Control method and apparatus for crushing device, device, medium, and product

Abstract

The present application discloses a control method and apparatus for a crushing device, a device, a medium, and a product. The method comprises: when a crushing device crushes a battery raw material for the N-th time, adjusting a model parameter of a crushing process parameter determination model on the basis of a first reward value, the first reward value being determined on the basis of first particle size data and preset standard particle size data, the first particle size data being particle size data of the battery raw material crushed for the (N-1)-th time by the crushing device, and N being an integer greater than 1; inputting first raw material data of a first battery raw material into the parameter-adjusted crushing process parameter determination model to obtain a crushing process parameter of the first battery raw material; and controlling the crushing device to crush the first battery raw material according to the crushing process parameter of the first battery raw material. In this way, self-feedback adjustment of crushing process parameters and automatic control of crushing devices are realized.

Description

Control method, device, equipment, medium and product of crushing equipment

[0001] Cross-reference to related applications

[0002] This application claims priority to Chinese Patent Application No. 202410826187.4, filed on June 25, 2024, entitled "Control method, device, equipment, medium and product of crushing equipment", the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0003] The present application relates to the technical field of batteries, in particular to a control method, device, equipment, medium and product of a crushing equipment. BACKGROUND

[0004] The semi-finished product of battery raw materials after a high-temperature sintering process needs to be crushed and classified to meet product standards. Different battery raw materials have different sintering temperatures. Some battery raw materials have more serious clumping due to high sintering temperatures, and different crushing process parameters need to be used to control the crushing equipment for crushing.

[0005] Therefore, there is a need for a control method of a crushing equipment to control the crushing of battery raw materials by the crushing equipment.

[0006] SUMMARY

[0007] The present application provides a control method, device, equipment, medium and product of a crushing equipment, which can realize self-feedback adjustment of crushing process parameters and automatic control of the crushing equipment.

[0008] In a first aspect, the present application provides a control method of a crushing equipment, comprising: in the case of Nth crushing of battery raw materials by the crushing equipment, adjusting model parameters of a crushing process parameter determination model based on a first reward value, the first reward value being determined based on first particle size data and preset standard particle size data, the first particle size data being particle size data of battery raw materials crushed by the crushing equipment in the (N-1)th time, N being an integer greater than 1; inputting first raw material data of the first battery raw materials to the crushing process parameter determination model with adjusted parameters to obtain crushing process parameters of the first battery raw materials; and controlling the crushing equipment to crush the first battery raw materials according to the crushing process parameters of the first battery raw materials.

[0009] Thus, in the case of N-th crushing of the battery raw material by the crushing device, the model parameters of the crushing process parameter determination model are adjusted based on a first reward value determined based on first particle size data and preset standard particle size data, the first particle size data being particle size data of the battery raw material crushed by the crushing device for the (N-1)-th time, the first raw material data of the first battery raw material is input into the crushing process parameter determination model with adjusted parameters to obtain the crushing process parameters of the first battery raw material, and then the crushing device is controlled to crush the first battery raw material according to the crushing process parameters of the first battery raw material. In this way, self-feedback adjustment of the crushing process parameters and automatic control of the crushing device are realized.

[0010] In some embodiments, the crushing process parameter determination model comprises a convolutional neural network and a deep neural network; and the inputting of the first raw material data of the first battery raw material into the crushing process parameter determination model with adjusted parameters to obtain the crushing process parameters of the first battery raw material comprises: inputting the first raw material data of the first battery raw material into the convolutional neural network, performing raw material feature extraction on the first raw material data by the convolutional neural network to obtain raw material features; and inputting the raw material features into the deep neural network to predict the crushing process parameters of the first battery raw material based on the raw material features by the deep neural network.

[0011] In this way, the raw material features are extracted from the first raw material data by the convolutional neural network, and the crushing process parameters of the first battery raw material are predicted based on the raw material features by the deep neural network, so that the crushing process parameters can be more accurately determined.

[0012] In some embodiments, the convolutional neural network comprises an input layer, a convolutional layer, a pooling layer, a fully connected layer, and a normalized exponential function layer; and the raw material features are extracted from the first raw material data by the convolutional neural network to obtain the raw material features, comprising: performing normalization processing on the first raw material data by the input layer to obtain target data; performing raw material feature extraction on the target data by the convolutional layer to obtain first features; performing pooling processing on the first features by the pooling layer to obtain second features; integrating the second features by the fully connected layer to obtain third features; and performing normalization processing on the third features by the normalized exponential function layer to obtain the raw material features.

[0013] In this way, the raw material features can be more accurately extracted through the above process.

[0014] In some embodiments, before the model parameters of the crushing process parameter determination model are adjusted based on the first reward value in the process of crushing the battery raw material for the Nth time by the crushing device, the method further comprises: obtaining a plurality of training samples, the training samples comprising raw material data samples of battery raw material samples and standard crushing process parameters corresponding to the raw material data samples; for each training sample in the plurality of training samples, the following steps are performed respectively: inputting the raw material data sample in the training sample into the preset crushing process parameter determination model to obtain predicted crushing process parameters of the battery raw material sample; determining a loss function value based on the predicted crushing process parameters and the standard crushing process parameters in the training sample; under the condition that a training stop condition is not reached, adjusting the model parameters of the crushing process parameter determination model based on the loss function value, and training the crushing process parameter determination model with the adjusted parameters based on the training sample until the training stop condition is reached, to obtain a trained crushing process parameter determination model.

