, in tobacco rolling and splicing equipment, the stable supply of cigarette paper is not only the basic condition to ensure the efficient and continuous operation of the entire production line, but also the key factor to ensure product quality and production efficiency. Cigarette paper tape is a key raw and auxiliary material of the rolling and splicing unit, and its automatic splicing is a key technology for replacing cigarette paper trays. However, in actual applications, the automatic splicing of paper tapes during the tray change process faces many challenges. Due to the influence of various factors such as paper tape material properties, environmental humidity, tension control, and splicing mechanical accuracy, it is difficult to determine and apply the optimal paper tape splicing process parameters at one time. This leads to frequent undesirable phenomena such as breakage, tailing, and tearing during the paper tape splicing process, which not only affects the continuous operation of the production line, but also increases the scrap rate and equipment downtime.

[0003] In summary, how to accurately predict the success rate of cigarette paper strip splicing under different process parameters is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a cigarette paper strip splicing prediction method, device, equipment and storage medium, which can accurately predict the success rate of cigarette paper strip splicing under different process parameters. The specific scheme is as follows:

[0005] In a first aspect, the present application provides a cigarette paper strip splicing prediction method, comprising:

[0006] Determining target process parameters of the cigarette paper strip splicing process, and determining parameter values corresponding to the target process parameters;

[0007] Determining a parameter value group based on the parameter value, and using a target tobacco rolling and joining device to perform an orthogonal test based on the parameter value group to obtain a target cigarette paper strip splicing success rate corresponding to each parameter value group;

[0008] Constructing a first target cigarette paper strip splicing prediction model based on a preset sparrow search algorithm, a preset back propagation neural network model, the parameter value group and the target cigarette paper strip splicing success rate corresponding to the parameter value group;

[0009] Optimize several algorithm stages in the preset sparrow search algorithm according to the preset phased collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group;

[0010] Use the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model to predict the success rate of cigarette paper splicing.

[0011] Optionally, determining the parameter value group based on the parameter value, and using the target tobacco rod-making and tipping equipment to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group, including:

[0012] Use the preset statistical analysis software to determine the parameter value group based on the preset orthogonal table, the target process parameters, and the parameter values corresponding to the target process parameters, and generate a corresponding target orthogonal experiment plan table based on the parameter value group; wherein, the target process parameters include the target gasket thickness, the target old disk speed difference, the target old disk deceleration angle, the target pre-tightening speed difference, and the target pre-tightening overspeed angle, the pre-tightening speed difference is the difference between the running speeds of the pre-tightening roller and the new disk in the target tobacco rod-making and tipping equipment, the pre-tightening overspeed angle is the angle difference between the pre-tightening roller when it overspeed and the cigarette paper splicing device in the target tobacco rod-making and tipping equipment, and the new disk and the old disk in the target tobacco rod-making and tipping equipment are the reels for storing cigarette paper;

[0013] Use the target tobacco rod-making and tipping equipment to conduct an orthogonal experiment based on each parameter value group in the target orthogonal experiment plan table to obtain the target cigarette paper splicing success rate corresponding to each parameter value group.

[0014] Optionally, after using the target tobacco rod-making and tipping equipment to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group, it further includes:

[0015] Screen out the parameter value groups with the target parameter value from each parameter value group to obtain each target parameter value group corresponding to the target parameter value; the target parameter value is any one of the parameter values;

[0016] Sum up the target cigarette paper splicing success rates corresponding to each target parameter value group corresponding to the target parameter value to obtain the target sum value corresponding to the target parameter value, and determine the target test times corresponding to the target parameter value based on the number of groups of the target parameter value group corresponding to the target parameter value;

[0017] Determine the average success rate of target cigarette paper splicing corresponding to the target parameter value based on the target number of trials corresponding to the target parameter value and the target sum value corresponding to the target parameter value;

[0018] Determine the maximum and minimum values among the average success rates of target cigarette paper splicing corresponding to each of the target parameter values, and determine the target range of the target process parameters based on the maximum value and the minimum value, so as to determine the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model based on the target range.

[0019] Optionally, the constructing of the first target cigarette paper splicing prediction model based on the preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group includes:

[0020] Determine the target input layer based on the target quantity corresponding to the target process parameter, and determine the target hidden layer and the target output layer;

[0021] Determine the first initial weight and the first initial threshold based on the preset random initialization method, and determine the target activation function;

[0022] Construct a first initial model according to the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the first initial weight, and the first initial threshold;

[0023] Optimize the first initial weight and the first initial threshold based on the preset sparrow search algorithm and the preset error threshold to obtain the second initial weight and the second initial threshold;

[0024] Use the first initial model to determine the first output result corresponding to the parameter value group based on the preset forward propagation method, and determine the first target error between the first output result and the target cigarette paper splicing success rate corresponding to the parameter value group based on the preset backpropagation method, so as to adjust the second initial weight and the second initial threshold based on the first target error to obtain the corresponding first target weight and the first target threshold;

[0025] Determine the first target cigarette paper splicing prediction model based on the first initial model, the first target weight, and the first target threshold.

[0026] Optionally, the optimizing of several algorithm stages in the preset sparrow search algorithm according to the preset phased collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm includes:

[0027] Initialize the sparrow population based on a preset chaotic map, and determine the target optimization dimension of each sparrow in the sparrow population based on a preset Levy flight strategy and a preset opposition-based learning mechanism during the exploration phase of the preset sparrow search algorithm;

[0028] Update the position information of each sparrow according to a preset grey wolf optimization algorithm, a preset differential evolution crossover mutation algorithm, and the target optimization dimension during the exploitation phase of the preset sparrow search algorithm;

[0029] Determine the target position information based on a preset gradient descent algorithm, a preset population dynamic reduction strategy, and the position information of each sparrow during the convergence phase of the preset sparrow search algorithm, so as to determine the fitness value of the sparrow based on the target position information, and determine the target solution corresponding to the preset sparrow search algorithm based on the fitness value and the preset error threshold.

[0030] Optionally, constructing a second target cigarette paper tape splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper tape splicing success rate corresponding to the parameter value group includes:

[0031] Determine the third initial weight and the third initial threshold based on the preset random initialization method, and construct a second initial model based on the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the third initial weight, and the third initial threshold;

[0032] Optimize the third initial weight and the third initial threshold based on the optimized preset sparrow search algorithm and the preset error threshold to obtain the fourth initial weight and the fourth initial threshold;

[0033] Use the second initial model to determine the second output result corresponding to the parameter value group based on the preset forward propagation method, and determine the second target error between the second output result and the target cigarette paper tape splicing success rate corresponding to the parameter value group based on the preset backpropagation method, so as to adjust the fourth initial weight and the fourth initial threshold based on the second target error to obtain the corresponding second target weight and the second target threshold;

[0034] Determine the second target cigarette paper tape splicing prediction model based on the second initial model, the second target weight, and the second target threshold.

