Rolling control method and equipment based on pole piece thickness prediction and medium

Abstract

The embodiment of the invention discloses a rolling control method and equipment based on pole piece thickness prediction and a medium. The rolling control method based on pole piece thickness prediction comprises the steps that rolling operation parameters and a data pair set corresponding to the pole piece thickness are obtained; according to the data pair set, determining a pole piece thickness prediction model capable of representing correlation characteristics between the rolling operation parameters and the pole piece thickness; inputting the current rolling operation parameters into the pole piece thickness prediction model to predict the numerical value of the current pole piece thickness; according to the difference between the numerical value of the current pole piece thickness and the target pole piece thickness threshold value, the current operation parameters of rolling are adjusted. According to the technical scheme provided by the embodiment of the invention, the thickness of the pole piece is accurately predicted through the pole piece thickness prediction model, so that the thickness consistency of the pole piece actually produced by rolling is relatively good, and the problem that the first-order optimal rate of the thickness of the pole piece is influenced due to the fact that the thickness of the pole piece manufactured by the existing pole piece manufacturing method is inconsistent is solved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of pole piece manufacturing, and in particular to a rolling control method, device and medium based on pole piece thickness prediction. Background Art

[0002] With the development of power battery technology, people have higher and higher requirements for the pole pieces of power batteries. Existing pole pieces are generally made by rolling. Rolling is the most important process before pole piece forming. Whether in theory or in actual production, there are a lot of items that can be improved. Rolling with a pulling function reduces pole piece wrinkles. Hot rolling reduces pole piece rebound and enhances battery performance. Rolling with bending controls the consistency of pole piece thickness to a greater extent. The pole pieces made by the existing pole piece manufacturing method have the problem of inconsistent thickness, which affects the first-time excellent rate of pole piece thickness. Summary of the invention

[0003] The embodiments of the present invention provide a rolling control method, device and medium based on pole piece thickness prediction to solve the problem that pole pieces made by the existing pole piece manufacturing method have inconsistent thickness, which affects the primary quality rate of the pole piece thickness.

[0004] In order to achieve the above technical problems, the present invention adopts the following technical solutions:

[0005] An embodiment of the present invention provides a rolling control method based on pole piece thickness prediction, comprising:

[0006] Obtaining a data pair set of roller pressing operating parameters and corresponding pole piece thickness;

[0007] Determining, according to the data pair set, a pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness;

[0008] Inputting the current operating parameters of the roller press into the pole piece thickness prediction model to predict the value of the current pole piece thickness;

[0009] According to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold, the current operating parameters of the rolling are adjusted.

[0010] Optionally, the obtaining of the operating parameters of the roller press includes at least one of the following:

[0011] Negative calibration value of roll pressure, positive calibration value, meters, roll pressure speed, roll pressure meter, lower bending cylinder pressure on auxiliary rolling transmission side, lower bending cylinder pressure on main rolling transmission side, upper bending cylinder pressure on main rolling operation side, lower bending cylinder pressure on auxiliary rolling operation side, unwinding tension and winding tension.

[0012] Optionally, determining, based on the data pair set, a pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness includes:

[0013] The data pair set is input into a stacked regression model for learning, and a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operation parameters and the pole piece thickness is determined.

[0014] Optionally, the stacked regression model includes a random forest classifier model, a redundant tree classifier model, an adaptive boosting classifier model, a gradient boosting classifier model and a decision tree regression model;

[0015] Accordingly, the inputting of the current operating parameters of the rolling process into the pole piece thickness prediction model to predict the value of the current pole piece thickness includes:

[0016] Substituting the current operating parameters of the roller press into the random forest classifier model to generate a first predicted thickness value of the pole piece;

[0017] Substituting the current operating parameters of the roller press into the redundant tree classifier model to generate a second predicted thickness value of the pole piece;

[0018] Substituting the current operating parameters of the roller press into the adaptive boosting classifier model to generate a third predicted thickness value of the pole piece;

[0019] Substituting the current operating parameters of the roller press into the gradient boosting classifier model to generate a fourth predicted thickness value of the pole piece;

[0020] Substituting the current operating parameters of the roller press into the decision tree regression model to generate a fifth predicted thickness value of the pole piece;

[0021] The first predicted thickness value, the second predicted thickness value, the third predicted thickness value, the fourth predicted thickness value and the fifth predicted thickness value are merged to determine the value of the current pole piece thickness.