[0015] In this way, through the above training process, a crushing process parameter determination model with higher prediction accuracy can be obtained.

[0016] In some embodiments, adjusting the model parameters of the crushing process parameter determination model based on the loss function value comprises: determining the gradient of the model parameters of the crushing process parameter determination model based on the loss function value; and adjusting the model parameters of the crushing process parameter determination model based on the gradient.

[0017] In this way, through the above process, the model parameters of the crushing process parameter determination model can be adjusted more accurately.

[0018] In some embodiments, the first raw material data comprises at least one of raw material coarse crushing parameters, raw material sintering parameters, and raw material state parameters.

[0019] In this way, based on more comprehensive raw material data of the battery raw material, the crushing process parameters of the battery raw material can be determined more accurately.

[0020] In some embodiments, the crushing process parameters comprise at least one of a feeding frequency, a crushing gas pressure, a classification wheel frequency, a classification wheel sealing gas pressure, an air induction frequency, and a mill bottom weight.

[0021] In this way, through more comprehensive crushing process parameters, the crushing device can be controlled more accurately.

[0022] In a second aspect, the application provides a control device of a crushing device, comprising: an adjustment module configured to adjust model parameters of a crushing process parameter determination model based on a first reward value in a case that the crushing device crushes battery raw materials for the Nth time, the first reward value being determined based on first granularity data and preset standard granularity data, the first granularity data being granularity data of battery raw materials crushed for the (N-1)th time by the crushing device, N being an integer greater than 1; a determination module configured to input first raw material data of first battery raw materials into the crushing process parameter determination model with the adjusted model parameters to obtain crushing process parameters of the first battery raw materials; and a control module configured to control the crushing device to crush the first battery raw materials according to the crushing process parameters of the first battery raw materials.

[0023] In this way, in a case that the crushing device crushes battery raw materials for the Nth time, the model parameters of the crushing process parameter determination model can be adjusted based on the first reward value, the first reward value being determined based on the first granularity data and the preset standard granularity data, the first granularity data being granularity data of battery raw materials crushed for the (N-1)th time by the crushing device, the first raw material data of the first battery raw materials being input into the crushing process parameter determination model with the adjusted model parameters to obtain the crushing process parameters of the first battery raw materials, and then the crushing device is controlled to crush the first battery raw materials according to the crushing process parameters of the first battery raw materials. In this way, self-feedback adjustment of the crushing process parameters and automatic control of the crushing device are realized.

[0024] In a third aspect, the application provides an electronic device, the device comprising: a processor and a memory storing computer program instructions;

[0025] The processor implements the control method of the crushing device as shown in any one of the embodiments of the first aspect when executing the computer program instructions.

[0026] In a fourth aspect, the application provides a computer storage medium, the computer storage medium storing computer program instructions, the computer program instructions being executed by a processor to implement the control method of the crushing device as shown in any one of the embodiments of the first aspect.

[0027] In a fifth aspect, the application provides a computer program product, instructions in the computer program product being executed by a processor of an electronic device to cause the electronic device to perform the control method of the crushing device as shown in any one of the embodiments of the first aspect.

[0028] The above description is only a summary of the technical solutions of the application. In order to more clearly understand the technical means of the application, the application can be implemented in accordance with the content of the description, and in order to make the above and other purposes, characteristics and advantages of the application more obvious and easy to understand, the following specific embodiments of the application are described. BRIEF DESCRIPTION OF DRAWINGS

[0029] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments with reference made to the accompanying drawings. The drawings provided herein are for illustrative purposes only and, therefore, should not be considered to be limiting in any way. Like reference characters in the drawings are denoted by like reference characters throughout the various figures. The drawings contain only preferred embodiments of the application and do not limit the application to the disclosed embodiments.

[0030] Fig. 1 is a flow diagram of a control method of a pulverizing device according to some embodiments of the present application;

[0031] Fig. 2 is a structural diagram of a pulverizing process parameter determination model according to some embodiments of the present application;

[0032] Fig. 3 is a flow diagram of a control method of a pulverizing device according to some embodiments of the present application;

[0033] Fig. 4 is a flow diagram of a training process of a pulverizing process parameter determination model according to some embodiments of the present application;

[0034] Fig. 5 is a flow diagram of a control method of a pulverizing device according to some embodiments of the present application;

[0035] Fig. 6 is a structural diagram of a control device of a pulverizing device according to some embodiments of the present application;

[0036] Fig. 7 is a structural diagram of an electronic device according to some embodiments of the present application. DETAILED DESCRIPTION

[0037] The embodiments of the technical solutions of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and therefore only serve as examples, and cannot limit the protection scope of the present application.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application; the terms "comprising" and "having," and any variations thereof, as used in the specification and claims herein, are intended to cover not only the recited elements but also any additional elements.

[0039] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly specified.

[0040] Reference to an "embodiment" in this document means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. As will be apparent to those of ordinary skill in the art, embodiments described herein can be combined with other embodiments.

