Straightening force prediction method and device of straightening machine
By training multiple machine learning models for straightening force prediction, the problem of high straightening force calculation cost in the existing technology is solved, and efficient and accurate control of straightening force prediction is achieved, ensuring production safety and The equipment is stable.
Patent Information
- Application Number
- CN202411030394.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
The existing straightening force calculation methods rely on complex mechanism models, have high calculation costs and are difficult to meet the real-time requirements of production, and cannot effectively predict and control the straightening force, resulting in the impact of straightening effect and equipment safety risks.
Using machine learning-based straightening force prediction method, a variety of coupled machine learning models (such as CATBOOST, LightGBM, RandomForestRF and ExtraTrees models) are trained to perform data standardization and regularization processing by obtaining straightening machine equipment parameters and steel plate attribute parameters. , build a preset straightening force prediction model, predict the straightening force in real time and send it to the control system.
It realizes efficient, timely and accurate straightening force prediction, effectively controls straightening force, improves production efficiency and ensures production safety.
Smart Images

Figure CN118966434A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence application technology, and in particular to a method and device for predicting the straightening force of a straightening machine. Background Art
[0002] The straightening machine plays a vital role in the straightening process of medium and thick plates. It can ensure the flatness and quality of the steel plates. However, the straightening machine may face a straightening force that exceeds its designed load capacity during operation, which will not only affect the straightening effect, but may also cause damage to the straightening machine equipment. Therefore, accurately predicting and controlling the straightening force to avoid exceeding the load limit of the straightening machine is the key to ensuring production efficiency and safety of the straightening machine equipment.
[0003] Existing straightening force calculation methods rely on complex mechanism models, which often need to consider the actual stiffness of the straightening machine and the accuracy of the rolling force formula. The calculation cost is high and it is difficult to meet the real-time requirements of production. Summary of the invention
[0004] In view of the problems in the prior art, an embodiment of the present invention provides a method and device for predicting the straightening force of a straightening machine, which can at least partially solve the problems in the prior art.
[0005] On the one hand, the present invention provides a straightening force prediction method of a straightening machine, comprising:
[0006] Obtain straightening machine equipment parameters and steel plate property parameters;
[0007] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0008] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0009] Wherein, obtaining the straightening force prediction sample of the straightening machine includes:
[0010] Collecting an original data set; the original data set is a data set consisting of straightening machine equipment parameters and steel plate property parameters used in the model training stage;
[0011] Each type of data in the original data set is preprocessed respectively, and each type of data after the preprocessing is used as a straightening force prediction sample of the straightening machine.
[0012] The preprocessing of each type of data in the original data set includes:
[0013] Performing data standardization processing on each type of data in the original data set;
[0014] Each type of data after data standardization is subjected to data regularization processing respectively to obtain each type of data after preprocessing.
[0015] Among them, a plurality of coupled machine learning models are trained according to the straightening force prediction samples of the straightening machine to obtain the straightening force prediction model of the preset straightening machine, including:
[0016] Determine a plurality of groups of model combination parameters corresponding to each coupled machine learning model respectively;
[0017] Training multiple groups of model combination parameters corresponding to each coupled machine learning model, and determining the coupled target machine learning model and the corresponding target model combination parameters according to the training results;
[0018] The target machine learning model, the target model combination parameters corresponding thereto, and the straightening force prediction samples of the straightening machine are subjected to linear regression processing to obtain the straightening force prediction model of the preset straightening machine.
[0019] Among them, each coupled machine learning model includes a CATBOOST model, a LightGBM model, a RandomForestRF model and an ExtraTrees model; accordingly, multiple sets of model combination parameters corresponding to each coupled machine learning model are determined respectively, including:
[0020] Determine multiple groups of model combination parameters with the same number of groups corresponding to the CATBOOST model, the LightGBM model, the RandomForestRF model, and the ExtraTrees model, respectively.
[0021] The step of determining the coupled target machine learning model and the target model combination parameters corresponding thereto according to the training results includes:
[0022] Selecting p groups of target model combination parameters with the highest prediction accuracy from the training results, and selecting q groups of target model combination parameters with the lowest prediction accuracy from the training results;
[0023] The p groups of target model combination parameters and the q groups of target model combination parameters are removed, and the retained machine learning model and the model combination parameters corresponding thereto are used as the target machine learning model and the target model combination parameters corresponding thereto for determining the coupling.
