Disk shear shearing parameter optimization method and device

Abstract

The invention discloses a disc shear shearing parameter optimization method and device, and relates to the technical field of cold-rolled strip steel production, and the method comprises the steps: collecting process data and production data, obtaining the collected process data and production data through the communication with a PLC, carrying out the image defect recognition of the edge part of the sheared strip steel through an industrial camera, and carrying out the image defect recognition of the edge part of the sheared strip steel; obtaining a classification result of the severity of the strip steel defects, and preprocessing the classification result; establishing a sample space, taking the trimming detection quality as a sample label, taking other data as sample features, forming a sample, selecting a KNN algorithm to carry out training completion, and storing a training model after training completion; calculating by constructing a parameter vector space and substituting the parameter vector space into the trained training model, selecting an optimal parameter set value, and maintaining an original value if the optimal parameter set value is not optimal. By means of the disc edge cutting device, the failure rate during disc edge cutting work can be reduced, the edge cutting production stability can be improved, and the edge cutting quality is improved.

Description

Technical Field

[0001] The invention relates to the technical field of cold-rolled strip steel production, and in particular to a method and device for optimizing shearing parameters of a disc shear. Background Art

[0002] In cold-rolled strip mills, the cold-rolling and pickling mill units are equipped with non-driven turret-type circular shears. The steel strip, pulled by the tension of the tension roller, enters the upper and lower shear blades of the circular shear for shearing. The shear blades move in a circular motion as the steel strip moves, shearing the edges of the strip. The turret-type circular shear is equipped with two sets of shear blades, which can be replaced offline during production to ensure continuous operation of the production line. Before the shear blades begin shearing, the steel strip undergoes plastic deformation, forming an arc area. As the shear blades move, they press into the steel strip to shear, forming a relatively smooth shear surface. Its width is approximately 1 / 3 to 1 / 5 of the thickness of the steel strip. During the shearing process, as the cracks migrate, a relatively rough tear surface is produced on the remaining portion of the steel strip.

[0003] The shear blade clearance and overlap are key control parameters for a circular shear. The adjustment ranges for these two parameters are determined by process engineers and require comprehensive consideration of multiple relevant data points. During production, operators modify and adjust the blade overlap and clearance based on actual production data, but such adjustments cannot guarantee reliability and accuracy. Summary of the Invention

[0004] The embodiments of the present application provide a method and device for optimizing the shearing parameters of a disc shear, which solves the problems of large adjustment lag, adjustment accuracy being extremely dependent on the operator's experience and unstable adjustment effect. A disc shear shearing model based on machine learning is designed based on production data (such as strip thickness, yield strength, production speed, and trimming detection quality) and process data (such as production speed, disc shear gap and overlap) to automatically find the optimal control parameters under the current state.

[0005] In a first aspect, the present application provides a method for optimizing shearing parameters of a disc shear, the method comprising:

[0006] Collect process data and production data, and obtain the collected process data and production data through communication with the PLC. Use industrial cameras to identify image defects on the edges of the sheared strip to obtain a classification result based on the severity of the strip defects. The classification results are preprocessed and stored in the system data. The preprocessing process includes deleting blank values, invalid values, and singular values.

[0007] Based on the preprocessing results stored in the system data, a sample space is established. The trimming detection quality is used as the sample label, and other data are used as sample features to form a sample. 75% of the data in the sample space is used as model training data, and 25% is used as test data. The KNN algorithm is selected for training and the training model is saved after training.

[0008] By constructing the parameter vector space and substituting it into the training model after training, the optimal parameter setting value is selected. If there is no optimal value, the original value is maintained.

[0009] Furthermore, the process data includes the disc shear gap and overlap, and the production data includes production speed, strip steel grade, strip steel thickness and trimming inspection quality evaluation, wherein the strip steel grade data is divided into six categories according to yield strength.