[0022] Optionally, adjusting the current operating parameters of the rolling process according to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold value includes:

[0023] Comparing the value of the current pole piece thickness with a target pole piece thickness threshold;

[0024] When the value of the current pole piece thickness is greater than or less than the target pole piece thickness threshold, the current operating parameters of the rolling process are adjusted.

[0025] Optionally, after adjusting the current operating parameters of the roller press, the method further includes:

[0026] When the value of the current electrode thickness is greater than or less than the target electrode thickness threshold, the primary quality rate of the produced electrode is judged, and an early warning is issued based on the judgment result.

[0027] Optionally, the data pair set for obtaining the rolling operation parameters and the corresponding pole piece thickness includes:

[0028] Use the panel data tool to query the data pair set of the rolling operation parameters and the corresponding pole piece thickness, and generate data pair sets for multiple time periods.

[0029] Optionally, before determining the pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness according to the data pair set, the method further includes:

[0030] For the data pair set, a variable distribution histogram is established for each operating parameter of the rolling process and the corresponding pole piece thickness;

[0031] Determining the weight value of the correlation between each operating parameter of the rolling and the thickness of the pole piece according to the variable distribution histogram;

[0032] Classifying a plurality of operating parameters into at least two categories of operating parameters according to the data pair set and the weight value;

[0033] According to the at least two types of operating parameters, based on feature engineering, a set of data pairs between each type of operating parameter and the pole piece thickness is determined.

[0034] According to another aspect of the present invention, this embodiment provides an electronic device, including:

[0035] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the rolling control method based on pole piece thickness prediction proposed in any item of the first aspect.

[0036] According to another aspect of the present invention, this embodiment provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement the rolling control method based on pole piece thickness prediction proposed in any item of the first aspect when executed.

[0037] According to another aspect of the present invention, this embodiment provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the rolling control method based on pole piece thickness prediction proposed in any of the above embodiments.

[0038] The rolling control method based on the prediction of the pole piece thickness provided by the embodiment of the present invention obtains a set of data pairs of the rolling operating parameters and the corresponding pole piece thickness, and determines a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operating parameters and the pole piece thickness according to the data pair set. Then, the current operating parameters of the rolling are input into the pole piece thickness prediction model to predict the value of the current pole piece thickness. Thus, the current operating parameters of the rolling are adjusted according to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold. This arrangement realizes the accurate prediction of the pole piece thickness through the pole piece thickness prediction model, and the operating parameters of the rolling are adjusted in real time according to the predicted value of the current pole piece thickness, so that the thickness consistency of the pole piece actually produced by the rolling is better, thereby improving the first-time excellent rate of the pole piece thickness. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the contents of the embodiments of the present invention and these drawings without paying any creative work.

[0040] Figure 1 is a flow chart of a rolling control method based on pole piece thickness prediction provided by an embodiment of the present invention;

[0041] Figure 2 is a flow chart of another rolling control method based on pole piece thickness prediction provided by an embodiment of the present invention;

[0042] Figure 3 It is a schematic diagram of a stacked regression model provided by an embodiment of the present invention outputting a predicted value of a corresponding electrode thickness when inputting target characteristic data;

[0043] Figure 4 It is a waveform diagram of the predicted value and the true value of the corresponding pole piece thickness outputted by a stacked regression model provided by an embodiment of the present invention when target characteristic data is input. DETAILED DESCRIPTION

[0044] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0045] Based on the above technical problems, this embodiment proposes the following solutions:

[0046] Figure 1is a flow chart of a rolling control method based on pole piece thickness prediction provided by an embodiment of the present invention. Figure 1 , the rolling control method based on pole piece thickness prediction provided by the embodiment of the present invention includes:

[0047] S101, obtaining a data pair set of rolling operation parameters and corresponding pole piece thickness.

