Method for determining steel passing amount of rough rolling working roller of newly-built hot continuous rolling line

The BP neural network model predicts the wear amount of work rollers in hot continuous rolling line and determines the appropriate oversteerage amount, which solves the problem of peeling the support rollers caused by wear of rough rolling work rollers in the production of new hot continuous rolling line, and improves production safety and rhythm stability.

CN118904907AActive Publication Date: 2024-11-08HUNAN VALIN LIANYUAN IRON & STEEL CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411116049.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-08
Estimated Expiration
2044-08-14

AI Technical Summary

Technical Problem

In the production of new hot continuous rolling lines, wear of rough rolling work rolls leads to an increase in local pressure of the support roller, which may cause roll surface peeling accidents. It is difficult for the existing technology to determine the appropriate amount of steel oversteer in a timely manner, affecting production safety and rhythm.

Method used

Based on the BP neural network, the work roller wear model is established. By collecting relevant data from multiple hot rolling lines, the work roller wear amount is predicted, and the appropriate oversteerage amount is determined to reduce the peeling accident of the supporting roller surface and avoid frequent replacement of work. Rollers affect production.

Benefits of technology

It effectively reduces the accident of partial roller surface peeling of the support roller, ensures that the working roller has a certain amount of steel, avoids excessive wear, and ensures the safety of the rolling roller and stable production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118904907A_ABST
    Figure CN118904907A_ABST
Patent Text Reader

Abstract

The invention provides a method for determining the steel passing amount of a rough rolling working roller of a newly-built hot continuous rolling line. The method comprises the following steps that firstly, related data of the hot continuous rolling line are collected; 2, wear loss data of a working roller are collected; 3, performing normalization processing on the data, and dividing the data into a training set and a test set; 4, establishing a working roll wear model, taking the collected data as input, and taking the working roll wear loss as prediction output; 5, training the working roll wear model, and adjusting the model according to an error back propagation algorithm; 6, testing the wear model of the working roll; step 7, carrying out reverse normalization processing on the output data; and 8, data of the newly-built hot continuous rolling line are input, and the appropriate steel passing amount is determined. The working roll wear model is established based on the BP neural network, so that the use safety of the rough rolling supporting roll is ensured while the production rhythm is prevented from being influenced by frequent replacement of the rough rolling working roll.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of hot rolling, and in particular to a method for determining the amount of steel overload of a rough rolling working roll of a newly built hot rolling line. Background Art

[0002] The rough rolling process of the hot rolling line rolls the raw slab into a relatively thin intermediate slab, with a large single-pass reduction and a large rolling force. The rough rolling work roll is in direct contact with the slab, and the rough rolling backup roll is only in contact with the rough rolling work roll. Under the same over-steel amount, the wear of the rough rolling backup roll is much less than that of the rough rolling work roll. Under normal circumstances, the rough rolling backup roll has good contact with the work roll, and the contact pressure between the rolls is evenly distributed. However, when the rough rolling work roll has been on the machine for a period of time and has obvious wear and tear, the uniform distribution of contact pressure between the rolls will be broken, and a contact pressure peak will appear, resulting in an increase in the local pressure of the backup roll. In severe cases, the backup roll will peel off, causing the backup roll to be scrapped, seriously affecting the safety of the roll.

[0003] For a newly built hot rolling production line, the product specifications are complex in the initial stage of production, the production plan is unstable, and the roll oversteel amount is not clearly specified. It is often the case that a set of rough rolling work rolls are on the machine for half a month or even longer, resulting in the rough rolling work rolls being unable to be replaced in time when they are worn, causing the rough rolling support rolls to be locally severely compressed and the roll surface to peel off. In the prior art, the appropriate oversteel amount of the newly built hot rolling production line is often determined by actual measurement during the production process on the machine to guide the replacement of the rough rolling work rolls. However, due to the large changes in products during the production process, it is difficult to determine the appropriate oversteel amount data within a short production cycle of a type of product, and it is difficult to avoid the problem of excessive wear of the rough rolling work rolls causing local roll surface peeling of the support rolls during the actual measurement process, which is prone to production accidents.