[0010] Furthermore, 75% of the data in the sample space is used as model training data, and 25% is used as test data. The KNN algorithm is selected for training, and the training model after training is saved, which includes:

[0011] 75% of the data in the sample space are randomly selected and allocated as the model training dataset for the model learning process; the remaining 25% of the data are randomly selected and allocated as the model testing dataset for evaluating the performance of the model;

[0012] The KNN algorithm is used to train the model training data set. The KNN algorithm predicts the category of unknown samples by calculating the distance between samples. The parameters of the KNN algorithm are adjusted during the training process to achieve the best classification effect.

[0013] After the model training is completed, the trained model parameters and structure are saved to the system hard disk.

[0014] Furthermore, the KNN algorithm is used to train the model training data set, and the specific steps are as follows:

[0015] Step 1: Initialize the parameters of the KNN algorithm, including selecting an appropriate number of neighbors K;

[0016] Step 2: For each sample in the training data set, calculate its distance to other samples in the training set;

[0017] Step 3: For each sample, select the K nearest neighbor samples based on the calculated distance;

[0018] Step 4: Determine the predicted label of the current sample by majority voting based on the labels of K neighbor samples;

[0019] Step 5: Repeat steps 2 to 4 until all training samples have been processed;

[0020] Step 6: Optimize the performance of the model by adjusting the K value and other possible parameters.

[0021] Furthermore, the parameter vector space is constructed and substituted into the training model after training to perform calculations, and the optimal parameter setting value is selected. If there is no optimal value, the original value is maintained, which includes:

[0022] Set the upper and lower limits of the disk shear gap and overlap, take M values ​​evenly, and construct the parameter vector space Ω;

[0023] Collect other process and production data, substitute each element in Ω into the model calculation, and obtain the trimming inspection quality classification value;

[0024] Select the parameter with the result of 0 as the new set value. If there are multiple parameters with the result of 0, select randomly. If there is no parameter with the result of 0, select the parameter with the result of 1. If there are multiple parameters with the result of 0, select randomly. If there is no parameter with the result of 0 or 1, do not change it and continue to use the original set value.

[0025] Furthermore, in the process of collecting process data and production data, the sensor detects and records the plate shape data of the position along the length direction. After the plate shape data is input into the input system, the system calculates the absolute value average of the corresponding data of the four positions of the entire roll of strip steel.

[0026] The present application provides a disc shear shear parameter optimization device, the device comprising:

[0027] Acquisition module: used to collect process data and production data, and obtain the collected process data and production data through communication with the PLC. The industrial camera is used to identify image defects on the edge of the sheared strip to obtain a classification result based on the severity of the strip defects. The classification result is pre-processed and stored in the system data. The pre-processing process includes deleting blank values, invalid values, and singular values.

[0028] Training module: used to establish a sample space based on the preprocessing results stored in the system data, using the trimming detection quality as the sample label and other data as sample features to form a sample. 75% of the data in the sample space is used as model training data and 25% as test data. The KNN algorithm is selected for training and the training model is saved after training.

[0029] Calculation module: used to construct a parameter vector space and substitute it into the training model after training to perform calculations, select the optimal parameter setting value, and maintain the original value if there is no optimal value.

[0030] Furthermore, the process data in the acquisition module includes the disc shear gap and overlap, and the production data includes production speed, strip steel grade, strip steel thickness and trimming inspection quality evaluation, among which the strip steel grade data is divided into six categories according to yield strength.

[0031] Furthermore, the training module includes:

[0032] Allocation unit: randomly selects and allocates 75% of the data in the sample space as the model training dataset for the model learning process; randomly selects and allocates the remaining 25% of the data as the model testing dataset for evaluating the performance of the model;

[0033] Training prediction unit: used to train the model training data set using the KNN algorithm, where the KNN algorithm predicts the category of unknown samples by calculating the distance between samples. During the training process, the parameters of the KNN algorithm are adjusted to achieve the best classification effect;

[0034] Saving unit: used to save the trained model parameters and structure to the system hard disk after the model training is completed.