[0041] In the description of the embodiments of the application, the term "and / or" is merely an association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0042] In the description of the embodiments of the application, the term "a plurality of" refers to more than two (including two), and similarly, "a plurality of groups" refers to more than two groups (including two groups), and "a plurality of pieces" refers to more than two pieces (including two pieces).

[0043] As in the background art, the semi-finished product of the lithium ion battery positive electrode material after the high-temperature sintering process needs to be crushed and classified to meet the product standard. Different positive electrode materials have different sintering temperatures. Some materials have more serious clumping due to high sintering temperature, and need to be controlled using different crushing process parameters. Usually, there are many control parameters of the crushing equipment, and real-time adjustment is mostly relied on manual experience, which has great management difficulty and high labor cost, and the consistency of the particle size after crushing is low. Manual adjustment of the crushing process parameters by personnel is prone to operation errors, resulting in batch product defects.

[0044] To solve the above technical problems, in the embodiments of the application, the model parameters of the crushing process parameter determination model can be adjusted based on a first reward value in the case of Nth crushing of the battery raw material by the crushing equipment, the first reward value is determined based on first particle size data and preset standard particle size data, the first particle size data is the particle size data of the battery raw material crushed by the crushing equipment for the (N-1)th time, the first raw material data of the first battery raw material is input into the crushing process parameter determination model after the parameter adjustment, to obtain the crushing process parameters of the first battery raw material, and then the crushing equipment is controlled to crush the first battery raw material according to the crushing process parameters of the first battery raw material. In this way, self-feedback adjustment of the crushing process parameters and automatic control of the crushing equipment are realized.

[0045] The control method of the crushing equipment provided by the embodiments of the application will be described in detail below in combination with FIG. 1.

[0046] FIG. 1 shows a flowchart of a control method of a crushing device according to an embodiment of the present application. It should be noted that the control method of the crushing device can be applied to a control device of the crushing device.

[0047] As shown in FIG. 1, the control method of the crushing device can include the following steps:

[0048] S110, in the case that the crushing device crushes the battery raw material for the Nth time, adjusting model parameters of a crushing process parameter determination model based on a first reward value;

[0049] S120, inputting first raw material data of the first battery raw material to the crushing process parameter determination model with the adjusted parameters to obtain crushing process parameters of the first battery raw material;

[0050] S130, controlling the crushing device to crush the first battery raw material according to the crushing process parameters of the first battery raw material.

[0051] Thus, in the case that the crushing device crushes the battery raw material for the Nth time, the model parameters of the crushing process parameter determination model can be adjusted based on the first reward value, the first reward value is determined based on the first particle size data and the preset standard particle size data, the first particle size data is the particle size data of the battery raw material crushed by the crushing device for the (N-1)th time, the first raw material data of the first battery raw material is input to the crushing process parameter determination model with the adjusted parameters to obtain the crushing process parameters of the first battery raw material, and then the crushing device is controlled to crush the first battery raw material according to the crushing process parameters of the first battery raw material. In this way, self-feedback adjustment of the crushing process parameters and automatic control of the crushing device are realized.

[0052] Regarding S110, the battery raw material can be a semi-finished product of lithium ion battery positive electrode material after a high-temperature sintering process.

[0053] The crushing process parameter determination model can be a deep reinforcement learning model (Deep Q-Learning, DQN). The crushing process parameter determination model can be used to determine the crushing process parameters of the battery raw material.

[0054] The model parameters of the crushing process parameter determination model can be adjusted based on the first reward value. Specifically, a corresponding relationship between the first reward value and the model parameters can be set in advance, so that the model parameters can be adjusted based on the corresponding relationship and the first reward value.

[0055] The first reward value can be determined based on the first particle size data and the preset standard particle size data. The first particle size data can be the particle size data of the battery raw material crushed by the crushing device for the (N-1)th time. N can be an integer greater than 1.

[0056] The first granularity data can be obtained by detecting the granularity of the battery raw material crushed for the N-1th time by the crushing device. The preset standard granularity data can be set according to actual needs.

[0057] S120 is related to determining the crushing process parameters of the first battery raw material based on the first raw material data of the first battery raw material through the adjusted parameters.

[0058] In some embodiments, in order to more accurately determine the crushing process parameters of the first battery raw material, the first raw material data can include at least one of the raw material coarse crushing parameters, the raw material sintering parameters and the raw material state parameters.

[0059] In addition, the first raw material data can also include parameters of other processes that the first battery raw material has undergone before the crushing process.

[0060] In this way, based on more comprehensive raw material data of the battery raw material, the crushing process parameters of the battery raw material can be more accurately determined.

[0061] In some embodiments, in order to more accurately control the crushing device, the crushing process parameters can include at least one of the feeding frequency, the crushing gas pressure, the classification wheel frequency, the classification wheel sealing gas pressure, the air induction frequency and the mill bottom weight.

[0062] In addition, according to actual needs, the crushing process parameters can also include other parameters.

[0063] In this way, through more comprehensive crushing process parameters, the crushing device can be more accurately controlled.