[0004] Application Contents

[0005] The present application is made in view of the above-mentioned subject, and its purpose is to provide a method for determining the oversteel amount of the roughing working roll of a newly built hot rolling line, establish a working roll wear model based on the BP neural network, collect relevant data of multiple hot rolling lines to train the model, and then use the model to predict the wear amount of the working roll of the newly built hot rolling line to select the appropriate oversteel amount. The appropriate oversteel amount can be selected before product production, which greatly reduces the local peeling of the supporting roll surface accidents. While ensuring that the roughing working roll has a certain oversteel amount and avoiding frequent replacement of the roughing working roll to affect the production rhythm, it avoids the peeling of the roughing supporting roll surface caused by excessive wear of the roughing working roll, thereby ensuring the safe use of the roughing supporting roll.

[0006] Specifically, the first aspect of the present application provides a method for determining the amount of steel overload of a rough rolling work roll of a newly built hot rolling line, comprising the following steps:

[0007] Step 1: Collect wear-related data of multiple hot rolling lines and steel overload data of rough rolling work rolls of each hot rolling line in multiple different online cycles;

[0008] The data that has a greater impact on the wear of the rough rolling work rolls on the hot rolling line are collected as input data for the work roll wear model to better predict the change of work roll wear on this hot rolling line with the increase of over-steel amount, so as to determine the appropriate over-steel amount.

[0009] Step 2: Collect the wear data of the rough rolling work rolls in the corresponding online cycle;

[0010] The actual rough rolling work roll wear data collected is used for model training and comparison of test results. The model is optimized based on the error between the model prediction results and the actual results to achieve better prediction results.

[0011] Step 3: Normalize the collected data and divide the collected data into training set and test set;

[0012] Data normalization is an important preprocessing technique in data analysis and machine learning. Its purpose is to convert the original data into a fixed range or make all data have the same scale, thereby eliminating the impact of dimension and order of magnitude differences and improving the training effect and generalization ability of the model.

[0013] The main function of the training set is to train the model and estimate the parameters in the model through training data. These parameters enable the model to reflect the data relationships in the real world and make predictions.

[0014] The purpose of the test set is to evaluate the performance of the model on unseen data, that is, the generalization ability of the model. The test set can be used to verify whether the model can continue to work effectively on new, unseen data. This helps to discover possible overfitting problems in the model, that is, the model performs well on the training data but poorly on new data.

[0015] Step 4: Establish a work roll wear model based on BP neural network, taking the data collected in step 1 as input and the work roll wear amount as predicted output;

[0016] Step 5: Train the work roll wear model with the training set data, compare the output value with the collected rough rolling work roll wear data, and adjust the model according to the error back propagation algorithm;

[0017] Step 6: Test the work roll wear model with the test set data until the error is less than the threshold or the maximum number of iterations is reached;

[0018] The error threshold is used to determine whether the model has converged to an acceptable level of accuracy. The maximum number of iterations refers to the maximum number of iterations allowed during the iteration process. When this number is reached, the algorithm will stop running regardless of whether the current error is small enough. This is mainly to control the running time and computing resources of the algorithm.

[0019] Step 7: Denormalize the output data and draw the relationship curve between the steel amount data of each hot rolling line and the wear amount of the working roll;

[0020] Denormalization is the process of restoring normalized data back to its original size. When normalizing, we usually scale the data using the minimum and maximum values, while denormalizing requires restoring the data to its original scale based on these parameters.

[0021] Step 8: Input the wear-related data and over-steel test data of the newly built hot rolling line into the working roll wear model, and determine the appropriate over-steel amount based on the slope data in the relationship curve.