[0035] Furthermore, the calculation module includes:

[0036] Construction unit: used to set the upper and lower limits of the disc shear gap and overlap, evenly taking M values ​​to construct the parameter vector space Ω;

[0037] Collection and calculation unit: used to collect other process and production data, substitute each element in Ω into the model calculation, and obtain the trimming inspection quality classification value;

[0038] Selection judgment unit: used to select the parameter with a result of 0 as the new set value. If there are multiple parameters with a result of 0, it will be randomly selected; if there is no parameter with a result of 0, the parameter with a result of 1 will be selected. If there are multiple parameters, it will be randomly selected; if there is no parameter with a result of 0 or 1, it will not be changed and the original set value will continue to be used.

[0039] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0040] This embodiment involves a method and device for optimizing the shearing parameters of a cold-rolled disc shear in cold rolling production. The system can automatically adjust and match the key shearing parameters of the disc shear according to the current production process data and production data to improve the quality of the disc shear edge, reduce the edge defects of the strip after shearing, and achieve stable production of cold-rolled strip.

[0041] The present invention provides a system and device for optimizing the shearing parameters of a disc shear based on strip shape data, on-site process data, and production data. This system searches for the optimal shearing parameters of the disc shear during production, which can not only reduce the failure rate of the disc shear during operation, but also improve the stability of shearing production and the quality of shearing.

[0042] The present invention adopts a graphical interface for human-computer interaction, which can facilitate the operation and processing of quality management personnel and on-site operators; "personalized" disc shear shearing parameters can be saved according to different varieties and specifications of strip steel products, and saved in the database for easy access at any time.

[0043] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 A schematic structural diagram of a device for optimizing shearing parameters of a disc shear provided in this embodiment;

[0046] Figure 2 A schematic flow chart of a method for optimizing shearing parameters of a disc shear provided in this embodiment;

[0047] Figure 3 A schematic diagram of gap and overlap adjustment provided in this embodiment;

[0048] Figure 4 Schematic diagram of the disc shear shearing parameter optimization system provided in this embodiment;

[0049] Figure 5 Schematic diagram of the hot-rolled plate shape data selection area provided in this embodiment;

[0050] Figure 6 Schematic diagram of the human-machine main interface of the disc shear shearing parameter optimization system provided in this embodiment;

[0051] Figure 7 Schematic diagram of the historical plate shape data selection and calculation interface provided in this embodiment. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0053] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner.In the following description, many specific details are provided so as to provide a full understanding of the embodiments of the present application. However, it will be appreciated by those skilled in the art that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps etc. can be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0054] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0055] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0056] In existing technology, cold-rolled strip edges must be trimmed to meet customer product width requirements, eliminate edge defects, and improve edge quality. The disc shears in the pickling mill are located before the rolling mill. The quality of these trimming operations not only affects product quality but also indirectly impacts mill production and unit output.

[0057] The shearing parameters of the disc shear mainly include gap amount and overlap amount. By adjusting their values, the shearing parameters are optimized, and combined with parameters such as shearing speed and blade installation accuracy, the best edge shearing effect can be achieved.

[0058] If the disc shear gap is not adjusted properly, it will not only shorten the service life of the disc shear blade, but also affect the quality of the steel strip edge. In production, this adjustment amount is within a range, generally 7% to 11% of the strip thickness.

[0059] The overlap primarily affects whether the strip can smoothly engage the upper and lower shear blades. If the overlap is too small, the strip edge bends during shearing, and the scrap strip becomes difficult to cut, causing uncut scrap to become stuck under the strip or to accumulate in the chute. If the overlap is too large, the shear blades wear out, making the strip impossible to cut or causing the strip edge to protrude from above the scrap guide chute.

[0060] In addition, the specifications of the strip (such as thickness), steel type (such as yield strength), production speed, edge shape defects of hot-rolled incoming materials, etc. will all affect the shearing quality of the disc shear. The final shearing quality is a comprehensive result.