[0024] Wherein, after the step of obtaining the straightening force prediction result, the straightening force prediction method of the straightening machine further includes:
[0025] The straightening force prediction result is sent to a control system so that the control system can control the straightening force of the straightening machine according to the straightening force prediction result.
[0026] On the one hand, the present invention provides a straightening force prediction device for a straightening machine, comprising:
[0027] An acquisition unit, used for acquiring equipment parameters of the straightening machine and property parameters of the steel plate;
[0028] A prediction unit, used for predicting the straightening machine equipment parameters and the steel plate property parameters based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0029] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0030] In another aspect, an embodiment of the present invention provides an electronic device, including: a processor, a memory, and a bus, wherein:
[0031] The processor and the memory communicate with each other via the bus;
[0032] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the following method:
[0033] Obtain straightening machine equipment parameters and steel plate property parameters;
[0034] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0035] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0036] An embodiment of the present invention provides a non-transitory computer-readable storage medium, including:
[0037] The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the following method:
[0038] Obtain straightening machine equipment parameters and steel plate property parameters;
[0039] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0040] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0041] The straightening force prediction method and device of a straightening machine provided in an embodiment of the present invention obtain the equipment parameters of the straightening machine and the property parameters of the steel plate; predict the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result; wherein the preset straightening force prediction model of the straightening machine is obtained by training multiple coupled machine learning models based on the straightening force prediction samples of the straightening machine, and can efficiently, timely and accurately predict the straightening force of the straightening machine, thereby effectively controlling the straightening force of the straightening machine, improving production efficiency and ensuring production safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. 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 these drawings without creative work. In the drawings:
[0043] Figure 1 It is a flow chart of a straightening force prediction method of a straightening machine provided in one embodiment of the present invention.
[0044] Figure 2 It is a flow chart of a straightening force prediction method of a straightening machine provided in another embodiment of the present invention.
[0045] Figure 3 It is a structural schematic diagram of a straightening force prediction device for a straightening machine provided in one embodiment of the present invention.
[0046] Figure 4 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, in the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily.
[0048] Figure 1 FIG. 1 is a flow chart of a straightening force prediction method for a straightening machine provided by an embodiment of the present invention. Figure 1 As shown, the straightening force prediction method of the straightening machine provided by the embodiment of the present invention includes:
[0049] Step S1: Obtain straightening machine equipment parameters and steel plate property parameters.
[0050] Step S2: predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0051] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0052] In the above step S1, the device obtains the straightening machine equipment parameters and the steel plate property parameters. The device may be a computer device that executes the method, for example, it may include a server, etc. The steel plate property parameters may include the steel plate thickness, in mm, the steel plate width, in mm, the steel plate temperature before straightening, in ° C, and the yield limit corresponding to the steel type, in MPa. The straightening machine equipment parameters may include the inlet absolute roll gap, in mm, the outlet absolute roll gap, in mm, and the bending roll amount, in mm. Here, it refers to the straightening machine equipment parameters and steel plate property parameters used in the model use stage.
[0053] In the above step S2, the device predicts the straightening machine equipment parameters and the steel plate property parameters based on the straightening force prediction model of the preset straightening machine to obtain a straightening force prediction result; the straightening machine equipment parameters and the steel plate property parameters can be input into the straightening force prediction model of the preset straightening machine, and the output result of the straightening force prediction model of the preset straightening machine is used as the straightening force prediction result.
[0054] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0055] Obtaining a straightening force prediction sample of the straightening machine, including:
[0056] Collecting an original data set; the original data set is a data set consisting of straightening machine equipment parameters and steel plate property parameters used in the model training stage;
[0057] Each type of data in the original data set is preprocessed respectively, and each type of data after the preprocessing is used as a straightening force prediction sample of the straightening machine.
[0058] The preprocessing of each type of data in the original data set comprises:
[0059] Performing data standardization processing on each type of data in the original data set;
[0060] Each type of data after data standardization is subjected to data regularization processing respectively to obtain each type of data after preprocessing.