[0064] In some embodiments, in order to more accurately determine the crushing process parameters, the crushing process parameter determination model can include a convolutional neural network and a deep neural network.

[0065] Based on this, S120 can include:

[0066] The first raw material data of the first battery raw material is input into the convolutional neural network, and the raw material features are obtained by performing raw material feature extraction on the first raw material data through the convolutional neural network.

[0067] The raw material features are input into the deep neural network, and the crushing process parameters of the first battery raw material are predicted based on the raw material features through the deep neural network.

[0068] Here, the convolutional neural network can be used for feature extraction on the first raw material data. The deep neural network can be used to predict the crushing process parameters of the first battery raw material.

[0069] Exemplarily, as shown in FIG. 2, the crushing process parameter determination model can include a convolutional neural network 210 and a deep neural network 220. The convolutional neural network 210 can output data iteratively, and the deep neural network 220 can enhance operation accuracy.

[0070] In this way, by performing material feature extraction on the first material data through the convolutional neural network, and predicting the crushing process parameters of the first battery material based on the material features through the deep neural network, the crushing process parameters can be determined more accurately.

[0071] In some embodiments, in order to more accurately extract material features, the convolutional neural network can include an input layer, a convolutional layer, a pooling layer, a fully connected layer, and a normalized exponential function layer.

[0072] Based on this, the above-mentioned material feature extraction on the first material data through the convolutional neural network can include:

[0073] The first material data is normalized through the input layer to obtain target data;

[0074] The target data is subjected to material feature extraction through the convolutional layer to obtain first features;

[0075] The first features are subjected to pooling processing through the pooling layer to obtain second features;

[0076] The second features are integrated through the fully connected layer to obtain third features;

[0077] The third features are normalized through the normalized exponential function layer to obtain material features.

[0078] Here, the first material data can be normalized through the input layer to standardize its input characteristics, so that data of different dimensions have the same distribution, making statistical analysis more effective. The convolutional layer performs cross-correlation operation to extract feature values of different dimensions of the target data, and uses a linear rectifier function to express its features. The pooling layer selects and filters the first features, compresses the amount of data and parameters, reduces overfitting, and forms a feature vector matrix. The fully connected layer integrates the second features. The soft max (normalized exponential function) obtains the final output.

[0079] Exemplarily, as shown in FIG. 2, the convolutional neural network 210 can include an input layer, a convolutional layer 211, a pooling layer 212, a fully connected layer 213, and a normalized exponential function layer 214.

[0080] In this way, the material features can be more accurately extracted through the above process.

[0081] S130, the process of the crushing device crushing the first battery raw material according to the crushing process parameters of the first battery raw material can be the process of the crushing device crushing the battery raw material for the Nth time.

[0082] After determining the crushing process parameters of the first battery raw material based on the crushing process parameter determination model, the crushing device can be directly controlled to crush the first battery raw material according to the crushing process parameters of the first battery raw material, realizing automatic control of the crushing device without manual control.

[0083] Exemplarily, as shown in FIG. 3, the raw material data such as raw material crushing parameters, raw material sintering parameters and raw material state parameters can be input into the crushing process parameter determination model, the crushing process parameter determination model can output the crushing process parameters such as feeding frequency, crushing gas pressure, classification wheel frequency, classification wheel sealing gas pressure, induced draft frequency and mill bottom weight, the crushing process parameters are input into the crushing device, the crushing device can be controlled to crush the battery raw material, the particle size of the crushed battery raw material is detected, and then the reward value is obtained, and the model parameters of the crushing process parameter determination model can be adjusted based on the reward value. Then the above process can be repeated based on the crushing process parameter determination model after the parameters are adjusted to perform the next crushing.

[0084] In the embodiments of the present application, through the effective connection of the crushing process parameter determination model and the crushing device, intelligent control of the crushing device is realized through deep reinforcement learning, automatic control and adjustment of the crushing process parameters are realized, thereby replacing manual adjustment, reducing production risk, improving manufacturing efficiency, realizing unmanned operation on site, greatly reducing labor cost, and through automatic calculation and prediction of the optimal strategy between the target parameters by big data, the accuracy of optimization is improved to improve product consistency.

[0085] In some embodiments, in order to obtain a crushing process parameter determination model with higher prediction accuracy, before S110, the method can further include:

[0086] Obtaining a plurality of training samples, the training samples can include raw material data samples of battery raw material samples and standard crushing process parameters corresponding to the raw material data samples;

[0087] For each training sample in the plurality of training samples, the following steps are performed respectively:

[0088] Inputting the raw material data sample in the training sample into the preset crushing process parameter determination model to obtain the predicted crushing process parameters of the battery raw material sample;

[0089] Based on the predicted crushing process parameters and the standard crushing process parameters in the training sample, a loss function value is determined;

[0090] The model parameter of the crushing process parameter determination model is adjusted based on the loss function value in a case where the training stop condition is not reached, and the crushing process parameter determination model after the parameter adjustment is trained based on the training sample until the training stop condition is reached, so as to obtain the trained crushing process parameter determination model.

[0091] Here, the standard crushing process parameter can be artificially set or obtained through other algorithms. The training stop condition can be that the number of training reaches a preset number, or the loss function value reaches a preset threshold.