[0061] In actual operation, the shearing parameters of the disc shear are set according to the process data table in the secondary system, and then manually adjusted according to the actual production situation. However, such adjustments have a large lag, and the adjustment accuracy is extremely dependent on the operator's experience and the adjustment effect is unstable. In order to solve this problem, the embodiment of the present application provides a method for optimizing the shearing parameters of a disc shear, designs a disc shear shearing model based on machine learning based on production data (such as strip thickness, yield strength, production speed, and trimming detection quality) and process data (such as production speed, disc shear gap and overlap), and is used to automatically find the optimal control parameters under the current state.

[0062] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0063] This embodiment first relies on Figure 1 The device for optimizing the shearing parameters of the disc shear shown provides the following Figure 2 A method for optimizing shearing parameters of a disc shear is shown, and the method includes steps S100 to S400.

[0064] S100, collecting process data and production data, and obtaining the collected process data and production data through communication with the PLC, performing image defect recognition on the edge of the sheared strip through an industrial camera to obtain a classification result of the severity of the strip defect, and preprocessing the classification result to obtain the preprocessing result and store it in the system data, wherein the preprocessing process includes deleting blank values, invalid values ​​and singular values.

[0065] It should be noted that in this step, the process data includes the disc shear gap and overlap, and the production data includes production speed, strip steel grade, strip steel thickness and trimming inspection quality evaluation, among which the strip steel grade data is divided into six categories according to yield strength.

[0066] It is understandable that, Figure 3As shown in Figure 1, the process data are the disc shear gap and overlap parameters. These parameters are communicated to the system through the PLC that controls the disc shear. The production data are production speed, strip steel grade, strip thickness, and trimming quality evaluation classification. The system communication of these data is shown in Table 1:

[0067] Table 1 Production data communication objects

[0068] Data Object Communication object Data Type Production speed Master PLC Integer type Strip steel grades Secondary unit Dummy variable coding Strip thickness Secondary unit Integer type Edge trimming quality inspection Edge defect detection system Categorical values

[0069] Because the production data contains a large number of strip steel grades, this system divides them into six categories based on yield strength and processes them using dummy variable processing. The specific processing is shown in Table 2.

[0070] Table 2 Dummy variable processing of strip steel grades

[0071]

[0072] The trimming quality parameters in production data serve as "label values" for machine learning. This system obtains this data through communication with an on-site edge defect detection system. This on-site edge defect detection system is specialized quality inspection equipment. It uses an industrial camera to photograph the edges of the sheared strip at regular intervals. It then identifies defects in these images and categorizes the severity of the strip defects, as shown in Table 3.

[0073] Table 3 Classification of cutting edge quality

[0074] category Numerical No defects 0 Occasionally visible defects, not obvious 1 There are defects, but they are not obvious 2 There are obvious defects such as holes, gaps, and notches 3

[0075] Hot rolled strip shape data is obtained by communicating with the secondary system. Because the disc shear only cuts the material part of the strip edge, the hot rolled strip shape data only uses the values ​​of the two sensors on the edge, namely the transmission side sensor 1, transmission side sensor 2, and the operation side sensor 1, operation side sensor 2, such as Figure 5 During data collection, the sensor detects and records the flatness data at each location along the length. After the flatness data is transmitted to the system, the system calculates the average absolute value of the corresponding data at these four locations across the entire coil, as shown in Table 4.

[0076] Table 4 Calculation of average value of strip edge shape data

[0077]

[0078] After collecting the data, the data is preprocessed, including deleting blank values, invalid values ​​and singular values, and saved in the system database.

[0079] S200. Based on the preprocessing results stored in the system data, a sample space is established, and the trimming detection quality is used as the sample label and other data are used as sample features to form a sample. 75% of the data in the sample space is used as model training data and 25% as test data. The KNN algorithm is selected for training and the training model is saved after the training is completed.