[0048] Specifically, rolling is a manufacturing process of pressing the pole piece. The rollers include an upper roller and a lower roller. After the pole piece material passes through the roller gap between the upper roller and the lower roller, the pole piece is pressed into a pole piece of a preset thickness under the pressure between the upper roller and the lower roller.

[0049] Exemplarily, the process flow chart of the pole piece rolling process is as follows: coating - rolling unwinding - rolling deviation correction - rolling - laser thickness measurement - pre-slitting (optional) - rolling rewinding - slitting - slitting unwinding - slitting - slitting rewinding - slitting, etc. It should be noted that the process flow of the pole piece rolling process may or may not include the pre-slitting process.

[0050] The operating parameters of the upper and lower rollers during the working process can include the main factors that affect the consistency of the thickness of the pole piece during the rolling and slitting process of the pole piece, such as: roller straightness, roller jump, roller bending, etc. The factors affecting the straightness of the roller are mostly due to long-term use and the wear of the roller. The roller jump value is related to the rigidity of the roller. The better the rigidity, the smaller the roller jump value. Roller bending requires tension and the deformation resistance of the rolled piece to determine. The greater the deformation tension of the rolled piece, the greater the roller bending. The roller gap and pressure act together on the pole piece to ensure the normal thickness of the pole piece.

[0051] The structure of the pole piece after rolling is more stable, and the gaps between the particles of the pole piece after rolling are smaller. Rolling will affect the gram capacity, first coulomb efficiency, rate performance and cycle performance of the pole piece. Since the pole piece rolling process is more complicated from a mechanism analysis, it will be affected by many factors. It is necessary to obtain the historical operating parameters of rolling so as to predict the thickness of the pole piece based on the historical operating parameters of rolling.

[0052] The operating parameters of rolling may include the basic specifications of the rolling equipment, the equipment parameters of the rolling rollers, the rolling speed when rolling different batches or the same batch of pole pieces, the rolling meter, the upper bending cylinder pressure on the main rolling drive side, the lower bending cylinder pressure on the main rolling drive side, the wrinkle removal tension, the unwinding tension, and the preheating temperature. Each rolling operating parameter and the pole piece thickness form a data pair. When the rolling operating parameters include at least two operating parameters, the rolling operating parameters and the corresponding pole piece thickness data pairs form a data pair set. When the rolling operating parameters include a single operating parameter, the rolling operating parameters at different times and the corresponding pole piece thickness data pairs can form a data pair set.

[0053] S102: Determine, based on the data pair set, a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operation parameters and the pole piece thickness.

[0054] Specifically, according to the data pair set, the correlation characteristics between the rolling operation parameters and the pole piece thickness are analyzed, and a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operation parameters and the pole piece thickness is determined. The pole piece thickness prediction model can be determined by simulation or training.

[0055] S103, inputting the current operating parameters of the rolling into the pole piece thickness prediction model to predict the value of the current pole piece thickness.

[0056] Specifically, the current operating parameters are input into the pole piece thickness prediction model. The pole piece thickness prediction model makes a prediction based on the input current operating parameters, and outputs the value of the current pole piece thickness corresponding to the current operating parameters.

[0057] Exemplarily, the pole piece thickness prediction model may include a stacking regression model (Stacking), through which multiple single models are combined in a certain way by using the stacking regression model and the stacking ensemble learning method. The stacking regression model can combine the advantages of each single prediction model and stack multiple models in multiple layers. Finally, according to the current operating parameters, the value of the pole piece thickness is predicted to improve the accuracy of the pole piece thickness prediction.

[0058] S104. Adjust the current operating parameters of the rolling process according to the difference between the current pole piece thickness value and the target pole piece thickness threshold value.