[0022] Furthermore, the wear-related data include rolling force, total number of rolling passes, steel grade hardness, rolling speed, strip width, and strip length.

[0023] Furthermore, the normalization process uses a Min-Max standardization method to convert the original data into data within the range of 0 to 1.

[0024] The Min-Max standardization method scales the data to the interval [0, 1], where Min is the minimum value 0 and Max is the maximum value 1.

[0025] Furthermore, the collected data is divided into a training set and a test set, specifically including: taking the wear-related data of a hot rolling line, the steel overload data of a machine cycle and the corresponding working roll wear data as a complete data sample, and performing data division based on the data sample.

[0026] Furthermore, the training set accounts for 70% to 80% of the total data.

[0027] This ratio division ensures that there is enough data to train the model while retaining a portion of the data for evaluating the model's generalization ability.

[0028] Furthermore, the BP neural network includes an input layer, a hidden layer and an output layer.

[0029] The input layer is the starting point of the entire neural network. It is responsible for receiving external input data signals. Its main function is to pass the received signals to the next layer, the hidden layer, to provide basic data for subsequent processing. The hidden layer is located between the input layer and the output layer and is the core part of the neural network. Its main function is to perform nonlinear transformation and feature extraction on the input data to adapt to complex practical problems. The hidden layer performs nonlinear transformation by weighted summation of input signals and applies activation functions to improve the network's expressive power. The output layer is the final output end of the neural network and is responsible for generating the final prediction results or classification results.

[0030] Furthermore, the learning rate of the BP neural network is set to 0.1.

[0031] The learning rate is one of the important hyperparameters for training neural networks. It represents the step size of the gradient moving toward the optimal solution of the loss function in each iteration. It is usually represented by the symbol η, and its size determines the speed of network learning. A suitable learning rate can ensure stable learning of the network while reducing training time with reasonable high-speed convergence. If the learning rate is too high, it may cause oscillation or even divergence during training; if the learning rate is too low, it may lead to slow training speed and difficulty in achieving the expected effect.

[0032] Furthermore, the number of nodes in the hidden layer of the BP neural network is determined by trial and error.

[0033] The trial and error method is used to determine the number of nodes in the hidden layer of the BP neural network, and the optimal solution is found by gradually increasing or decreasing the number of nodes and changing the error.

[0034] The number of neurons in the input layer is equal to the dimension of the input data, and the number of neurons in the output layer is also equal to the dimension of the output data.

[0035] Furthermore, the adjusting the model according to the error back propagation algorithm specifically includes adjusting the weights and biases of each layer in the model according to the error back propagation algorithm to reduce the error.

[0036] The core idea of ​​the error back propagation algorithm is to calculate the error between the actual output value and the expected output value, and propagate this error from the output layer to the input layer layer by layer, so as to adjust the weight matrix and bias vector between each layer, so as to make the network output as close to the actual value as possible; based on the chain rule, the partial derivative of the error with respect to the parameters is decomposed into a series of simple steps, starting from the output layer, and calculating the partial derivative of the error with respect to the parameters between each layer layer by layer, and finally reaching the input layer.

[0037] Furthermore, determining the appropriate amount of steel passing according to the slope data in the relationship curve specifically includes: finding a starting point on the curve with a positive slope and a maximum slope of the curve, and taking a point adjacent to this point as a point with an appropriate amount of steel passing.

[0038] The part of the curve with a positive slope and the maximum slope of the curve is the part where the wear of the rough rolling working roll increases sharply. In the absence of abnormal operation, it means that the workload of the working roll is too high, which significantly increases the wear rate. Therefore, the point where the amount of steel passing increases by 5% from the starting point of this part is selected as the appropriate point for the amount of steel passing, so as to ensure the quality of the working roll and support roll surface while taking into account the production rhythm as much as possible.