[0080] It should be noted that step S200 includes S201, S202 and S203, wherein:

[0081] S201, randomly selecting and allocating 75% of the data in the sample space as a model training dataset for the model learning process; randomly selecting and allocating the remaining 25% of the data as a model testing dataset for evaluating the performance of the model;

[0082] S202, using the KNN algorithm to train the model training data set, wherein the KNN algorithm predicts the category of unknown samples by calculating the distance between samples, wherein the parameters of the KNN algorithm are adjusted during the training process to achieve the best classification effect;

[0083] S203: After the model training is completed, the model parameters and structure obtained by training are saved to the system hard disk.

[0084] In this step, after data preprocessing, a sample space is established, with trimming quality used as the sample label and other data used as sample features. The sample features and sample labels constitute a sample, and all samples within a continuous rolling period constitute the sample space. 75% of the data in this sample space is used as model training data, and 25% as test data for model building and testing. The KNN algorithm is selected as the machine learning algorithm. The KNN algorithm is a general machine learning algorithm that performs well when the data volume or dimensionality are small, making it well-suited for this business scenario. After model training is complete, it is saved to the system hard drive.

[0085] Therefore, in step S202, the KNN algorithm is used to train the model training data set. The specific steps are as follows:

[0086] Step 1: Initialize the parameters of the KNN algorithm, including selecting an appropriate number of neighbors K;

[0087] Step 2: For each sample in the training data set, calculate its distance to other samples in the training set;

[0088] Step 3: For each sample, select the K nearest neighbor samples based on the calculated distance;

[0089] Step 4: Determine the predicted label of the current sample by majority voting based on the labels of K neighbor samples;

[0090] Step 5: Repeat steps 2 to 4 until all training samples have been processed;

[0091] Step 6: Optimize the performance of the model by adjusting the K value and other possible parameters.

[0092] Finally, model evaluation and saving: Use the model test dataset to evaluate the trained KNN model. Evaluation indicators may include accuracy, recall rate, and F1 score. After the evaluation is completed, the trained model parameters and structure with the best performance are saved to the system hard disk for subsequent testing and practical application.

[0093] S300: Calculate the optimal parameter setting value by constructing a parameter vector space and substituting it into the training model after training. If there is no optimal value, maintain the original value.

[0094] In this step S300, including S301, S302 and S303, wherein:

[0095] S301, setting the upper and lower limits of the disk shear gap and overlap, taking M values ​​evenly, and constructing the parameter vector space Ω;

[0096] S302, collecting other process and production data, substituting each element in Ω into the model calculation, and obtaining the trimming inspection quality classification value;

[0097] S303. Select a parameter with a result of 0 as a new set value. If there are multiple parameters with a result of 0, select randomly. If there is no parameter with a result of 0, select a parameter with a result of 1. If there are multiple parameters with a result of 1, select randomly. If there is no parameter with a result of 0 or 1, do not change it and continue to use the original set value.

[0098] It is understandable that in the content provided by the present invention, during production, the system can optimize the parameters of the disc shear gap and overlap. Assume that the upper and lower limits of the disc shear gap are (Gmax, Gmin), and the upper and lower limits of the disc shear overlap are (Lmax, Lmin). In the intervals of (Gmax, Gmin) and (Lmax, Lmin), M values ​​are uniformly taken, recorded as (G1, G2, G3...GM), (L1, L2, L3...LM), and (G1, G2, G3...GM), (L1, L2, L3...LM) are arranged and combined in the form of (Gi, Lj), where 1≤i≤M, 1≤j≤M, thus constructing M2 parameter vector spaces Ω.

[0099] The system collects other process and production data, and substitutes each element in this vector space Ω into the model calculation in turn, so that M2 trimming inspection quality classification values ​​can be obtained.

[0100] The parameter with a result of 0 is selected as the new setting value of the disk shear gap and overlap. If there is more than one parameter with a result of 0, it is randomly selected.