[0059] Specifically, the operating parameters of the rolling mill will affect the thickness of the electrode sheet produced after rolling. The target electrode sheet thickness threshold can be preset according to the batch of electrode sheets, the purpose of the electrode sheets, etc. The electrode sheets whose predicted electrode sheet thickness is within the target electrode sheet thickness threshold range are electrode sheets that meet the requirements, and the electrode sheets outside the target electrode sheet thickness threshold range are electrode sheets that cannot meet the production needs. It can be set that when the value of the current electrode sheet thickness is outside the target electrode sheet thickness threshold range, the operating parameters of the rolling are adjusted so that the thickness of the electrode sheet actually produced by the rolling is within the target electrode sheet thickness threshold range to meet the production needs, thereby improving the first-time excellent rate of the electrode sheet thickness.

[0060] The rolling control method based on the prediction of the pole piece thickness provided in this embodiment obtains a set of data pairs of the rolling operating parameters and the corresponding pole piece thickness, and determines a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operating parameters and the pole piece thickness based on the data pair set. Then, the current operating parameters of the rolling are input into the pole piece thickness prediction model to predict the value of the current pole piece thickness. Thus, the current operating parameters of the rolling are adjusted according to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold. This arrangement realizes the accurate prediction of the pole piece thickness through the pole piece thickness prediction model, and the operating parameters of the rolling are adjusted in real time according to the predicted value of the current pole piece thickness, so that the thickness consistency of the pole piece actually produced by the rolling is better, thereby improving the first-time excellent rate of the pole piece thickness.

[0061] Optionally, the operating parameters of the rolling include at least one of the following: negative calibration value of the rolling, positive calibration value, number of meters, rolling speed, rolling recorded in meters, lower bending cylinder pressure on the auxiliary rolling transmission side, lower bending cylinder pressure on the main rolling transmission side, upper bending cylinder pressure on the main rolling operation side, lower bending cylinder pressure on the auxiliary rolling operation side, unwinding tension and winding tension.

[0062] Specifically, the unit of negative calibration of roller pressure can be set to um. The unit of positive calibration of roller pressure is um. The unit of meter is m, and the unit of roller speed is m / min. The unit of roller pressure meter is m. The unit of lower bending cylinder pressure on auxiliary rolling transmission side is t. The unit of lower bending cylinder pressure on main rolling transmission side is t. The unit of upper bending cylinder pressure on main rolling operation side is t. The unit of lower bending cylinder pressure on auxiliary rolling operation side is t. The unit of unwinding tension is N. Winding tension can include winding tension 0 and winding tension 1. The unit of winding tension 0 is N. The unit of winding tension 1 is N.

[0063] Optionally, based on the above embodiments, determining a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operating parameters and the pole piece thickness according to the data pair set includes: inputting the data pair set into a stacked regression model for learning, and determining a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operating parameters and the pole piece thickness.

[0064] Specifically, exemplarily, the pole piece thickness prediction model may include a stacking regression model. Take a Stacking integration framework of a stacking regression model with 2 layers as an example for explanation. Generally, the first layer model of the Stacking integration framework is composed of multiple base classifiers. The input of each base classifier is training data, and the second layer model of the Stacking integration framework is a meta classifier. The training data of the meta classifier is composed of the output of the base classifier of the first layer model and the label of the original training data. After the meta classifier is trained, a complete Stacking integration model can be obtained.

[0065] The stacked regression model uses a combination strategy to combine individual learners. The stacked regression model is to stack multiple sub-models on the same layer, and finally output the predicted value of the current electrode thickness. The stacked regression model relies on the results of multiple sub-models, that is, organizing the results of multiple weak learners, which often outperforms a single strong model.