[0092] Specifically, the raw material data sample and the standard crushing process parameter corresponding to the raw material data sample can be collected, a plurality of training samples can be constructed, and a preset crushing process parameter determination model can be constructed, and then the deep reinforcement learning crushing process parameter determination model can be trained based on the plurality of training samples.

[0093] In this way, through the above training process, the crushing process parameter determination model with higher prediction accuracy can be obtained.

[0094] In some embodiments, in order to more accurately adjust the model parameter of the crushing process parameter determination model, the above adjustment of the model parameter of the crushing process parameter determination model based on the loss function value can include:

[0095] determining the gradient of the model parameter of the crushing process parameter determination model based on the loss function value;

[0096] adjusting the model parameter of the crushing process parameter determination model based on the gradient.

[0097] The gradient can be used to determine the adjustment amplitude of the model parameter.

[0098] In this way, through the above process, the model parameter of the crushing process parameter determination model can be more accurately adjusted.

[0099] For example, as shown in FIG. 4, during the training process, the experience replay mechanism can be used for the data output by the convolutional neural network. In addition to using the deep convolutional network to approximate the current value function, another network is used to generate the target Q value. The reward value, i.e. the online test granularity data and the error term, is reduced to a limited interval, so that the Q value and the gradient value are within a reasonable range, and the stability of the algorithm is improved.

[0100] s is the state, a is the action, r is the reward, γ is the discount factor for punishing uncertainty in future rewards, Q(s, a) is the action-value function, and Q is the crushing process parameter.

[0101] Specifically, Q(s, a|θ i) can represent the output of the current value network, which can be used to evaluate the value function corresponding to the current state-action pair; ) can represent the output of the target value network, which is generally used to approximate the optimization target of the value function, i.e., the target Q value. The parameters θ of the current value network can be updated in real time, and after every M rounds of iterations, the parameters of the current value network can be copied to the target value network. The network parameters are updated by minimizing the mean square error between the current Q value and the target Q value. The error function can be: i s,a,r,s′ i i 2

[0102] where i is the iteration number, L(θ i ) is the error function, E s，a，r，s′ is the state, action and reward corresponding to a certain time step in a trajectory, r is the reward, Y i is the initial state of the i-th iteration, s' is the state of the next iteration, is the target network parameter of the i-th iteration, a is the action, and θ i is the network parameter of the DQN.

[0103] Taking the partial derivative of the parameter θ, the following gradient is obtained:

[0104] where is the derivative of the network Q value, is the derivative of the network parameter of the DQN.

[0105] After introducing the target value network, the target Q value remains unchanged for a period of time, which reduces the correlation between the current Q value and the target Q value to some extent and improves the stability of the algorithm.

[0106] The algorithm description of the DQN can be:

[0107] where is the derivative of the parameter θ, L i (θ i ) is the mean square error function, E s，a～ρ(·)；s′～ε is the greedy policy, ε is the discount factor that determines the field of view of the agent, and θ i-1 is the parameter of the previous iteration.

[0108] In some embodiments, after the training of the crushing process parameter determination model is completed, the crushing process parameter determination model can also be tested.

[0109] ​​​​​In some embodiments, the comminution process parameter determination model can also be iteratively upgraded periodically after it is put into use.

[0110] Exemplarily, as shown in FIG. 5, the control method of the comminution device can include the following steps:

[0111] S501, determining a key parameter in a key process.

[0112] The key process can be a comminution process, and the key parameter can include a feeding frequency, a comminution gas pressure, a classification wheel frequency, a classification wheel sealing gas pressure, an induced draft frequency, and a mill base weight.

[0113] S502, obtaining raw data of structured data.

[0114] The structured data can include raw material state data, raw material sintering parameters, and raw material coarse crushing parameters.

[0115] S503, establishing an association between the key device parameters and the raw data to obtain initial data sets A1-A N .

[0116] The feeding frequency, the comminution gas pressure, the classification wheel frequency, the classification wheel sealing gas pressure, the induced draft frequency, and the mill base weight are correlated with the raw material state data, the raw material sintering parameters, and the raw material coarse crushing parameters to obtain initial data sets A1-A N .

[0117] S504, constructing a training set and a test set.

[0118] The raw data is divided into a training set and a test set in a ratio of 70% and 30%, and the obtained feature data (i.e., comminution particle size) corresponding to the training set is used to train the comminution process parameter determination model.

[0119] S505, using a deep reinforcement learning agent to establish a comminution process parameter determination model.

[0120] The set parameter configuration is obtained, and the raw data is input into the deep reinforcement learning agent for learning to obtain feature data and a hyperparameter set.

[0121] S506, training the deep reinforcement learning comminution process parameter determination model.

[0122] The feature data and the hyperparameters of this round are used to train and test the crushing process parameter determination model, and the predicted parameters are output; the predicted parameters of the feature data and the hyperparameter set are returned to the crushing process parameter determination model for parameter training, and the weight parameters inside the crushing process parameter determination model are adjusted; the above steps are repeatedly iterated, and when the number of iterations reaches the set number of learning times, the iteration ends, and the final crushing process parameter determination model is obtained.

[0123] S507, test the accuracy of the crushing process parameter determination model of the deep reinforcement learning.