[0061] The data standardization formula is as follows:
[0062]
[0063] Among them, X represents each type of data, X norm Represents each type of data after data standardization.
[0064] The data regularization processing formula is as follows:
[0065] X′=log e (1+X norm )
[0066] Where X′ represents each type of data after preprocessing, i.e., the subsequent straightening force prediction sample of the straightening machine. Before continuing to explain the training of the straightening force prediction model of the preset straightening machine, the training model architecture is first explained as follows: Figure 2 As shown in the figure, since the data in the training process is less than that in the neural network training, the prediction model is regressed and predicted by machine learning. The model is modified by several methods of ensemble learning. To improve the accuracy, the overall Stacking method is adopted. By coupling multiple models together, the four groups of models with the highest and lowest accuracy are discarded in the training results to prevent overfitting of the model and inaccurate prediction. The data is preprocessed in the selected parameters and characterization is done in advance.
[0067] According to the straightening force prediction samples of the straightening machine, a plurality of coupled machine learning models are trained to obtain the straightening force prediction model of the preset straightening machine, including:
[0068] Determine a plurality of groups of model combination parameters corresponding to each coupled machine learning model respectively;
[0069] Training multiple groups of model combination parameters corresponding to each coupled machine learning model, and determining the coupled target machine learning model and the corresponding target model combination parameters according to the training results;
[0070] The target machine learning model, the target model combination parameters corresponding thereto, and the straightening force prediction samples of the straightening machine are subjected to linear regression processing to obtain the straightening force prediction model of the preset straightening machine.
[0071] Each coupled machine learning model includes a CATBOOST model, a LightGBM model, a RandomForestRF model, and an ExtraTrees model; accordingly, multiple sets of model combination parameters corresponding to each coupled machine learning model are determined respectively, including:
[0072] Determine multiple groups of model combination parameters with the same number of groups corresponding to the CATBOOST model, LightGBM model, RandomForestRF model and ExtraTrees model. Take the four models with 5 groups as an example, as follows:
[0073] The five groups of model combination parameters corresponding to the CATBOOST model are shown in Table 1:
[0074] Table 1
[0075]
[0076] Parameters not set in Table 1 are default values, and parameters are set to default values.
[0077] The default parameters and their values for the CATBOOST model include:
[0078] 1.iterations: The default number of iterations is 10,000.
[0079] 2.learning_rate: The default learning rate.
[0080] 3.depth: The default tree depth.
[0081] 4.l2_leaf_reg: default L2 regularization parameter.
[0082] 5.model_size_reg: The default model size regularization parameter is `1`.
[0083] 6.rsm: The default random state mode is `1`.
[0084] 7.colsample_bylevel: The default feature sampling ratio for each layer is `1`.
[0085] 8.max_ctr_complexity: default CTR complexity.
[0086] 9.min_data_in_leaf: The default minimum amount of data for each leaf node is `1`.
[0087] 10.one_hot_max_size: The default maximum number of categories for one-hot encoding is `16`.
[0088] 11.border_count: The default number of borders is `255`.
[0089] 12.random_seed: The default random seed is `0`.
[0090] 13.loss_function: The default loss function is `'Logloss'`.
[0091] 14.eval_metric: The default evaluation metric is `'Accuracy'`.
[0092] 15.od_type: The default overfitting detector type is `'Iter'`.
[0093] For the five groups of model combination parameters corresponding to the LightGBM model, as shown in Table 2:
[0094] Table 2
[0095]
[0096] Parameters not set in Table 2 are default values. Other default parameters include:
[0097] boosting_type or boosting: The default is gbdt, which is the traditional gradient boosted decision tree.
[0098] Objective: The default is regression, which is used for regression tasks.
[0099] num_iterations or num_tree: The default value is 100, which indicates the number of boosting iterations or the number of trees.
[0100] learning_rate: represents the shrinkage rate of each tree.
[0101] num_leaves: indicates the maximum number of leaf nodes in a tree.
[0102] tree_learner: The default is serial, which means a tree learner for a single machine.