[0101] If the result is not 0, select the parameter with the result of 1. If there is more than one parameter with the result of 1, select it randomly.

[0102] If the result is neither 0 nor 1, no change is made and the setting value issued by the original secondary machine is still used.

[0103] In some embodiments, the present invention proposes a method and apparatus for optimizing shearing parameters of a disc shear based on production and process data, the main implementation methods of which are as follows:

[0104] (1) Hardware configuration

[0105] Firstly, the hardware of the disk shear shearing parameter optimization system consists of a system database, a system server and a network connection device, such as Figure 4 As shown in the figure, data modeling and disk shear parameter optimization operations are performed on the server's human-machine interface. The disk shear parameter optimization system communicates with the edge defect detection system, the factory's secondary machine server, the on-site PLC control system, and the disk shear control PLC via a TCP / IP network.

[0106] (2) Control system operation

[0107] Among them, the control system operation interface is as follows Figure 6 The "Process Parameters" area displays the system's current process data, production data, and circular shear parameters. Circular shear parameters, including gap and overlap, display the original settings issued by the secondary level. DS01 and DS02 represent the average absolute values ​​of the flatness recorded by two sensors on the transmission side of the coil in production. OS01 and OS02 represent the average absolute values ​​of the flatness recorded by two sensors on the operating side of the coil in production.

[0108] Therefore, after clicking "Optimal Calculation", the system will display the calculation results of the disc shear control parameters, namely the gap and overlap, in the right column based on the trained model and the optimal parameter search process.

[0109] As production continues, the working conditions and environmental parameters on site also change, and the prediction accuracy of the old model for the input data under the new working conditions will continue to decline. The system provides an option to update the model. Click "Data Selection" on the operation interface to enter the historical flatness data selection and calculation interface. Figure 7 Select the historical data that needs to be trained in the selection box, click "Calculate", and the system will recalculate the plate shape data based on the selected data.

[0110] Finally, after returning to the main interface, click "Model Update" and the system will recalculate the model based on the new data and save it in the system database.

[0111] This embodiment provides a disc shear shearing parameter optimization device, such as Figure 2 As shown, the device includes:

[0112] Acquisition module: used to collect process data and production data, and obtain the collected process data and production data through communication with the PLC. The industrial camera is used to identify image defects on the edge of the sheared strip to obtain a classification result based on the severity of the strip defects. The classification result is pre-processed and stored in the system data. The pre-processing process includes deleting blank values, invalid values, and singular values.

[0113] Training module: used to establish a sample space based on the preprocessing results stored in the system data, using the trimming detection quality as the sample label and other data as sample features to form a sample. 75% of the data in the sample space is used as model training data and 25% as test data. The KNN algorithm is selected for training and the training model is saved after training.

[0114] Calculation module: used to construct a parameter vector space and substitute it into the training model after training to perform calculations, select the optimal parameter setting value, and maintain the original value if there is no optimal value.

[0115] Specifically, the process data in the acquisition module includes the disc shear gap and overlap, and the production data includes production speed, strip steel grade, strip steel thickness and trimming inspection quality evaluation, among which the strip steel grade data is divided into six categories according to yield strength.

[0116] Specifically, the training module includes:

[0117] Allocation unit: randomly selects and allocates 75% of the data in the sample space as the model training dataset for the model learning process; randomly selects and allocates the remaining 25% of the data as the model testing dataset for evaluating the performance of the model;

[0118] Training prediction unit: used to train the model training data set using the KNN algorithm, where the KNN algorithm predicts the category of unknown samples by calculating the distance between samples. During the training process, the parameters of the KNN algorithm are adjusted to achieve the best classification effect;

[0119] Saving unit: used to save the trained model parameters and structure to the system hard disk after the model training is completed.