[0035] Optionally, predicting the success rate of cigarette paper tape splicing using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model includes:

[0036] Determine a parameter value group to be verified from the parameter value groups in the orthogonal experiment based on a preset quantity condition, and use the first target cigarette paper tape splicing prediction model to determine the first cigarette paper tape splicing prediction result corresponding to the parameter value group to be verified, and use the second target cigarette paper tape splicing prediction model to determine the second cigarette paper tape splicing prediction result corresponding to the parameter value group to be verified;

[0037] Determine the target cigarette paper tape splicing success rate corresponding to the parameter value group to be verified, determine the first prediction error between the first cigarette paper tape splicing prediction result and the target cigarette paper tape splicing success rate, and determine the second prediction error between the second cigarette paper tape splicing prediction result and the target cigarette paper tape splicing success rate;

[0038] Evaluate the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model based on the first prediction error and the second prediction error.

[0039] Second, this application provides a cigarette paper tape splicing prediction device, including:

[0040] A parameter value determination module, configured to determine the target process parameters of the cigarette paper tape splicing process and determine the parameter values corresponding to each of the target process parameters;

[0041] A success rate determination module, configured to determine a parameter value group based on the parameter values, and use a target tobacco rod-making and tipping machine to perform an orthogonal experiment based on the parameter value group to obtain the target cigarette paper tape splicing success rate corresponding to each parameter value group;

[0042] A first model construction module, configured to construct a first target cigarette paper tape splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper tape splicing success rate corresponding to the parameter value group;

[0043] A second model construction module, configured to optimize several algorithm stages in the preset sparrow search algorithm according to a preset staged collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper tape splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper tape splicing success rate corresponding to the parameter value group;

[0044] A cigarette paper tape splicing prediction module, configured to predict the cigarette paper tape splicing success rate by using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model.

[0045] Third, this application provides an electronic device, including:

[0046] A memory for storing a computer program;

[0047] A processor for executing the computer program to implement the aforementioned cigarette paper splicing prediction method.

[0048] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the aforementioned cigarette paper splicing prediction method is implemented.

[0049] In the present application, first, the target process parameters of the cigarette paper splicing process are determined, and the parameter values corresponding to each of the target process parameters are determined; then, based on the parameter values, a parameter value group is determined, and an orthogonal experiment is carried out based on the parameter value group by using a target tobacco making and tipping machine to obtain the target cigarette paper splicing success rate corresponding to each parameter value group; then, a first target cigarette paper splicing prediction model is constructed based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; then, according to a preset staged collaborative optimization mechanism, several algorithm stages in the preset sparrow search algorithm are optimized to obtain the optimized preset sparrow search algorithm, and a second target cigarette paper splicing prediction model is constructed based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; finally, the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model are used to predict the cigarette paper splicing success rate. As can be seen from the above, in the present application, first, the target process parameters in the cigarette paper splicing process are determined, secondly, the parameter values of each target process parameter are determined, an orthogonal experiment is designed and carried out based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group, then a first target cigarette paper splicing prediction model is constructed according to the preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group, and an optimized preset sparrow search algorithm is introduced to construct a second target cigarette paper splicing prediction model to predict the cigarette paper splicing success rate by using the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model. In this way, the cigarette paper splicing prediction model constructed in the present application can accurately predict the cigarette paper splicing success rate under different process parameters, and can effectively improve the cigarette paper splicing success rate of the tobacco making and tipping machine. Description of the Drawings

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on the provided drawings.

[0051] Figure 1 Flowchart of a method for predicting the splicing of cigarette paper tapes provided by this application;

[0052] Figure 2 Schematic diagram of a specific failure of cigarette paper tape splicing provided by this application;

[0053] Figure 3 Schematic diagram of a specific tobacco rod-making and tipping machine provided by this application;

[0054] Figure 4 Schematic diagram of a specific backpropagation neural network structure provided by this application;

[0055] Figure 5 Schematic diagram of a specific model construction process provided by this application;

[0056] Figure 6 Schematic diagram of a specific model construction process provided by this application;

[0057] Figure 7 Schematic diagram of a specific model evaluation result provided by this application;

[0058] Figure 8 Flowchart of a method for predicting the splicing of cigarette paper tapes provided by this application;

[0059] Figure 9 Flowchart of a method for predicting the splicing of cigarette paper tapes provided by this application;

[0060] Figure 10 Schematic diagram of a specific optimization algorithm evaluation result provided by this application;

[0061] Figure 11 Schematic diagram of the structure of a device for predicting the splicing of cigarette paper tapes provided by this application;

[0062] Figure 12 Schematic diagram of the structure of an electronic device provided by this application. Detailed implementation manners

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] At present, in tobacco rolling and splicing equipment, the stable supply of cigarette paper is not only the basic condition to ensure the efficient and continuous operation of the entire production line, but also a key factor to ensure product quality and production efficiency. As the key raw and auxiliary materials of the rolling and splicing unit, the automated splicing of cigarette paper tape is the key technology for replacing cigarette paper trays. However, in actual applications, the automated splicing of paper tapes during the tray change process faces many challenges. Due to the influence of various factors such as paper tape material properties, environmental humidity, tension control, and splicing mechanical accuracy, it is difficult to determine and apply the optimal paper tape splicing process parameters at one time. This leads to frequent undesirable phenomena such as breakage, tailing, and tearing during the paper tape splicing process, which not only affects the continuous operation of the production line, but also increases the scrap rate and equipment downtime. To this end, the present application provides a cigarette paper tape splicing prediction scheme that can accurately predict the success rate of cigarette paper tape splicing under different process parameters.

[0065] See also Figure 1 As shown, the embodiment of the present invention discloses a cigarette paper strip splicing prediction method, which may include:

[0066] Step S11, determining target process parameters of the cigarette paper strip splicing process, and determining parameter values corresponding to the target process parameters.