[0103] num_threads: The default value is OpenMP_default, which means using the default number of threads of OpenMP.
[0104] min_data_in_leaf: indicates the minimum amount of data in a leaf node, used to deal with overfitting.
[0105] min_sum_hessian_in_leaf: The default value is 1e-3, which means the minimum sum of Hessians in a leaf node, and is also used to deal with overfitting.
[0106] feature_fraction: represents the sampling ratio of the feature.
[0107] bagging_fraction: The default value is 1.0, which indicates the sampling ratio of the data.
[0108] max_depth: The default value is -1, which means there is no limit on the maximum depth of the tree.
[0109] min_data_per_group: The default value is 100, which indicates the minimum amount of data for each classification group.
[0110] extra_trees: Indicates whether to use extreme random trees.
[0111] For the 5 groups of model combination parameters corresponding to the RandomForestRF model, as shown in Table 3:
[0112] Table 3
[0113]
[0114] Parameters not set in Table 3 are default values:
[0115] n_estimators: The number of trees in the forest (default is 300).
[0116] Criterion: Function to measure the quality of the split (default is 'squared_error', i.e. mean squared error).
[0117] max_depth: Maximum depth of the tree (default is None, i.e. the tree will grow until all leaves are pure or contain fewer than min_samples_split samples).
[0118] min_samples_split: The minimum number of samples required for internal node re-split (default is 2).
[0119] min_samples_leaf: The minimum number of samples required for a leaf node (default is 1).
[0120] min_weight_fraction_leaf: The minimum sum of sample weights required for a leaf node (default is 0.0).
[0121] max_features: The number of features to consider when finding the best split (default is 1.0, i.e. all features are considered).
[0122] max_leaf_nodes: Maximum number of leaf nodes (default is None, i.e. no limit).
[0123] min_impurity_decrease: Minimum impurity reduction required to split a node (default is 0.0).
[0124] bootstrap: Whether to use the bootstrap method when building the tree (default is True).
[0125] oob_score: Whether to use out-of-bag samples to estimate generalization accuracy (default is False).
[0126] n_jobs: The number of CPUs to use for parallel work for training and prediction (default is None, i.e. use a single CPU).
[0127] random_state: random number seed (default is None, i.e. no seed is set).
[0128] Verbose: Parameter that controls the level of detail (the default is 0, which means no detailed information is output).
[0129] warm_start: When set to True, reuse the solution from the previous call to fit additional trees (default is False).
[0130] ccp_alpha: Parameter for complexity pruning (default is 0.0).
[0131] max_samples: The maximum number of samples used for sampling when building each tree (the default is None, that is, all samples are used).
[0132] The five groups of model combination parameters corresponding to the ExtraTrees model are shown in Table 4:
[0133] Table 4
[0134]
[0135] Parameters not set in Table 4 are default values:
[0136] n_estimators=100: The number of trees in the forest.
[0137] criterion = 'squared_error': splitting criterion (mean squared error).
[0138] max_depth=None: Maximum depth of the tree.
[0139] min_samples_split=2: The minimum number of samples required to split an internal node.
[0140] min_samples_leaf=1: The minimum number of samples required for a leaf node.
[0141] min_weight_fraction_leaf=0.0: The minimum weighted fraction of sample weights required for leaf nodes.
[0142] max_features = 'auto' : Maximum number of features to consider at each split.
[0143] max_leaf_nodes=None: maximum number of leaf nodes.
[0144] min_impurity_decrease=0.0: Minimum impurity reduction at a split point.
[0145] bootstrap=False: Whether to use bootstrap samples.
[0146] oob_score=False: Whether to use out-of-bag samples to estimate generalization accuracy.
[0147] n_jobs=None: The number of parallel jobs.
[0148] random_state=None: random seed.
[0149] verbose=0: Whether to enable detailed output.
[0150] warm_start=False: When set to True, reuse the solution from the previous call to increase the fit
[0151] ccp_alpha=0.0: complexity penalty.
[0152] max_samples=None: The number of samples to train each tree.