[0124] The predicted value of the predicted target parameter is output, the obtained feature data is imported into the trained crushing process parameter determination model for verification with the test set, the predicted value of the predicted target parameter is compared with the true value of the predicted target parameter to obtain the prediction accuracy parameter.

[0125] S508, feature information storage.

[0126] The hyperparameter information and the prediction accuracy parameter information are recorded in the information storage.

[0127] S509, output the prediction result.

[0128] The central control system automatically inputs the particle size online test result, the raw material coarse crushing parameter, the raw material sintering parameter, the raw material state data and other structured original data as the predicted data into the crushing process parameter determination model to obtain the final prediction result of the predicted target parameter.

[0129] S510, intelligently adjust the parameters of the crushing equipment according to the prediction result.

[0130] The prediction result is input into the crushing equipment to realize intelligent adjustment of the parameters of the crushing equipment.

[0131] The crushing process parameter determination model can include a discrete data agent, a continuous data agent and a hyperparameter agent, the discrete data agent and the continuous data agent can adopt a multi-head self-attention network structure with the same topological structure, and the hyperparameter agent can adopt a recurrent neural network. The original data can be split into discrete data and continuous data, the discrete data is input into the discrete data agent to output a discrete feature generation dictionary, the continuous data is input into the continuous data agent to output a continuous feature generation dictionary, the initially set hyperparameters are input into the hyperparameter data agent to output updated hyperparameters, and the updated hyperparameters form a new hyperparameter set; then the discrete feature generation dictionary is combined with the discrete feature generation function to generate discrete features, and the continuous feature generation dictionary is combined with the continuous feature generation function to generate continuous features as feature data, further improving the accuracy of data output.

[0132] In the embodiment of the application, a deep reinforcement learning algorithm is used to perform deep learning on big data, so as to obtain a crushing process parameter determination model with high accuracy.

[0133] In the embodiment of the application, a network structure based on deep reinforcement learning is first established, then a state, action and reward function calculation method of the deep reinforcement learning network are determined, then the deep reinforcement learning network is initialized, and finally the deep reinforcement learning network is trained according to the system state to obtain an optimal strategy for parameter allocation and device system control. Through the organic combination of the deep reinforcement learning method and the device control system, intelligent adjustment and control of the crushing device are realized.

[0134] Based on the same inventive concept, the embodiment of the application also provides a control device of a crushing device. The control device of the crushing device provided by the embodiment of the application will be described in detail below with reference to FIG. 6.

[0135] FIG. 6 shows a structural schematic diagram of a control device of a crushing device according to an embodiment of the application.

[0136] As shown in FIG. 6, the control device of the crushing device can include:

[0137] An adjusting module 601 is configured to adjust model parameters of a crushing process parameter determination model based on a first reward value in the case that the crushing device crushes battery raw materials for the Nth time, the first reward value being determined based on first granularity data and preset standard granularity data, the first granularity data being granularity data of battery raw materials crushed by the crushing device for the (N-1)th time, N being an integer greater than 1.

[0138] A determining module 602 is configured to input first raw material data of the first battery raw materials to the crushing process parameter determination model with adjusted parameters to obtain crushing process parameters of the first battery raw materials.

[0139] A control module 603 is configured to control the crushing device to crush the first battery raw materials according to the crushing process parameters of the first battery raw materials.

[0140] In this way, in the case that the crushing device crushes battery raw materials for the Nth time, the model parameters of the crushing process parameter determination model can be adjusted based on the first reward value, the first reward value being determined based on the first granularity data and the preset standard granularity data, the first granularity data being granularity data of battery raw materials crushed by the crushing device for the (N-1)th time, the first raw material data of the first battery raw materials is input to the crushing process parameter determination model with adjusted parameters to obtain the crushing process parameters of the first battery raw materials, and then the crushing device is controlled to crush the first battery raw materials according to the crushing process parameters of the first battery raw materials. In this way, self-feedback adjustment of the crushing process parameters and automatic control of the crushing device are realized.

[0141] In some embodiments, in order to more accurately determine the crushing process parameters, the crushing process parameter determination model comprises a convolutional neural network and a deep neural network;

[0142] Based on this, the determination module 602 can comprise:

[0143] The extraction submodule is configured to input the first raw material data of the first battery raw material into the convolutional neural network, perform raw material feature extraction on the first raw material data through the convolutional neural network, and obtain the raw material feature.

[0144] The prediction submodule is configured to input the raw material feature into the deep neural network, and predict the crushing process parameters of the first battery raw material based on the raw material feature through the deep neural network.

[0145] In some embodiments, in order to more accurately extract the raw material feature, the convolutional neural network comprises an input layer, a convolutional layer, a pooling layer, a fully connected layer, and a normalized exponential function layer;

[0146] Based on this, the extraction submodule can comprise:

[0147] The first normalization unit is configured to perform normalization processing on the first raw material data through the input layer to obtain target data.

[0148] The extraction unit is configured to perform raw material feature extraction on the target data through the convolutional layer to obtain first features.

[0149] The pooling unit is configured to perform pooling processing on the first features through the pooling layer to obtain second features.