[0067] See also Figure 2 As shown in the figure, due to the lack of suitable cigarette paper strip splicing process parameters in the tobacco rolling equipment, the paper strip is prone to breakage, tailing, tearing, etc. during splicing. It is understandable that in the cigarette paper strip splicing process, there are many process parameters that have a significant impact on the success rate of cigarette paper strip splicing. Figure 3 As shown, the components used in the target tobacco rolling and joining equipment include a tension control device, a pre-tightening roller, a tape-free splicing device, etc. After research, the gasket thickness H, the old disc speed difference Vs, and the old disc deceleration angle can be , preload speed difference Vw, preload overspeed angle The five process parameters are determined as the target process parameters of the cigarette paper strip splicing process, wherein the preload speed difference is the difference in running speed between the preload roller and the new disc in the target tobacco rolling and splicing equipment, the preload overspeed angle is the angle difference between the preload roller in the target tobacco rolling and splicing equipment and the cigarette paper strip splicing device when the preload roller is overspeeding, and the new disc and the old disc in the target tobacco rolling and splicing equipment are the reels for storing cigarette paper. Specifically, the gasket thickness H can be set to: , the speed difference Vs of the old disk can be set to: , the deceleration angle of the old disk can be set to: , the pre-tightening speed difference Vw can be set to: , the pre-tightening overspeed angle can be set to: , the cigarette paper splicing technical parameters in the tobacco rod-making and -tipping equipment are shown in Table 1.

[0068] Table 1

[0069]

[0070] Step S12: Determine a parameter value group based on the parameter values, and use the target tobacco rod-making and -tipping equipment to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper tape splicing success rate corresponding to each parameter value group.

[0071] In this embodiment, the above-mentioned determining a parameter value group based on the parameter values and using the target tobacco rod-making and -tipping equipment to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper tape splicing success rate corresponding to each parameter value group may include: First, use a preset statistical analysis software to determine the parameter value group based on a preset orthogonal table, the target process parameters, and the parameter values corresponding to the target process parameters, and generate a corresponding target orthogonal experiment plan table based on the parameter value group; then use the target tobacco rod-making and -tipping equipment to conduct an orthogonal experiment based on each parameter value group in the target orthogonal experiment plan table to obtain the target cigarette paper tape splicing success rate corresponding to each parameter value group. Specifically, as shown in Table 2, first, an orthogonal table of five factors and three levels can be designed, and an orthogonal experiment of five factors and three levels can be designed using SPSS software to generate 27 groups of parameter value groups to obtain the target orthogonal experiment plan table. Then, as shown in Table 3, use the target tobacco rod-making and -tipping equipment to conduct an orthogonal experiment based on each parameter value group in the target orthogonal experiment plan table, with each group being conducted 20 times, to obtain the target cigarette paper tape splicing success rate.

[0072] Table 3

[0073]

[0074] Table 2

[0075]

[0076] Step S13: Construct a first target cigarette paper tape splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper tape splicing success rate corresponding to the parameter value group.

[0077] In this embodiment, constructing the first target cigarette paper splicing prediction model based on the preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group may include: First, determine the target input layer based on the target quantity corresponding to the target process parameters, and determine the target hidden layer and the target output layer; Then, determine the first initial weight and the first initial threshold based on the preset random initialization method, and determine the target activation function; Then, construct the first initial model according to the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the first initial weight, and the first initial threshold; Then, optimize the first initial weight and the first initial threshold based on the preset sparrow search algorithm and the preset error threshold to obtain the second initial weight and the second initial threshold; Next, use the first initial model to determine the first output result corresponding to the parameter value group based on the preset forward propagation method, and determine the first target error between the first output result and the target cigarette paper splicing success rate corresponding to the parameter value group based on the preset backpropagation method, so as to adjust the second initial weight and the second initial threshold based on the first target error to obtain the corresponding first target weight and the first target threshold; Finally, determine the first target cigarette paper splicing prediction model based on the first initial model, the first target weight, and the first target threshold. Specifically, a BP neural network (Backpropagation Neural Network, that is, a backpropagation neural network) and a sparrow search algorithm can be used to construct the first target cigarette paper splicing prediction model. See Figure 4 As shown, the BP neural network, as a typical model of a multi-layer feedforward neural network, has two key training stages: one is the forward propagation of data, which realizes the layer-by-layer transfer of information from the input layer to the output layer; the other is the backward propagation of errors, which calculates the difference between the loss function and the target output, and adjusts the network weights to reduce the prediction error. The backpropagation algorithm consists of an input layer, a hidden layer, and an output layer, and the layers are connected by weights and thresholds. In the initialization stage of the algorithm, the weights and thresholds are randomly set, the built-in activation function is used to predict the target parameters, and the weights and thresholds are adjusted according to the error between the actual data and the model parameters in the model until the expected goal is reached, and the establishment of the prediction model is completed. According to the characteristics of the test data with multiple inputs and a single output, it can be determined that the target input layer has five layers and the target output layer has one layer. The number of hidden layer nodes is an important factor affecting the prediction ability of the BP neural network. When the number of hidden layer nodes is too small, the learning ability and information processing ability will be significantly weakened. When the number of hidden layer nodes is too large, the structural complexity will increase, resulting in an extended learning cycle, a reduced convergence speed, a decreased prediction accuracy, and an easy occurrence of overfitting. According to experience, the calculation formula for the number of hidden layer nodes is as follows:

[0078] ;

[0079] where is the number of hidden layer nodes, n is the number of input layer nodes, m is the number of output layer nodes, and a is a constant within 10. The strong capture ability of the BP neural network enables it to map complex non-linear relationships and is suitable for predicting the success rate of paper tape splicing. However, its weights are gradually fine-tuned along the local optimization path, resulting in a stronger local search ability than the global search ability, and it is easy to fall into the local optimum rather than the global optimum, which affects data prediction. Therefore, the sparrow search algorithm can be used to improve the performance of the BP neural network. SSA (Sparrow Search Algorithm) is a new swarm intelligence optimization strategy based on the foraging and anti-predation behaviors of sparrows. The sparrow search algorithm simulates three key roles in the foraging process of sparrows: seekers, joiners, and scouts. Among them, the seekers are responsible for exploring the food source. Once the warning signal exceeds the safety threshold, they will lead the whole group to transfer to a safe area. During each iteration, the position change of the seekers is as follows:

[0080] ;

[0081] where t represents the current iteration number; represents the maximum number of iterations; represents the position information of the i-th sparrow in the j-th dimension at the t-th iteration. The joiners will follow the seekers to compete for food and dynamically adjust the fitness value. The position of the joiners will also change with the competition for food. The position update of the joiners is as follows:

[0082] ;

[0083] where represents the best position of the seekers at the (t + 1)-th iteration; represents the worst position in the current sparrow population. In addition, randomly select individuals in the group as scouts. Once the scouts sense a dangerous situation and send a signal, the population will migrate to a new foraging point according to this signal and adjust and optimize the current position of the scouts. The optimization formula is as follows:

[0084] ;

[0085] where represents the best position of the sparrows at the t-th iteration; k represents a random number; is the compensation control parameter, which is a random number obeying a normal distribution with a mean of 0 and a variance of 1. SeeFigure 5 As shown in the figure, the basic idea of the SSA-BP algorithm is to use the initial weights and thresholds of the BP neural network as the optimization targets of the SSA algorithm, and the fitness value in the prediction algorithm as the fitness value in the SSA algorithm. After multiple iterations, the obtained results are assigned to the BP neural network algorithm to enhance the optimization ability of the BP neural network and thus improve the accuracy of the prediction results.