[0153] The step of determining the coupled target machine learning model and the target model combination parameters corresponding thereto according to the training results includes:
[0154] Selecting p groups of target model combination parameters with the highest prediction accuracy from the training results, and selecting q groups of target model combination parameters with the lowest prediction accuracy from the training results;
[0155] Remove the p group target model combination parameters and the q group target model combination parameters, and use the remaining machine learning model and the corresponding model combination parameters as the target machine learning model and the corresponding target model combination parameters for determining the coupling. The values of p and q can be set independently according to actual conditions, and the two can be equal or unequal, and can be selected as 4. Referring to the above example, the following is explained:
[0156] Remove the 4 models with the highest prediction accuracy and the 4 models with the lowest prediction accuracy from the training results, and then there are 12 models left, with 12 prediction values, which together with the previous input values form a matrix of length 12. Combined with the number of collected data, the matrix of length 7 can be regarded as 19 output parameters. Then perform linear regression, and the abstract formula can be expressed as follows:
[0157]
[0158] in, represents the final predicted value, f i ′(x i ) represents the abstract function of the i-th input parameter xi, and n is equal to the number of collected data, 7. i (y i ) represents the predicted value of the i-th prediction model. m refers to the number of prediction models, 12, that is, the number of target machine learning models and their corresponding target model combination parameters.
[0159] After the step of obtaining the straightening force prediction result, the straightening force prediction method of the straightening machine further includes:
[0160] The straightening force prediction result is sent to the control system so that the control system can control the straightening force of the straightening machine according to the straightening force prediction result. The straightening force prediction result can be sent to the control system, and actual production data will be generated in real time during online production. The actual production data can be used to retrain the model regularly in a scheduled manner to continuously improve the prediction accuracy of the model.
[0161] The machine learning model can learn the relationship between the straightening force and various influencing factors from historical data. With the accumulation of training data and the iteration of the model, its prediction ability will gradually increase. In addition, the flexibility and adaptability of the machine learning model enable it to quickly respond to changes in the production process and adjust the straightening parameters in real time, thereby protecting the straightening machine from damage while improving production efficiency and product quality.
[0162] Predicting the maximum straightening force of the straightening machine through machine learning can play a protective role in the production process of the straightening machine. Machine learning can discover the objective laws of straightening force calculation from the data. At the same time, new data can be continuously used to train the model during the production process to improve the accuracy of the model. Through the combination of multiple different learners, the objective laws of the straightening machine can be better extracted, which can maximize the applicability and accuracy of the prediction model.
[0163] The straightening force prediction method of a straightening machine provided in an embodiment of the present invention obtains the straightening machine equipment parameters and the steel plate property parameters; the straightening machine equipment parameters and the steel plate property parameters are predicted based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result; wherein the preset straightening force prediction model of the straightening machine is obtained by training multiple coupled machine learning models based on the straightening force prediction samples of the straightening machine, and can efficiently, timely and accurately predict the straightening force of the straightening machine, thereby effectively controlling the straightening force of the straightening machine, improving production efficiency and ensuring production safety.
[0164] Further, obtaining a straightening force prediction sample of the straightening machine includes:
[0165] Collecting an original data set; the original data set is a data set consisting of straightening machine equipment parameters and steel plate property parameters used in the model training stage; the above-mentioned embodiment can be referred to for description and will not be repeated here.
[0166] Each type of data in the original data set is preprocessed respectively, and each type of data after preprocessing is used as a straightening force prediction sample of the straightening machine. The above description can be referred to in the embodiment, which will not be repeated here.
[0167] Furthermore, the preprocessing of each type of data in the original data set includes:
[0168] Data standardization is performed on each type of data in the original data set; please refer to the above embodiment for details and will not be described in detail.
[0169] Each type of data after data standardization is subjected to data regularization processing to obtain each type of data after preprocessing. The above description can be referred to in the above embodiment and will not be repeated here.
[0170] Furthermore, a plurality of coupled machine learning models are trained according to the straightening force prediction samples of the straightening machine to obtain the straightening force prediction model of the preset straightening machine, including:
[0171] Determine the multiple groups of model combination parameters corresponding to each coupled machine learning model respectively; refer to the above embodiment for description and no further details will be given.