[0120] Specifically, the calculation module includes:

[0121] Construction unit: used to set the upper and lower limits of the disc shear gap and overlap, evenly taking M values ​​to construct the parameter vector space Ω;

[0122] Collection and calculation unit: used to collect other process and production data, substitute each element in Ω into the model calculation, and obtain the trimming inspection quality classification value;

[0123] Selection judgment unit: used to select the parameter with a result of 0 as the new set value. If there are multiple parameters with a result of 0, it will be randomly selected; if there is no parameter with a result of 0, the parameter with a result of 1 will be selected. If there are multiple parameters, it will be randomly selected; if there is no parameter with a result of 0 or 1, it will not be changed and the original set value will continue to be used.

[0124] It should be noted that, regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0125] Corresponding to the above method embodiment, this embodiment also provides a device for optimizing the shearing parameters of a disc shear. The device for optimizing the shearing parameters of a disc shear described below and the method for optimizing the shearing parameters of a disc shear described above can refer to each other.

[0126] Since the electronic device described in this embodiment is an electronic device used to implement the information processing method in the embodiment of this application, based on the information processing method described in the embodiment of this application, those skilled in the art will be able to understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be described in detail here. As long as those skilled in the art implement the electronic device used by the information processing method in the embodiment of this application, it falls within the scope of protection to be provided by this application.

[0127] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, 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 magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), 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 a 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.

[0129] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work 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 The function specified in one or more boxes.

[0130] 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 the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0131] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0132] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

[0133] In summary, the present invention provides a method for optimizing the shearing parameters of a disc shear, a system and device for optimizing the shearing parameters of a disc shear based on strip plate shape data, on-site process data and production data. This system searches for the optimal shearing parameters of the disc shear during production, which can not only reduce the failure rate of the disc shearing edge during operation, but also improve the stability of the cutting production and improve the cutting quality; a graphical interface is used for human-computer interaction, which can facilitate the operation and processing of quality management personnel and on-site operators; "personalized" disc shear shearing parameters can be saved according to different varieties and specifications of strip steel products, and saved in the database for call at any time.

[0134] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for optimizing shearing parameters of a disc shear, characterized in that: include: Collect process data and production data, and obtain the collected process data and production data through communication with PLC, use industrial camera to identify image defects on the edge of the sheared strip to obtain classification results of the severity of strip defects, and pre-process the classification results to obtain pre-processing results and store them in system data, wherein the pre-processing process includes deleting blank values, invalid values ​​and singular values; Based on the preprocessing results stored in the system data, a sample space is established, the trimming detection quality is used as the sample label, and other data are used as sample features to form a sample. 75% of the data in the sample space is used as model training data, and 25% is used as test data. The KNN algorithm is selected for training, and the training model is saved after the training is completed. By constructing the parameter vector space and substituting it into the training model after training, the optimal parameter setting value is selected. If there is no optimal value, the original value is maintained.

2. The method for optimizing shearing parameters of a disc shear according to claim 1, characterized in that: The process data include the disc shear gap and overlap, and the production data include production speed, strip steel grade, strip steel thickness and edge trimming inspection quality evaluation, wherein the strip steel grade data are divided into six categories according to yield strength.

3. The method for optimizing shearing parameters of a disc shear according to claim 1, characterized in that: The method uses 75% of the data in the sample space as model training data and 25% as test data, selects the KNN algorithm for training, and saves the training model after training, including: 75% of the data in the sample space are randomly selected and allocated as the model training data set for the model learning process; the remaining 25% of the data are randomly selected and allocated as the model testing data set for evaluating the performance of the model; The KNN algorithm is used to train the model training data set. The KNN algorithm predicts the category of unknown samples by calculating the distance between samples. The parameters of the KNN algorithm are adjusted during the training process to achieve the best classification effect. After the model training is completed, the trained model parameters and structure are saved to the system hard disk.