[0086] Step S14, optimizing several algorithm stages in the preset sparrow search algorithm according to a preset staged collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and constructing a second target cigarette paper strip splicing prediction model based on the optimized preset sparrow search algorithm, the preset back propagation neural network model, the parameter value group and the target cigarette paper strip splicing success rate corresponding to the parameter value group.

[0087] It is understandable that SSA randomly sets the initial population in the search space, which may cause problems such as uneven population distribution and insufficient search space. This embodiment can initialize the sparrow population based on a preset chaotic map. Specifically, chaos is a method with ergodic and random properties. The Tent chaotic map can better initialize the sparrow population of SSA to achieve the purpose of uniformly distributing the initial solution, enhancing population diversity and preventing the population from falling into the local optimum. At the same time, the rand random variable can be increased to reduce its instability. The formula is as follows:

[0088] ;

[0089] It should be noted that, in the present embodiment, the target optimization dimension of each sparrow in the sparrow population can be determined based on the preset Levy flight strategy and the preset reverse learning mechanism during the exploration phase of the preset sparrow search algorithm. Specifically, Levy flight can be responsible for global coarse search, expanding the search range by long-distance jumps to cover potential areas far from the current solution. Reverse learning can be responsible for local fine search, expanding the search area in a symmetrical direction for individuals with poor fitness, and exploring the optimal solution. The implementation steps of Levy flight are: first select the seeker, sort it according to fitness, and select the front The optimal individuals are used as discoverers, and N is the population size. The Levy flight step length is as follows:

[0090] ;

[0091] in, is the adaptive perturbation coefficient, which increases linearly with the number of iterations t. The perturbation is small in the initial stage and the exploration is enhanced in the later stage. is a random step size that follows the Levy distribution. In this embodiment, the position information of each sparrow can be updated according to a preset grey wolf optimization algorithm, a preset differential evolution crossover mutation algorithm, and the target optimization dimension during the development stage of the preset sparrow search algorithm. Specifically, the GWO multi-level guidance mechanism (GWO, i.e., Grey Wolf Optimizer) can utilize the three-level leading individuals to guide the followers to update and balance exploration and development. The first step: Select the leading individuals: According to the fitness ranking, the top 3 are respectively the best wolf, the second-best wolf, the third-best wolf. The second step: Calculate the weight coefficients: Dynamically allocate the weights of . The third step: Multi-level position update: The followers move towards the weighted center of the three-level leading individuals and add random perturbations. The position update formula for the joiners is as follows:

[0092] ;

[0093] In this embodiment, new solutions can be generated through DE mutation and crossover to enhance the local development ability. The first step: Mutation operation: Randomly select individuals to generate mutation vectors. The second step: Crossover operation: Mix the mutation vectors with the original individuals to generate trial vectors. The third step: Selection operation: Retain the better solutions to enter the next generation. And in this embodiment, the target position information can be determined based on a preset gradient descent algorithm, a preset population dynamic reduction strategy, and the position information of each sparrow during the convergence stage of the preset sparrow search algorithm, so as to determine the fitness value of the sparrows based on the target position information, and determine the target solution corresponding to the preset sparrow search algorithm based on the fitness value and the preset error threshold. Specifically, the gradient descent algorithm can be introduced in the local development stage of SSA to accelerate the process of approaching the optimal solution and reduce the blindness of random search. The first step: Select elite individuals: In each iteration, select several individuals with the best fitness in the current population, such as the first as candidates; The second step: Calculate the gradient direction: Calculate the gradient of the objective function for each elite individual. For high-dimensional problems, the finite difference method can be used to approximate the gradient; The third step: Gradient descent update: Update the positions of the elite individuals along the negative gradient direction; The fourth step: Boundary processing and fitness evaluation: Ensure that the new positions are within the search space and calculate the fitness. If the new positions are better, replace the original individuals. The population dynamic reduction strategy can gradually reduce the population size as the algorithm progresses, concentrate resources on optimizing elite individuals, and improve the calculation efficiency. The first step: Initial setting: Set the maximum population size and the minimum size. The second step: Reduction condition: Dynamically adjust the population size according to the number of iterations or fitness changes. The third step: Individual elimination mechanism: Randomly retain a few individuals with low fitness but large differences, such as , to avoid population homogenization.

[0094] In this embodiment, there is a problem that the first target cigarette paper tape splicing prediction model constructed based on the SSA and BP neural network is prone to converge to a local optimum during the training process, resulting in a decrease in the accuracy of the results. The basic idea of improving the SSA-BP algorithm is to use the initial weights and thresholds of the BP neural network as the optimization objectives of the optimized SSA algorithm, and use the fitness value in the prediction algorithm as the fitness value in the optimized SSA algorithm. After multiple iterations, the obtained results are assigned to the BP neural network to enhance the global search ability and convergence accuracy of the SSA-BP neural network, thereby improving the accuracy of the prediction results. Specifically, see Figure 6 As shown, first, determine the third initial weights and the third initial threshold based on a preset random initialization method, and construct a second initial model based on a preset backpropagation neural network model, a target input layer, a target hidden layer, a target output layer, a target activation function, the third initial weights, and the third initial threshold; then optimize the third initial weights and the third initial threshold based on the optimized preset sparrow search algorithm and a preset error threshold to obtain the fourth initial weights and the fourth initial threshold; then use the second initial model to determine the second output result corresponding to the parameter value group based on a preset forward propagation method, and determine the second target error between the second output result and the target cigarette paper tape splicing success rate corresponding to the parameter value group based on a preset backpropagation method, so as to adjust the fourth initial weights and the fourth initial threshold based on the second target error to obtain the corresponding second target weights and the second target threshold; finally, determine the second target cigarette paper tape splicing prediction model based on the second initial model, the second target weights, and the second target threshold.