[0172] Train multiple groups of model combination parameters corresponding to each coupled machine learning model, and determine the coupled target machine learning model and the corresponding target model combination parameters based on the training results; refer to the above-mentioned embodiment for description and no further details will be given.
[0173] The target machine learning model and the target model combination parameters corresponding thereto, as well as the straightening force prediction samples of the straightening machine are subjected to linear regression processing to obtain the straightening force prediction model of the preset straightening machine. The above-mentioned embodiments can be referred to for explanation and will not be described in detail.
[0174] Furthermore, each coupled machine learning model includes a CATBOOST model, a LightGBM model, a RandomForestRF model, and an ExtraTrees model; accordingly, multiple groups of model combination parameters corresponding to each coupled machine learning model are determined respectively, including:
[0175] Determine multiple groups of model combination parameters with the same number of groups corresponding to the CATBOOST model, the LightGBM model, the RandomForestRF model and the ExtraTrees model, respectively. The above embodiments can be referred to for explanation and will not be repeated here.
[0176] Furthermore, the step of determining the coupled target machine learning model and the target model combination parameters corresponding thereto according to the training results includes:
[0177] Select p groups of target model combination parameters with the highest prediction accuracy from the training results, and select q groups of target model combination parameters with the lowest prediction accuracy from the training results; refer to the above embodiment for description and no further details will be given.
[0178] The p groups of target model combination parameters and the q groups of target model combination parameters are removed, and the remaining machine learning model and the model combination parameters corresponding thereto are used as the target machine learning model and the target model combination parameters corresponding thereto for determining the coupling. The above-mentioned embodiments can be referred to for explanation and will not be described in detail.
[0179] Furthermore, after the step of obtaining the straightening force prediction result, the straightening force prediction method of the straightening machine further includes:
[0180] The straightening force prediction result is sent to a control system so that the control system can control the straightening force of the straightening machine according to the straightening force prediction result.
[0181] Figure 3 FIG. 1 is a schematic diagram of the structure of a straightening force prediction device for a straightening machine provided by an embodiment of the present invention. Figure 3As shown, the straightening force prediction device of the straightening machine provided by the embodiment of the present invention includes an acquisition unit 301 and a prediction unit 302, wherein:
[0182] The acquisition unit 301 is used to acquire the straightening machine equipment parameters and the steel plate property parameters; the prediction unit 302 is used to predict the straightening machine equipment parameters and the steel plate property parameters based on the straightening force prediction model of the preset straightening machine to obtain the straightening force prediction result; wherein, the straightening force prediction model of the preset straightening machine is obtained by training multiple coupled machine learning models based on the straightening force prediction samples of the straightening machine.
[0183] Specifically, the acquisition unit 301 in the device is used to acquire the straightening machine equipment parameters and the steel plate property parameters; the prediction unit 302 is used to predict the straightening machine equipment parameters and the steel plate property parameters based on the straightening force prediction model of the preset straightening machine to obtain the straightening force prediction result; wherein, the straightening force prediction model of the preset straightening machine is obtained by training multiple coupled machine learning models based on the straightening force prediction samples of the straightening machine.
[0184] The straightening force prediction device of the straightening machine provided in the embodiment of the present invention obtains the equipment parameters of the straightening machine and the property parameters of the steel plate; predicts the equipment parameters of the straightening machine and the property parameters of the steel plate based on the straightening force prediction model of the preset straightening machine, and obtains the straightening force prediction result; wherein, the straightening force prediction model of the preset straightening machine is obtained by training multiple coupled machine learning models according to the straightening force prediction samples of the straightening machine, and can efficiently, timely and accurately predict the straightening force of the straightening machine, thereby effectively controlling the straightening force of the straightening machine, improving production efficiency and ensuring production safety.
[0185] The embodiment of the straightening force prediction device of the straightening machine provided by the embodiment of the present invention can be specifically used to execute the processing flow of the above-mentioned method embodiments. Its functions are not repeated here, and reference can be made to the detailed description of the above-mentioned method embodiments.