[0066] For example, the first-layer model can be set to use a model with high fitting degree in order to fully learn the training data. Since different sub-models differ in principle and training set. In order to reduce the risk of overfitting, the second-layer model tends to use a simple model. Therefore, the first-layer model can obtain output values ​​with differences (low correlation) and good predictive ability for the original training data, so that after further learning through the second-layer model, it can learn from the strengths of multiple first-layer models and improve the accuracy and stability of the prediction.

[0067] In the stacked regression model, the individual learners can be called primary learners, and the learners used for combination are called secondary learners or meta-learners. The data used by the secondary learner for training is called the secondary training set, which is obtained by using the primary learner on the training set.

[0068] First, train the individual learner, that is, the primary learner. Then use the trained individual learner to generate the predicted results. The predicted results generated by the trained individual learner serve as the training set for the secondary learner. Based on the secondary learner, the final trained stacked regression model is obtained.

[0069] The data pair set is input into the trained stacked regression model, predicted by the primary learner, and then predicted by the secondary learner to output the value of the current pole piece thickness.

[0070] Optional, Figure 2 is a flow chart of another rolling control method based on pole piece thickness prediction provided by an embodiment of the present invention. Figure 2 ,The stacked regression models include random forest classifier model, redundant tree classifier model, adaptive boosting classifier model, gradient boosting classifier model and decision tree regression model.

[0071] The rolling control method based on pole piece thickness prediction provided by the embodiment of the present invention includes:

[0072] S101, obtaining a data pair set of rolling operation parameters and corresponding pole piece thickness.

[0073] S102: Determine, based on the data pair set, a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operation parameters and the pole piece thickness.

[0074] S301. Substitute the current operating parameters of rolling into the random forest classifier model to generate the first predicted thickness value of the pole piece. Substitute the current operating parameters of rolling into the redundant tree classifier model to generate the second predicted thickness value of the pole piece. Substitute the current operating parameters of rolling into the adaptive boosting classifier model to generate the third predicted thickness value of the pole piece. Substitute the current operating parameters of rolling into the gradient boosting classifier model to generate the fourth predicted thickness value of the pole piece. Substitute the current operating parameters of rolling into the decision tree regression model to generate the fifth predicted thickness value of the pole piece.

[0075] Specifically, the current operating parameters of the roller press are substituted into the stacking regression model. The first predicted thickness value of the pole piece is determined by the random forest classifier model. The second predicted thickness value of the pole piece is determined by the redundant tree classifier model. The third predicted thickness value of the pole piece is determined by the adaptive boosting classifier model. The fourth predicted thickness value of the pole piece is determined by the gradient boosting classifier model. And the fifth predicted thickness value of the pole piece is determined by the decision tree regression model.

[0076] S302, fusing the first predicted thickness value, the second predicted thickness value, the third predicted thickness value, the fourth predicted thickness value and the fifth predicted thickness value to determine a value of the current pole piece thickness.

[0077] Specifically, the first predicted thickness value, the second predicted thickness value, the third predicted thickness value, the fourth predicted thickness value, and the fifth predicted thickness value are fused by a stacked regression model, such as averaging or weighted averaging. The first predicted thickness value, the second predicted thickness value, the third predicted thickness value, the fourth predicted thickness value, and the fifth predicted thickness value can be combined to improve the accuracy and stability of the prediction, thereby outputting the value of the current pole piece thickness. The stacked regression model can output multiple values ​​of the current pole piece thickness to achieve multi-input and multi-output pole piece thickness prediction.

[0078] S104. Adjust the current operating parameters of the rolling process according to the difference between the current pole piece thickness value and the target pole piece thickness threshold value.

[0079] Optionally, on the basis of the above embodiment, adjusting the current operating parameters of the rolling according to the difference between the numerical value of the current pole piece thickness and the target pole piece thickness threshold may include: comparing the numerical value of the current pole piece thickness with the target pole piece thickness threshold; when the numerical value of the current pole piece thickness is greater than or less than the target pole piece thickness threshold, adjusting the current operating parameters of the rolling.