[0095] In a specific implementation manner, see Table 4 and Figure 7 As shown, use CSSA-BPNN to represent the second target cigarette paper tape splicing prediction model, input the orthogonal experimental data into BPNN and CSSA-BPNN respectively for prediction, randomly select 22 groups of orthogonal experimental data as the training set, put them into the two models for training respectively, and select 5 groups of orthogonal experimental data as the test set, input them into the two models respectively, and compare the true value with the predicted value of the cigarette paper tape splicing success rate output by the model to obtain the corresponding model evaluation results.

[0096] Table 4

[0097]

[0098] Step S15: Use the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model to predict the success rate of cigarette paper tape splicing.

[0099] In this embodiment, in order to verify the performance of the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model, first, a parameter value group to be verified can be determined from the parameter value groups in the orthogonal experiment based on a preset quantity condition, and the first cigarette paper tape splicing prediction result corresponding to the parameter value group to be verified can be determined by using the first target cigarette paper tape splicing prediction model, and the second cigarette paper tape splicing prediction result corresponding to the parameter value group to be verified can be determined by using the second target cigarette paper tape splicing prediction model; then, the success rate of the target cigarette paper tape splicing corresponding to the parameter value group to be verified is determined, the first prediction error between the first cigarette paper tape splicing prediction result and the success rate of the target cigarette paper tape splicing is determined, and the second prediction error between the second cigarette paper tape splicing prediction result and the success rate of the target cigarette paper tape splicing is determined; finally, the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model are evaluated based on the first prediction error and the second prediction error.

[0100] As can be seen from the above, in this embodiment, the target process parameters of the cigarette paper splicing process are first determined, and the parameter values corresponding to each of the target process parameters are determined; then, a parameter value group is determined based on the parameter values, and an orthogonal experiment is carried out based on the parameter value group by using the target tobacco rolling and tipping equipment to obtain the target cigarette paper splicing success rate corresponding to each parameter value group; then, a first target cigarette paper splicing prediction model is constructed based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; then, according to a preset phased collaborative optimization mechanism, several algorithm stages in the preset sparrow search algorithm are optimized to obtain the optimized preset sparrow search algorithm, and a second target cigarette paper splicing prediction model is constructed based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; finally, the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model are used to predict the cigarette paper splicing success rate. As can be seen from the above, in this embodiment, the target process parameters in the cigarette paper splicing process are first determined, and then the parameter values of each target process parameter are determined. An orthogonal experiment is designed and carried out based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group. Then, a first target cigarette paper splicing prediction model is constructed according to a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group, and an optimized preset sparrow search algorithm is introduced to construct a second target cigarette paper splicing prediction model to predict the cigarette paper splicing success rate by using the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model. In this way, the cigarette paper splicing prediction model constructed in this embodiment can accurately predict the cigarette paper splicing success rate under different process parameters and can effectively improve the cigarette paper splicing success rate of the tobacco rolling and tipping equipment.

[0101] See Figure 8 As shown, in order to improve the cigarette paper splicing success rate of the tobacco rolling and tipping equipment, an embodiment of the present invention further discloses a cigarette paper splicing prediction method, which may include:

[0102] Step S21: Determine the target process parameters of the cigarette paper splicing process, and determine the parameter values corresponding to each of the target process parameters.

[0103] Step S22: Determine a parameter value group based on the parameter values, and use the target tobacco rolling and tipping equipment to perform an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group.

[0104] In this embodiment, after using the target tobacco tipping equipment to conduct an orthogonal experiment based on the parameter value groups to obtain the target cigarette paper splicing success rates corresponding to the parameter value groups, the following steps may further be included: First, screen out the parameter value groups with target parameter values from each of the parameter value groups to obtain each target parameter value group corresponding to the target parameter value; the target parameter value is any one of the parameters in the parameters; then sum up the target cigarette paper splicing success rates corresponding to each of the target parameter value groups corresponding to the target parameter value to obtain a target sum value corresponding to the target parameter value, and determine the target number of experiments corresponding to the target parameter value based on the number of the target parameter value groups corresponding to the target parameter value; then determine the average value of the target cigarette paper splicing success rate corresponding to the target parameter value based on the target number of experiments corresponding to the target parameter value and the target sum value corresponding to the target parameter value; finally, determine the maximum and minimum values among the average values of the target cigarette paper splicing success rates corresponding to each of the target parameter values, and determine the target range of the target process parameters based on the maximum and minimum values, so as to determine the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model based on the target range. In a specific implementation manner, as shown in Table 5, the range R of the gasket thickness H is 32.5, the range R of the old disc speed difference Vs is 5.63, the range R of the old disc deceleration angle is 7.99, the range R of the pre-tightening speed difference Vw is 9.03, the range R of the pre-tightening overspeed angle is 4.44. Therefore, in this embodiment, by comparing the ranges of each process parameter, the significant degree of the influence of each process parameter on the cigarette paper splicing success rate can be judged. The larger the range, the more significant the influence of the process parameter on the cigarette paper splicing success rate.

[0105] Table 5

[0106]

[0107] Step S23: Construct a first target cigarette paper splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value groups, and the target cigarette paper splicing success rates corresponding to the parameter value groups.

[0108] Step S24: Optimize several algorithm stages in the preset sparrow search algorithm according to a preset phased collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value groups, and the target cigarette paper splicing success rates corresponding to the parameter value groups.

[0109] Step S25: Predict the splicing success rate of the cigarette paper tape by using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model.

[0110] In a specific embodiment, as shown in Figure 9 The process of the cigarette paper tape splicing prediction method may specifically be as follows: First, in the cigarette paper tape splicing process, determine several main process parameters that affect the splicing success rate of the cigarette paper tape; Secondly, determine the level values of each process parameter, design and conduct an orthogonal experiment in SPSS software, record the splicing success rate of each group of experiments, and then perform a range analysis on the collected data to evaluate the significance of each factor; Then, according to the characteristics of the experimental data having multiple inputs and a single output, construct an SSA-BP neural network model to predict the splicing success rate, and introduce a phased collaborative optimization mechanism to improve the sparrow search algorithm to optimize the BP neural network; Finally, randomly select 5 groups of orthogonal experimental data, substitute them into the SSA-BP neural network and the improved SSA-BP neural network respectively for verification, and compare the cigarette paper tape splicing prediction errors of the two models.