4. The method for optimizing shearing parameters of a disc shear according to claim 3, characterized in that: The KNN algorithm is used to train the model training data set. The specific steps are as follows: Step 1: Initialize the parameters of the KNN algorithm, including selecting an appropriate number of neighbors K; Step 2: For each sample in the training data set, calculate its distance from other samples in the training set; Step 3: For each sample, select the K nearest neighbor samples based on the calculated distance; Step 4: Determine the predicted label of the current sample by majority voting based on the labels of K neighbor samples; Step 5: Repeat steps 2 to 4 until all training samples have been processed; Step 6: Optimize the performance of the model by adjusting the K value and other possible parameters.

5. The method for optimizing shearing parameters of a disc shear according to claim 1, characterized in that: The method constructs a parameter vector space and substitutes it into the training model after training to calculate and select the optimal parameter setting value. If there is no optimal value, the original value is maintained, which includes: Set the upper and lower limits of the disk shear gap and overlap, take M values ​​evenly, and construct the parameter vector space Ω; Collect other process and production data, substitute each element in Ω into the model calculation, and obtain the quality classification value of the trimming inspection; Select the parameter with the result of 0 as the new set value. If there are multiple parameters with the result of 0, select randomly. If there is no parameter with the result of 0, select the parameter with the result of 1. If there are multiple parameters with the result of 0, select randomly. If there is no parameter with the result of 0 or 1, do not change it and continue to use the original set value.

6. The method for optimizing shearing parameters of a disc shear according to claim 1, characterized in that: In the process of collecting process data and production data, the sensor detects and records the plate shape data of the position along the length direction. After the plate shape data is input into the input system, the system calculates the absolute value average of the corresponding data of the four positions of the entire roll of strip steel.

7. A disc shear shearing parameter optimization device, based on the disc shear shearing parameter optimization method according to claim 1, characterized in that: include: Acquisition module: used to collect process data and production data, and obtain the collected process data and production data through communication with PLC, use industrial camera to identify image defects on the edge of the sheared strip to obtain the classification result of the severity of the strip defects, and pre-process the classification results to obtain the pre-processing results and store them in the system data, where the pre-processing process includes deleting blank values, invalid values ​​and singular values; Training module: used to establish sample space based on the preprocessing results stored in the system data, use the trimming detection quality as the sample label, and other data as sample features to form the sample, and use 75% of the data in the sample space as model training data and 25% as test data, select the KNN algorithm for training, and save the training model after training; Calculation module: used to construct parameter vector space and substitute it into the training model after training to perform calculations, select the optimal parameter setting value, and maintain the original value if there is no optimal one.

8. The device for optimizing shearing parameters of a disc shear according to claim 6, characterized in that: The process data in the acquisition module include the disc shear gap and overlap, and the production data include production speed, strip steel grade, strip steel thickness and trimming inspection quality evaluation, wherein the strip steel grade data is divided into six categories according to yield strength.

9. The device for optimizing shearing parameters of a disc shear according to claim 6, characterized in that: The training module includes: Allocation unit: used to randomly select and allocate 75% of the data in the sample space as the model training data set for the model learning process; randomly select and allocate the remaining 25% of the data as the model testing data set for evaluating the performance of the model; Training prediction unit: used to train the model training data set using the KNN algorithm, where the KNN algorithm predicts the category of unknown samples by calculating the distance between samples, and the parameters of the KNN algorithm are adjusted during the training process to achieve the best classification effect; Saving unit: used to save the trained model parameters and structure to the system hard disk after the model training is completed.

10. The device for optimizing shearing parameters of a disc shear according to claim 6, characterized in that: The computing module includes: Construction unit: used to set the upper and lower limits of the disc shear gap and overlap, evenly taking M values ​​to construct the parameter vector space Ω; Collection and calculation unit: used to collect other process and production data, substitute each element in Ω into the model calculation, and obtain the trimming inspection quality classification value; Selection judgment unit: used to select the parameter with the result of 0 as the new set value. If there are multiple parameters with the result of 0, select randomly; if there is no parameter with the result of 0, select the parameter with the result of 1. If there are multiple parameters, select randomly; if there is no parameter with the result of 0 or 1, do not change it and continue to use the original set value.

Images