[0186] Figure 4 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as Figure 4 As shown, the electronic device includes: a processor (processor) 401, a memory (memory) 402 and a bus 403;
[0187] The processor 401 and the memory 402 communicate with each other via a bus 403;
[0188] The processor 401 is used to call the program instructions in the memory 402 to execute the methods provided by the above method embodiments, for example, including:
[0189] Obtain straightening machine equipment parameters and steel plate property parameters;
[0190] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0191] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0192] This embodiment discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can perform the methods provided by the above method embodiments, for example, including:
[0193] Obtain straightening machine equipment parameters and steel plate property parameters;
[0194] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0195] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0196] This embodiment provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program enables the computer to execute the methods provided by the above method embodiments, for example, including:
[0197] Obtain straightening machine equipment parameters and steel plate property parameters;
[0198] Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result;
[0199] The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
[0200] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0201] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or 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, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0202] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0203] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0204] In the description of this specification, the description with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0205] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for predicting the straightening force of a straightening machine, characterized in that: include: Obtain straightening machine equipment parameters and steel plate property parameters; Predicting the equipment parameters of the straightening machine and the property parameters of the steel plate based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result; The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
2. The straightening force prediction method of a straightening machine according to claim 1, characterized in that: Obtaining a straightening force prediction sample of the straightening machine, including: Collecting an original data set; the original data set is a data set consisting of straightening machine equipment parameters and steel plate property parameters used in the model training stage; Each type of data in the original data set is preprocessed respectively, and each type of data after the preprocessing is used as a straightening force prediction sample of the straightening machine.
3. The straightening force prediction method of a straightening machine according to claim 2, characterized in that: The preprocessing of each type of data in the original data set comprises: Performing data standardization processing on each type of data in the original data set; Each type of data after data standardization is subjected to data regularization processing respectively to obtain each type of data after preprocessing.
4. The straightening force prediction method of a straightening machine according to claim 1, characterized in that: According to the straightening force prediction samples of the straightening machine, a plurality of coupled machine learning models are trained to obtain the straightening force prediction model of the preset straightening machine, including: Determine a plurality of groups of model combination parameters corresponding to each coupled machine learning model respectively; Training multiple groups of model combination parameters corresponding to each coupled machine learning model, and determining the coupled target machine learning model and the corresponding target model combination parameters according to the training results; The target machine learning model, the target model combination parameters corresponding thereto, and the straightening force prediction samples of the straightening machine are subjected to linear regression processing to obtain the straightening force prediction model of the preset straightening machine.
5. The straightening force prediction method of a straightening machine according to claim 4, characterized in that: Each coupled machine learning model includes a CATBOOST model, a LightGBM model, a RandomForestRF model, and an ExtraTrees model; accordingly, multiple sets of model combination parameters corresponding to each coupled machine learning model are determined respectively, including: Determine multiple groups of model combination parameters with the same number of groups corresponding to the CATBOOST model, the LightGBM model, the RandomForestRF model, and the ExtraTrees model, respectively.
6. The straightening force prediction method of a straightening machine according to claim 5, characterized in that: The step of determining the coupled target machine learning model and the target model combination parameters corresponding thereto according to the training results includes: Selecting p groups of target model combination parameters with the highest prediction accuracy from the training results, and selecting q groups of target model combination parameters with the lowest prediction accuracy from the training results; The p groups of target model combination parameters and the q groups of target model combination parameters are removed, and the retained machine learning model and the model combination parameters corresponding thereto are used as the target machine learning model and the target model combination parameters corresponding thereto for determining the coupling.
7. The straightening force prediction method of a straightening machine according to any one of claims 1 to 6, characterized in that: After the step of obtaining the straightening force prediction result, the straightening force prediction method of the straightening machine further includes: The straightening force prediction result is sent to a control system so that the control system can control the straightening force of the straightening machine according to the straightening force prediction result.
8. A straightening force prediction device for a straightening machine, characterized in that: include: An acquisition unit, used for acquiring equipment parameters of the straightening machine and property parameters of the steel plate; A prediction unit, used for predicting the straightening machine equipment parameters and the steel plate property parameters based on a preset straightening force prediction model of the straightening machine to obtain a straightening force prediction result; The straightening force prediction model of the preset straightening machine is obtained by training a plurality of coupled machine learning models according to the straightening force prediction samples of the straightening machine.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