[0111] In a specific embodiment, as shown in Table 6, 6 different benchmark test functions can be used to evaluate different optimization algorithms. The benchmark test functions include unimodal functions and multimodal functions. The average fitness values of the optimization algorithms PSO (Particle Swarm Optimization), WOA (Whale Optimization Algorithm), GWO (Grey Wolf Optimizer), SSA (Sparrow Search Algorithm), and CSSA (Chaotic Sparrow Search Algorithm) changing with the number of iterations on different benchmark test functions are shown in Figure 10 As shown. For unimodal functions and multimodal functions, the convergence speed and optimization accuracy of CSSA are the highest. At the beginning of the search, it can quickly traverse the search space, making the search period of the algorithm shorter; during the search process, it can get closer to the optimal value of the function, making the search accuracy of the algorithm higher. And during the optimization process of the above 6 benchmark test functions, the fitness convergence of CSSA is roughly the same, which can show the strong robustness of CSSA.

[0112] Table 6

[0113]

[0114] Among them, for the more specific processing procedures of the above steps S21, S23, and S24, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0115] As can be seen from the above, in this embodiment, each target parameter value group can be determined from each parameter value group based on the target parameter value. Then, the average value of the target cigarette paper tape splicing success rates corresponding to each target parameter value group corresponding to the target parameter value is determined, and the maximum and minimum values among the average values of the target cigarette paper tape splicing success rates corresponding to each target parameter value are determined to obtain the target range of the target process parameter. In this way, in this embodiment, by comparing the ranges of each process parameter, the significant degree of the influence of each process parameter on the cigarette paper tape splicing success rate can be judged.

[0116] Correspondingly, as shown in Figure 11 the embodiments of the present application further provide a cigarette paper tape splicing prediction device, which may include:

[0117] A parameter value determination module 11, configured to determine the target process parameters of the cigarette paper tape splicing process and determine the parameter values corresponding to each of the target process parameters;

[0118] A success rate determination module 12, configured to determine parameter value groups based on the parameter values, and use a target tobacco rod-making and tipping machine to perform an orthogonal experiment based on the parameter value groups to obtain the target cigarette paper tape splicing success rates corresponding to each of the parameter value groups;

[0119] A first model construction module 13, configured to construct a first target cigarette paper tape splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value groups, and the target cigarette paper tape splicing success rates corresponding to the parameter value groups;

[0120] A second model construction module 14, configured to optimize several algorithm stages in the preset sparrow search algorithm according to a preset staged collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper tape splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value groups, and the target cigarette paper tape splicing success rates corresponding to the parameter value groups;

[0121] A cigarette paper tape splicing prediction module 15, configured to predict the cigarette paper tape splicing success rate by using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model.

[0122] As can be seen from the above, in this application, the target process parameters of the cigarette paper splicing process are first determined, and the parameter values corresponding to each of the target process parameters are determined; then, based on the parameter values, a parameter value group is determined, and an orthogonal experiment is carried out using the target tobacco tipping machine based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group; then, a first target cigarette paper splicing prediction model is constructed based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; then, according to a preset phased collaborative optimization mechanism, several algorithm stages in the preset sparrow search algorithm are optimized to obtain the optimized preset sparrow search algorithm, and a second target cigarette paper splicing prediction model is constructed based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; finally, the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model are used to predict the cigarette paper splicing success rate. As can be seen from the above, in this application, the target process parameters in the cigarette paper splicing process are first determined, and then the parameter values of each target process parameter are determined. An orthogonal experiment is designed and carried out based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group. Then, a first target cigarette paper splicing prediction model is constructed according to the preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group, and an optimized preset sparrow search algorithm is introduced to construct a second target cigarette paper splicing prediction model to predict the cigarette paper splicing success rate using the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model. In this way, the cigarette paper splicing prediction model constructed in this application can accurately predict the cigarette paper splicing success rate under different process parameters and can effectively improve the cigarette paper splicing success rate of the tobacco tipping machine.

[0123] In some specific embodiments, the success rate determination module 12 may include:

[0124] A target orthogonal test plan table generation unit is configured to use a preset statistical analysis software to determine a parameter value group based on a preset orthogonal table, the target process parameters, and the parameter values corresponding to the target process parameters, and generate a corresponding target orthogonal test plan table based on the parameter value group; wherein, the target process parameters include a target gasket thickness, a target old disk speed difference, a target old disk deceleration angle, a target pre-tightening speed difference, and a target pre-tightening overspeed angle, the pre-tightening speed difference is the difference between the running speeds of the pre-tightening roller and the new disk in the target tobacco making and tipping equipment, the pre-tightening overspeed angle is the angle difference between the pre-tightening roller when it is overspeed and the cigarette paper splicing device in the target tobacco making and tipping equipment, and the new disk and the old disk in the target tobacco making and tipping equipment are reels for storing cigarette paper;

[0125] A success rate determination unit is configured to use the target tobacco making and tipping equipment to perform orthogonal tests based on each parameter value group in the target orthogonal test plan table to obtain the target cigarette paper splicing success rate corresponding to each parameter value group.

[0126] In some specific embodiments, the cigarette paper splicing prediction device may further include:

[0127] A target parameter value group determination module is configured to screen out the parameter value groups with target parameter values from each parameter value group to obtain each target parameter value group corresponding to the target parameter value; the target parameter value is any one of the parameter values;

[0128] A target test times determination module is configured to sum the target cigarette paper splicing success rates corresponding to each target parameter value group corresponding to the target parameter value to obtain a target sum value corresponding to the target parameter value, and determine the target test times corresponding to the target parameter value based on the number of groups of the target parameter value groups corresponding to the target parameter value;

[0129] A success rate mean determination module is configured to determine the target cigarette paper splicing success rate mean corresponding to the target parameter value based on the target test times corresponding to the target parameter value and the target sum value corresponding to the target parameter value;

[0130] A target range determination module is configured to determine the maximum value and the minimum value among the target cigarette paper splicing success rate means corresponding to each target parameter value, and determine the target range of the target process parameters based on the maximum value and the minimum value, so as to determine the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model based on the target range.

[0131] In some specific embodiments, the first model construction module 13 may include:

[0132] A target input layer determination unit, configured to determine a target input layer based on the target quantity corresponding to the target process parameter, and determine a target hidden layer and a target output layer;

[0133] A target activation function determination unit, configured to determine a first initial weight and a first initial threshold based on a preset random initialization method, and determine a target activation function;

[0134] A first initial model construction unit, configured to construct a first initial model according to the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the first initial weight, and the first initial threshold;

[0135] A first initial weight optimization unit, configured to optimize the first initial weight and the first initial threshold based on the preset sparrow search algorithm and a preset error threshold to obtain a second initial weight and a second initial threshold;

[0136] A first target error determination unit, configured to use the first initial model to determine a first output result corresponding to the parameter value group based on a preset forward propagation method, and determine a first target error between the first output result and the target splicing success rate of the cigarette paper tape corresponding to the parameter value group based on a preset backpropagation method, so as to adjust the second initial weight and the second initial threshold based on the first target error to obtain corresponding first target weights and first target thresholds;

[0137] A first model construction unit, configured to determine the first target cigarette paper tape splicing prediction model based on the first initial model, the first target weight, and the first target threshold.