[0080] Specifically, the numerical value of the current pole piece thickness is compared with the target pole piece thickness threshold, and the numerical value of the current pole piece thickness and the target pole piece thickness threshold can be subtracted or divided. When the numerical value of the current pole piece thickness is greater than or less than the target pole piece thickness threshold, it means that the thickness of the pole piece is too thick or too thin. The thickness of the pole piece actually produced can be made consistent with the target pole piece thickness threshold by adjusting the operating parameters of the rolling press related to the thickness of the pole piece, thereby improving the first-time excellent rate of the pole piece thickness. Exemplarily, the target pole piece thickness threshold can include a thickness error of ±2μm.

[0081] Optionally, on the basis of the above embodiments, after adjusting the current operating parameters of the rolling, it may also include: when the value of the current electrode thickness is greater than or less than the target electrode thickness threshold, the primary quality rate of the produced electrode is judged, and an early warning is issued based on the judgment result.

[0082] Specifically, the primary optimum rate of the thickness of the electrode sheet is calculated according to the predicted value of the current electrode sheet thickness. A threshold of the primary optimum rate can be set. When the predicted primary optimum rate of the thickness of the electrode sheet is less than the threshold of the primary optimum rate, an early warning is issued. This setting facilitates timely monitoring of the primary optimum rate of the thickness of the electrode sheet produced according to the predicted value of the current electrode sheet thickness, and further improves the primary optimum rate of the thickness of the electrode sheet.

[0083] Figure 3 FIG. 1 is a schematic diagram of a stacked regression model provided by an embodiment of the present invention, which outputs a predicted value of the corresponding electrode thickness when inputting target feature data. Figure 3 , Figure 3 The stacking regression model includes multiple input values, namely, negative calibration value, positive calibration value, meter, rolling speed, rolling record, auxiliary rolling transmission side lower bending cylinder pressure, main rolling transmission side lower bending cylinder pressure, main rolling operation side upper bending cylinder pressure, auxiliary rolling operation side lower bending cylinder pressure, unwinding tension, winding tension 0 and winding tension 1. The stacking regression model includes multiple output values, namely, thickness 1, thickness 2, ..., thickness 15 and thickness 16 of the pole piece. Thickness 1, thickness 2, ..., thickness 15 and thickness 16 represent 16 different thickness prediction results respectively.

[0084] Figure 4 It is a waveform diagram of the predicted value and the true value of the corresponding electrode thickness output when the stacked regression model provided by the embodiment of the present invention inputs the target characteristic data. Figure 4 ,The blue color in the figure represents the true thickness value, and the red color represents the predicted thickness value. Figure 4 (a) in the figure shows the waveform diagram of the average thickness prediction value and the average thickness actual value of thickness 1. Figure 4 (b) in the figure shows the waveform diagram of the average thickness prediction value and the average thickness actual value of thickness 2. Figure 4(c) in the figure shows the waveform diagram of the average thickness prediction value and the average thickness actual value of thickness 3. Figure 4 (d) in the figure shows the waveform diagram of the average thickness prediction value and the average thickness actual value of thickness 15. Figure 4 (e) in FIG. 1 shows a waveform diagram of the average thickness prediction value and the average thickness true value of thickness 16. The errors between the average thickness prediction value and the average thickness true value corresponding to the thickness 1, thickness 2, ..., thickness 15 and thickness 16 of the pole piece are 0.9958, 0.9894, 0.686, 0.22, 0.8645, 0.9961, 0.9993, 0.2603, 0.3552, 0.3779, 0.3214, 0.2343, 0.3178, 0.2735, 0.3378 and 0.3237, respectively. It can be seen that when the stacked regression model inputs the current operating parameters, the deviation between the value of the current pole piece thickness corresponding to the output and the true value is small.

[0085] The rolling control method based on pole piece thickness prediction provided in this embodiment has a high prediction accuracy for the pole piece thickness, and can better compare the current pole piece thickness value with the target pole piece thickness threshold, and timely adjust the rolling operating parameters, thereby better improving the first-time optimization rate of the pole piece thickness.