[0150] The integration unit is configured to integrate the second features through the fully connected layer to obtain third features.

[0151] The second normalization unit is configured to perform normalization processing on the third features through the normalized exponential function layer to obtain the raw material feature.

[0152] In some embodiments, in order to obtain a crushing process parameter determination model with higher prediction accuracy, the control device of the crushing equipment can further comprise:

[0153] The acquisition module is configured to, before adjusting the model parameters of the crushing process parameter determination model based on the first reward value during the Nth crushing of the battery raw material by the crushing equipment, acquire a plurality of training samples, wherein the training samples comprise raw material data samples of battery raw material samples and standard crushing process parameters corresponding to the raw material data samples.

[0154] The training module is configured to, for each training sample in the plurality of training samples, perform the following steps:

[0155] The raw material data sample in the training sample is input into a preset crushing process parameter determination model to obtain a predicted crushing process parameter of the battery raw material sample;

[0156] Based on the predicted crushing process parameter and the standard crushing process parameter in the training sample, a loss function value is determined.

[0157] If the training stop condition is not reached, the model parameter of the crushing process parameter determination model is adjusted based on the loss function value, and the adjusted crushing process parameter determination model is trained based on the training sample until the training stop condition is reached, and a trained crushing process parameter determination model is obtained.

[0158] In some embodiments, in order to more accurately adjust the model parameter of the crushing process parameter determination model, the training module can include:

[0159] A determination sub-module is configured to determine the gradient of the model parameter of the crushing process parameter determination model based on the loss function value.

[0160] An adjustment sub-module is configured to adjust the model parameter of the crushing process parameter determination model based on the gradient.

[0161] In some embodiments, in order to more accurately determine the crushing process parameter of the first battery raw material, the first raw material data includes at least one of a raw material coarse crushing parameter, a raw material sintering parameter, and a raw material state parameter.

[0162] In some embodiments, in order to more accurately control the crushing equipment, the crushing process parameter includes at least one of a feeding frequency, a crushing gas pressure, a classification wheel frequency, a classification wheel sealing gas pressure, an air induction frequency, and a mill bottom weight.

[0163] FIG. 7 shows a structural schematic diagram of an electronic device according to an embodiment of the present application.

[0164] As shown in FIG. 7, the electronic device 7 can realize an exemplary hardware architecture structure diagram of the electronic device of the control method and the control device of the crushing equipment according to the embodiments of the present application. The electronic device can refer to the electronic device in the embodiments of the present application.

[0165] The electronic device 7 can include a processor 701 and a memory 702 storing computer program instructions.

[0166] Specifically, the processor 701 can include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or can be configured as one or more integrated circuits that implement the embodiments of the present application.

[0167] The memory 702 can include mass storage for data or instructions. As an example and not by way of limitation, the memory 702 can include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disc (e.g., a compact disc (CD) or a digital versatile disc (DVD)), a solid-state drive (SSD), a USB drive, or a combination of two or more of these. Where appropriate, the memory 702 can include removable or non-removable (or fixed) media, where appropriate. The memory 702 can be internal or external to the integrated gateway disaster recovery appliance. In particular embodiments, the memory 702 is non-volatile, solid-state memory. In particular embodiments, the memory 702 can include read-only memory (ROM), random-access memory (RAM), a magnetic disk storage medium, an optical storage medium, flash memory devices, electrical, optical, or other physically tangible storage device. Accordingly, in general, the memory 702 includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software that, when executed (by one or more processors), is operable to

[0168] The processor 701 implements the control method of the pulverizing device in any of the above embodiments by reading and executing computer program instructions stored in the memory 702.

[0169] In one example, the electronic device can further include a communication interface 703 and a bus 704. As shown in FIG. 7, the processor 701, the memory 702, and the communication interface 703 are connected through the bus 704 and complete communication with each other.

[0170] The communication interface 703 is mainly used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the application.

[0171] Bus 704 includes a hardware, software, or both that couples components of electronic device to each other. As an example and not by way of limitation, bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand (IB) interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association local (VLB) bus, or another suitable bus or a combination of two or more of these. Where appropriate, bus 704 can include one or more buses. Although this application describes and shows a particular bus, this application contemplates any suitable bus or interconnect.

[0172] The electronic device can execute the control method of the pulverizing device in the embodiments of the present application, thereby realizing the control method and device of the pulverizing device described in combination with FIGS. 1 to 6.

[0173] In addition, in combination with the control method of the pulverizing device in the above embodiments, the embodiments of the present application can provide a computer storage medium to realize. The computer storage medium has computer program instructions stored thereon; the computer program instructions are executed by a processor to realize any one of the control methods of the pulverizing device in the above embodiments.

[0174] It needs to be clear that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0175] The functions of the elements of the described structural diagrams shown in the above-described embodiments can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, functional cards, and the like. When implemented in software, the elements of the present application are program or code segments used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link, by way of data signals carried in a carrier wave. The "machine-readable medium" can include any medium that can store or transfer information. Examples of the machine-readable medium include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, and the like. The code segments can be downloaded via a computer network such as the Internet, an intranet, and the like.

[0176] It is also necessary to note that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in an order different from that mentioned in the embodiments, or several steps can be performed simultaneously.