[0138] In some specific embodiments, the second model construction module 14 may include:

[0139] An exploration phase optimization unit, configured to initialize a sparrow population based on a preset chaotic mapping, and determine a target optimization dimension of each sparrow in the sparrow population based on a preset Levy flight strategy and a preset reverse learning mechanism in the exploration phase of the preset sparrow search algorithm;

[0140] A development phase optimization unit, configured to update the position information of each sparrow according to a preset grey wolf optimization algorithm, a preset differential evolution crossover mutation algorithm, and the target optimization dimension in the development phase of the preset sparrow search algorithm;

[0141] A convergence stage optimization unit, which is used to determine target position information based on a preset gradient descent algorithm, a preset population dynamic reduction strategy, and the position information of each sparrow during the convergence stage of the preset sparrow search algorithm, so as to determine the fitness value of the sparrow based on the target position information, and determine the target solution corresponding to the preset sparrow search algorithm based on the fitness value and the preset error threshold.

[0142] In some specific embodiments, the second model construction module 14 may include:

[0143] A second initial model construction unit, which is used to determine a third initial weight and a third initial threshold based on the preset random initialization method, and construct a second initial model based on the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the third initial weight, and the third initial threshold;

[0144] A third initial weight optimization unit, which is used to optimize the third initial weight and the third initial threshold based on the optimized preset sparrow search algorithm and the preset error threshold to obtain a fourth initial weight and a fourth initial threshold;

[0145] A second target error determination unit, which is used to determine a second output result corresponding to the parameter value group based on the preset forward propagation method by using the second initial model, and determine a second target error between the second output result and the target cigarette paper tape splicing success rate corresponding to the parameter value group based on the preset backpropagation method, so as to adjust the fourth initial weight and the fourth initial threshold based on the second target error to obtain corresponding second target weights and second target thresholds;

[0146] A second model construction unit, which is used to determine the second target cigarette paper tape splicing prediction model based on the second initial model, the second target weight, and the second target threshold.

[0147] In some specific embodiments, the cigarette paper tape splicing prediction module 15 may include:

[0148] A to-be-verified parameter value group determination unit, which is used to determine a to-be-verified parameter value group from the parameter value groups in the orthogonal experiment based on a preset quantity condition, and use the first target cigarette paper tape splicing prediction model to determine a first cigarette paper tape splicing prediction result corresponding to the to-be-verified parameter value group, and use the second target cigarette paper tape splicing prediction model to determine a second cigarette paper tape splicing prediction result corresponding to the to-be-verified parameter value group;

[0149] A prediction error determination unit is configured to determine the target cigarette paper tape splicing success rate corresponding to the group of parameter values to be verified, determine a first prediction error between the first cigarette paper tape splicing prediction result and the target cigarette paper tape splicing success rate, and determine a second prediction error between the second cigarette paper tape splicing prediction result and the target cigarette paper tape splicing success rate;

[0150] A model evaluation unit is configured to evaluate the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model based on the first prediction error and the second prediction error.

[0151] Furthermore, an embodiment of the present application also discloses an electronic device. Figure 12 FIG. 20 is a structural diagram of an electronic device 20 shown according to an exemplary embodiment. The content in the figure should not be regarded as any limitation on the scope of use of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the cigarette paper tape splicing prediction method disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0152] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows is any communication protocol applicable to the technical solution of the present application, and no specific limitation is imposed thereon here; the input / output interface 25 is used to obtain external input data or output data to the outside, and its specific interface type can be selected according to specific application needs, and no specific limitation is made here.

[0153] In addition, as a carrier for resource storage, the memory 22 may be a read-only memory, a random access memory, a magnetic disk, or an optical disc, etc. The resources stored thereon may include an operating system 221, a computer program 222, etc., and the storage method may be temporary storage or permanent storage.

[0154] Among them, the operating system 221 is used to manage and control each hardware device and the computer program 222 on the electronic device 20, and it may be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program that can be used to complete the cigarette paper tape splicing prediction method executed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program that can be used to complete other specific tasks.

[0155] Further, the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the cigarette paper splicing prediction method disclosed above. For the specific steps of this method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be repeated here.

[0156] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and reference can be made to the description of the method part for related parts.

[0157] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0158] The steps of the method or algorithm described in combination with the embodiments disclosed herein can be directly implemented by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, register, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0159] Finally, it should also be noted that in this article, 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 variation thereof is intended to cover a non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or 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 element.

[0160] The above has introduced the technical solution provided by this application in detail. Specific examples are used in this article to elaborate on the principle and implementation manner of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A prediction method for splicing cigarette paper tapes, characterized in that, Including: Determine the target process parameters of the cigarette paper splicing process, and determine the parameter values corresponding to each of the target process parameters; Determine a parameter value group based on the parameter values, and use a target tobacco making and tipping machine to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group; Construct a first target cigarette paper splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; Optimize several algorithm stages in the preset sparrow search algorithm according to a preset phased collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper splicing success rate corresponding to the parameter value group; Predict the cigarette paper splicing success rate using the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model.

2. The cigarette paper tape splicing prediction method according to claim 1, characterized in that The determining a parameter value group based on the parameter values, and using a target tobacco making and tipping machine to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group includes: Use a preset statistical analysis software to determine the parameter value group based on a preset orthogonal table, the target process parameters, and the parameter values corresponding to the target process parameters, and generate a corresponding target orthogonal experiment plan table based on the parameter value group; wherein, the target process parameters include a target gasket thickness, a target old disk speed difference, a target old disk deceleration angle, a target pre-tightening speed difference, and a target pre-tightening overspeed angle, the pre-tightening speed difference is the difference between the running speeds of the pre-tightening roller and the new disk in the target tobacco making and tipping machine, the pre-tightening overspeed angle is the angle difference between the pre-tightening roller when it overspeed and the cigarette paper splicing device in the target tobacco making and tipping machine, and the new disk and the old disk in the target tobacco making and tipping machine are the reels for storing cigarette paper; Use the target tobacco making and tipping machine to conduct an orthogonal experiment based on each parameter value group in the target orthogonal experiment plan table to obtain the target cigarette paper splicing success rate corresponding to each parameter value group.