[0086] Optionally, based on the above embodiments, obtaining a data pair set of rolling operating parameters and corresponding pole piece thickness includes: using a panel data tool to query a data pair set of rolling operating parameters and corresponding pole piece thickness, and generating a data pair set of multiple time periods.

[0087] Specifically, by setting the panel data tool to query the rolling operation parameters, a data pair set of multiple time periods is generated, and the panel data tool is used for visual query. By checking the data missing of the data pair set of the multiple time periods, it is convenient to supplement the missing data through the mean or maximum likelihood method, or eliminate the missing data set to obtain the data pair set after the missing processing. And the data pair set after the missing processing of the multiple time periods is statistically processed to achieve the preprocessing of the current rolling operation parameters and generate the preprocessed current rolling operation parameters. This setting is convenient to improve the accuracy of the current pole piece thickness value predicted according to the current rolling operation parameters, thereby improving the accuracy of the pole piece thickness prediction.

[0088] The rolling control method based on pole piece thickness prediction provided in the embodiment of the present application realizes the prediction of the value of the current pole piece thickness during rolling. The established Stacking prediction model is stable and reliable. The Stacking prediction model continuously optimizes the prediction capability through sustainable data analysis and training, with accurate prediction and high portability. The rolling control method based on pole piece thickness prediction provided in the embodiment of the present application is conducive to further promoting the large-scale application of machine learning and deep learning in various processes in lithium battery manufacturing, promoting the construction of digital twins, and laying a solid foundation for the construction of lighthouse factories.

[0089] Optionally, based on the above embodiments, before determining the pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operating parameters and the pole piece thickness according to the data pair set, it also includes: for the data pair set, for each rolling operating parameter and the corresponding pole piece thickness, a variable distribution histogram is established. According to the variable distribution histogram, the weight value of the correlation between each rolling operating parameter and the pole piece thickness is determined. According to the data pair set and the weight value, multiple operating parameters are classified into at least two categories of operating parameters. According to the at least two categories of operating parameters, based on feature engineering, a data pair set between each type of operating parameter and the pole piece thickness is determined.

[0090] Specifically, for the data pair set, a variable distribution histogram is established for each operating parameter of the rolling and the corresponding electrode thickness. For example, a variable distribution histogram is established for the negative calibration value, the positive calibration value, the meter, the rolling speed, the rolling meter, the lower bending cylinder pressure on the auxiliary rolling transmission side, the lower bending cylinder pressure on the main rolling transmission side, the upper bending cylinder pressure on the main rolling operation side, the lower bending cylinder pressure on the auxiliary rolling operation side, the unwinding tension, and the winding tension and the corresponding electrode thickness.

[0091] According to the variable distribution histogram, the weight value of the correlation between each operating parameter of the rolling and the pole piece thickness is determined. According to the data pair set and the weight value, multiple operating parameters are classified into at least two categories of operating parameters. Multiple operating parameters can be classified according to the size of the weight value. The weight values ​​of operating parameters of different categories are different. The larger the weight value can be set, the greater the influence of the operating parameter on the pole piece thickness. According to at least two categories of operating parameters, based on feature engineering, mathematics, statistics and domain knowledge are used to extract useful features from the classified operating parameters. The weight value can be used as a feature to determine the data pair set between each type of operating parameter and the pole piece thickness. In this way, the determined data pair set between each type of operating parameter and the pole piece thickness can improve the quality of the pole piece thickness prediction model, thereby improving the accuracy of the predicted value of the current pole piece thickness.

[0092] This embodiment provides an electronic device, comprising at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the rolling control method based on pole piece thickness prediction proposed in any item of the first aspect.

[0093] This embodiment provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement the rolling control method based on pole piece thickness prediction proposed in any item of the first aspect when executed.

[0094] This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the rolling control method based on pole piece thickness prediction proposed in any of the above embodiments.