[0177] The above-described aspects of the present application are described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application. It should be understood that each block of the flowchart and / or block diagram, and combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which are executed via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / acts specified in the flowchart and / or block diagram block or blocks. The processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special-purpose application processor, or a field programmable logic circuit. It can also be understood that each block of the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can also be implemented by dedicated hardware to perform the specified functions or acts, or can be implemented by a combination of dedicated hardware and computer instructions.

[0178] Although the present application has been described with reference to preferred embodiments, various modifications can be made to it without departing from the scope of the present application, and components thereof can be replaced with equivalents, especially, the technical features mentioned in each embodiment can be combined in any manner, provided that there is no structural conflict. The present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A control method for a crushing device, comprising: In the case of the battery raw material being crushed by the crushing equipment for the Nth time, the model parameters of the model are determined by adjusting the crushing process parameters based on the first reward value. The first reward value is determined based on the first particle size data and the preset standard particle size data. The first particle size data is the particle size data of the battery raw material crushed by the crushing equipment for the N-1th time, where N is an integer greater than 1. The first raw material data of the first battery raw material is input into the pulverization process parameter determination model after parameter adjustment to obtain the pulverization process parameters of the first battery raw material. The crushing equipment is controlled to crush the first battery raw material according to the crushing process parameters of the first battery raw material.

2. The method according to claim 1, wherein, The model for determining the pulverization process parameters includes convolutional neural networks and deep neural networks; The step of inputting the first raw material data of the first battery raw material into the parameter-adjusted pulverization process parameter determination model to obtain the pulverization process parameters of the first battery raw material includes: The first raw material data of the first battery raw material is input into the convolutional neural network, and the raw material features are extracted from the first raw material data by the convolutional neural network to obtain the raw material features; The raw material characteristics are input into the deep neural network, and the deep neural network predicts the crushing process parameters of the first battery raw material based on the raw material characteristics.

3. The method according to claim 2, wherein, The convolutional neural network includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and a normalized exponential function layer; The step of extracting raw material features from the first raw material data using the convolutional neural network to obtain raw material features includes: The first raw material data is normalized through the input layer to obtain the target data. The target data is subjected to raw material feature extraction through the convolutional layer to obtain the first feature; The first feature is pooled using the pooling layer to obtain the second feature; The second feature is integrated through the fully connected layer to obtain the third feature; The raw material characteristics are obtained by normalizing the third feature through the normalized exponential function layer.

4. The method according to any one of claims 1-3, before determining the model parameters of the model based on the first reward value during the Nth pulverization of battery raw materials in the pulverizing equipment, the method further includes: Multiple training samples are obtained, including raw material data samples of battery raw materials and standard crushing process parameters corresponding to the raw material data samples; For each of the plurality of training samples, perform the following steps: The raw material data samples in the training samples are input into a preset crushing process parameter determination model to obtain the predicted crushing process parameters of the battery raw material samples. Based on the predicted pulverizing process parameters and the standard pulverizing process parameters in the training samples, the loss is determined. Loss of function value; If the training stopping condition is not met, the model parameters of the crushing process are adjusted based on the loss function value to determine the model parameters, and the model is determined based on the adjusted crushing process parameters trained on the training samples, until the training stopping condition is met, thus obtaining the trained crushing process parameter determination model.

5. The method according to claim 4, wherein, The process of adjusting the pulverizing process parameters based on the loss function value to determine the model parameters includes: The gradient of the model parameters is determined based on the loss function value to determine the crushing process parameters. The model parameters are determined by adjusting the pulverizing process parameters based on the gradient.

6. The method according to any one of claims 1-5, wherein, The first raw material data includes at least one of the following: raw material coarse crushing parameters, raw material sintering parameters, and raw material state parameters.

7. The method according to any one of claims 1-6, wherein, The crushing process parameters include at least one of the following: feeding frequency, crushing air pressure, classifying wheel frequency, classifying wheel sealing air pressure, induced draft frequency, and grinding bottom material weight.

8. A control device for a crushing equipment, the device comprising: An adjustment module is used to adjust the model parameters of the model based on a first reward value when the battery raw materials are crushed by the crushing equipment for the Nth time. The first reward value is determined based on the first particle size data and the preset standard particle size data. The first particle size data is the particle size data of the battery raw materials crushed by the crushing equipment for the N-1th time, where N is an integer greater than 1. The determination module is used to input the first raw material data of the first battery raw material into the pulverization process parameter determination model after parameter adjustment, so as to obtain the pulverization process parameters of the first battery raw material. The control module is used to control the crushing equipment to crush the first battery raw material according to the crushing process parameters of the first battery raw material.

9. An electronic device, the device comprising: Processor and memory storing computer program instructions; When the processor executes the computer program instructions, it implements the control method for the crushing equipment as described in any one of claims 1-7.

10. A computer storage medium storing computer program instructions, wherein the computer program instructions, when executed by a processor, implement the control method of the crushing equipment as described in any one of claims 1-7.

11. A computer program product, wherein instructions in the computer program product, when executed by a processor of an electronic device, cause the electronic device to perform the control method for a crushing device as described in any one of claims 1-7.

Images