3. The cigarette paper tape splicing prediction method according to claim 1, characterized in that After the using the target tobacco making and tipping machine to conduct an orthogonal experiment based on the parameter value group to obtain the target cigarette paper splicing success rate corresponding to each parameter value group, it further includes: Screen out the parameter value groups with target parameter values from each parameter value group to obtain each target parameter value group corresponding to the target parameter value; the target parameter value is any one of the parameter values; Sum the target cigarette paper splicing success rates corresponding to each target parameter value group corresponding to the target parameter value to obtain a target sum value corresponding to the target parameter value, and determine the target test times corresponding to the target parameter value based on the number of groups of the target parameter value groups corresponding to the target parameter value; Determine the average success rate of cigarette paper splicing corresponding to the target parameter value based on the target number of trials corresponding to the target parameter value and the target sum value corresponding to the target parameter value; Determine the maximum and minimum values among the average success rates of cigarette paper splicing corresponding to each of the target parameter values, and determine the target range of the target process parameters based on the maximum and minimum values, so as to determine the first target prediction model for cigarette paper splicing and the second target prediction model for cigarette paper splicing based on the target range.

4. The cigarette paper tape splicing prediction method according to claim 1, characterized in that, The construction of the first target prediction model for cigarette paper splicing based on the preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the success rate of cigarette paper splicing corresponding to the parameter value group includes: Determine the target input layer based on the target quantity corresponding to the target process parameter, and determine the target hidden layer and the target output layer; Determine the first initial weight and the first initial threshold based on the preset random initialization method, and determine the target activation function; Construct a first initial model according to the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the first initial weight, and the first initial threshold; Optimize the first initial weight and the first initial threshold based on the preset sparrow search algorithm and the preset error threshold to obtain the second initial weight and the second initial threshold; Use the first initial model to determine the first output result corresponding to the parameter value group based on the preset forward propagation method, and determine the first target error between the first output result and the success rate of cigarette paper splicing corresponding to the parameter value group based on the preset backpropagation method, so as to adjust the second initial weight and the second initial threshold based on the first target error to obtain the corresponding first target weight and the first target threshold; Determine the first target prediction model for cigarette paper splicing based on the first initial model, the first target weight, and the first target threshold.

5. The cigarette paper tape splicing prediction method according to claim 4, characterized in that, The optimization of several algorithm stages in the preset sparrow search algorithm according to the preset phased collaborative optimization mechanism to obtain the optimized preset sparrow search algorithm includes: Initialize the sparrow population based on the preset chaotic mapping, and determine the target optimization dimension of each sparrow in the sparrow population based on the preset Levy flight strategy and the preset reverse learning mechanism in the exploration stage of the preset sparrow search algorithm; Update the position information of each sparrow according to the preset grey wolf optimization algorithm, the preset differential evolution crossover mutation algorithm, and the target optimization dimension in the development stage of the preset sparrow search algorithm; Determine the target position information based on the preset gradient descent algorithm, the preset population dynamic reduction strategy, and the position information of each sparrow in the convergence stage of the preset sparrow search algorithm, so as to determine the fitness value of the sparrow based on the target position information, and determine the target solution corresponding to the preset sparrow search algorithm based on the fitness value and the preset error threshold.

6. The cigarette paper tape splicing prediction method according to claim 5, characterized in that, Constructing a second target cigarette paper splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the corresponding target cigarette paper splicing success rate of the parameter value group, including: Determining a third initial weight and a third initial threshold based on the preset random initialization method, and constructing a second initial model based on the preset backpropagation neural network model, the target input layer, the target hidden layer, the target output layer, the target activation function, the third initial weight, and the third initial threshold; Optimizing the third initial weight and the third initial threshold based on the optimized preset sparrow search algorithm and the preset error threshold to obtain a fourth initial weight and a fourth initial threshold; Using the second initial model to determine a second output result corresponding to the parameter value group based on the preset forward propagation method, and determining a second target error between the second output result and the corresponding target cigarette paper splicing success rate of the parameter value group based on the preset backpropagation method, so as to adjust the fourth initial weight and the fourth initial threshold based on the second target error to obtain corresponding second target weights and second target thresholds; Determining the second target cigarette paper splicing prediction model based on the second initial model, the second target weight, and the second target threshold.

7. The cigarette paper tape splicing prediction method according to any one of claims 1 to 6, characterized in that, Predicting the cigarette paper splicing success rate using the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model, including: Determining a parameter value group to be verified from the parameter value groups in the orthogonal experiment based on a preset quantity condition, and using the first target cigarette paper splicing prediction model to determine a first cigarette paper splicing prediction result corresponding to the parameter value group to be verified, and using the second target cigarette paper splicing prediction model to determine a second cigarette paper splicing prediction result corresponding to the parameter value group to be verified; Determining the corresponding target cigarette paper splicing success rate of the parameter value group to be verified, determining a first prediction error between the first cigarette paper splicing prediction result and the target cigarette paper splicing success rate, and determining a second prediction error between the second cigarette paper splicing prediction result and the target cigarette paper splicing success rate; Evaluating the first target cigarette paper splicing prediction model and the second target cigarette paper splicing prediction model based on the first prediction error and the second prediction error.

8. A cigarette paper tape splicing prediction device, characterized in that, Including: A parameter value determination module for determining target process parameters of the cigarette paper splicing process and determining parameter values corresponding to each of the target process parameters; A success rate determination module for determining a parameter value group based on the parameter values, and using a target tobacco tipping machine to perform an orthogonal experiment based on the parameter value group to obtain a target cigarette paper splicing success rate corresponding to each of the parameter value groups; A first model construction module for constructing a first target cigarette paper splicing prediction model based on a preset sparrow search algorithm, a preset backpropagation neural network model, the parameter value group, and the corresponding target cigarette paper splicing success rate of the parameter value group; A second model construction module, configured to optimize several algorithm stages in the preset sparrow search algorithm according to a preset phased collaborative optimization mechanism, obtain the optimized preset sparrow search algorithm, and construct a second target cigarette paper tape splicing prediction model based on the optimized preset sparrow search algorithm, the preset backpropagation neural network model, the parameter value group, and the target cigarette paper tape splicing success rate corresponding to the parameter value group; A cigarette paper tape splicing prediction module, configured to predict the success rate of cigarette paper tape splicing by using the first target cigarette paper tape splicing prediction model and the second target cigarette paper tape splicing prediction model.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory; wherein, the memory is used to store a computer program, and the computer program is loaded and executed by the processor to implement the cigarette paper tape splicing prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that For storing a computer program, which when executed by a processor implements the cigarette paper tape splicing prediction method according to any one of claims 1 to 7.

Images