[0095] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A rolling control method based on pole piece thickness prediction, characterized in that: include: Obtaining a data pair set of roller pressing operating parameters and corresponding pole piece thickness; Determining, according to the data pair set, a pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness; Inputting the current operating parameters of the roller press into the pole piece thickness prediction model to predict the value of the current pole piece thickness; According to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold, the current operating parameters of the rolling are adjusted.

2. The method according to claim 1, characterized in that The operating parameters of the roller press include at least one of the following: Negative calibration value of roll pressure, positive calibration value, meters, roll pressure speed, roll pressure meter, lower bending cylinder pressure on auxiliary rolling transmission side, lower bending cylinder pressure on main rolling transmission side, upper bending cylinder pressure on main rolling operation side, lower bending cylinder pressure on auxiliary rolling operation side, unwinding tension and winding tension.

3. The method according to claim 1 or 2, characterized in that: Determining, based on the data pair set, a pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness includes: The data pair set is input into a stacked regression model for learning, and a pole piece thickness prediction model that can characterize the correlation characteristics between the rolling operation parameters and the pole piece thickness is determined.

4. The method according to claim 3, characterized in that The stacked regression model includes a random forest classifier model, a redundant tree classifier model, an adaptive boosting classifier model, a gradient boosting classifier model and a decision tree regression model; Accordingly, the inputting of the current operating parameters of the rolling process into the pole piece thickness prediction model to predict the value of the current pole piece thickness includes: Substituting the current operating parameters of the roller press into the random forest classifier model to generate a first predicted thickness value of the pole piece; Substituting the current operating parameters of the roller press into the redundant tree classifier model to generate a second predicted thickness value of the pole piece; Substituting the current operating parameters of the roller press into the adaptive boosting classifier model to generate a third predicted thickness value of the pole piece; Substituting the current operating parameters of the roller press into the gradient boosting classifier model to generate a fourth predicted thickness value of the pole piece; Substituting the current operating parameters of the roller press into the decision tree regression model to generate a fifth predicted thickness value of the pole piece; The first predicted thickness value, the second predicted thickness value, the third predicted thickness value, the fourth predicted thickness value and the fifth predicted thickness value are merged to determine the value of the current pole piece thickness.

5. The method according to claim 1 or 2, characterized in that: The adjusting the current operating parameters of the rolling press according to the difference between the value of the current pole piece thickness and the target pole piece thickness threshold value includes: Comparing the value of the current pole piece thickness with a target pole piece thickness threshold; When the value of the current pole piece thickness is greater than or less than the target pole piece thickness threshold, the current operating parameters of the rolling process are adjusted.

6. The method according to claim 5, characterized in that After adjusting the current operating parameters of the roller press, the method further includes: When the value of the current electrode thickness is greater than or less than the target electrode thickness threshold, the primary quality rate of the produced electrode is judged, and an early warning is issued based on the judgment result.

7. The method according to claim 1 or 2, characterized in that: The data set for obtaining the rolling operating parameters and the corresponding pole piece thickness includes: Use the panel data tool to query the data pair set of the rolling operation parameters and the corresponding pole piece thickness, and generate data pair sets for multiple time periods.

8. The method according to claim 1 or 2, characterized in that: Before determining a pole piece thickness prediction model capable of characterizing the correlation characteristics between the rolling operation parameters and the pole piece thickness according to the data pair set, the method further includes: For the data pair set, a variable distribution histogram is established for each operating parameter of the rolling process and the corresponding pole piece thickness; Determining the weight value of the correlation between each operating parameter of the rolling and the thickness of the pole piece according to the variable distribution histogram; Classifying a plurality of operating parameters into at least two categories of operating parameters according to the data pair set and the weight value; According to the at least two types of operating parameters, based on feature engineering, a set of data pairs between each type of operating parameter and the pole piece thickness is determined.

9. An electronic device, characterized in that: include: at least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the rolling control method based on pole piece thickness prediction as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the rolling control method based on pole piece thickness prediction as described in any one of claims 1 to 8 when executed.

Images