[0039] In the second aspect, the present application also provides a computing device, which has the function of implementing the method described in the first aspect above. The beneficial effects can be found in the description of the first aspect, which will not be repeated here. The function can be implemented by hardware, or by executing the corresponding software implementation through hardware. The hardware or software includes one or more modules corresponding to the above functions. In one possible design, the structure of the device includes an acquisition module, a training module, and optionally, a construction module. These modules can implement the function of the training node in the method example of the first aspect above. Please refer to the detailed description in the method example for details, which will not be repeated here.

[0040] In the third aspect, the present application also provides a computing device, which is used to implement the functions of the method described in the first aspect above. The beneficial effects can be found in the description of the first aspect and will not be repeated here. The structure of the computing device includes a processor and a memory, and the memory is used to store instructions and / or data. The memory is coupled to the processor, and when the processor executes the program instructions stored in the memory, the function of the training node in the example of the first aspect above can be implemented. The structure of the computing device also includes a communication interface for communicating with other devices.

[0041] In a fourth aspect, the present application also provides a computer-readable storage medium, in which instructions are stored, and when the computer-readable storage medium is run on a computer, the computer executes the method in the above-mentioned first aspect and various possible designs of the first aspect.

[0042] In a fifth aspect, the present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the method in the above-mentioned first aspect and various possible designs of the first aspect.

[0043] In a sixth aspect, the present application also provides a computing chip, which is connected to a memory, and the chip is used to read and execute a software program stored in the memory, and to execute the methods in the above-mentioned first aspect and various possible implementation methods of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present drawings or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present drawings. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0045] Figure 1 A flowchart of the steps of a method for determining the over-steel amount of a rough rolling work roll of a newly built hot rolling line;

[0046] Figure 2 This is a curve chart showing the relationship between the excess steel data and the predicted wear amount of the newly built hot rolling line.

[0047] The purpose, features and advantages of this figure will be further described in conjunction with the embodiments with reference to the accompanying drawings. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0049] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0050] If not otherwise specified, all embodiments and optional embodiments of the present application can be combined with each other to form a new technical solution.

[0051] Unless otherwise specified, all technical features and optional technical features of this application can be combined with each other to form a new technical solution.

[0052] If there is no special explanation, all steps of the present application can be performed sequentially or randomly, preferably sequentially. For example, the method includes steps (a) and (b), which means that the method may include steps (a) and (b) performed sequentially, or may include steps (b) and (a) performed sequentially. For example, the method may further include step (c), which means that step (c) may be added to the method in any order. For example, the method may include steps (a), (b) and (c), or may include steps (a), (c) and (b), or may include steps (c), (a) and (b), etc.

[0053] If there is no special explanation, the "include" and "comprising" mentioned in this application are open-ended or closed-ended. For example, the "include" and "comprising" may mean that other components not listed may also be included or only the listed components may be included or only the listed components may be included.

[0054] If not specifically stated, in this application, the term "or" is inclusive. For example, the phrase "A or B" means "A, B, or both A and B". More specifically, any of the following conditions satisfies the condition "A or B": A is true (or exists) and B is false (or does not exist); A is false (or does not exist) and B is true (or exists); or both A and B are true (or exist).

[0055] In order to better understand the solutions of the embodiments of the present application, some relevant terms and concepts that may be involved in the embodiments of the present application are first introduced below.

[0056] (1) Artificial intelligence (AI), also known as intelligent machinery or machine intelligence, refers to machines that are made by humans and can display intelligence. Generally speaking, AI refers to the technology that displays human intelligence through ordinary computer programs.

[0057] (2) Neural networks (NN) are mathematical models that imitate the behavioral characteristics of animal neural networks and perform distributed parallel information processing. The purpose of information processing can be achieved by adjusting the interconnected relationships between a large number of nodes within the neural network. Neural networks have the ability to self-learn and adapt. Neural networks are usually used in artificial intelligence model training and data derivation processing.

[0058] Specifically, a neural network may usually include multiple layers connected end to end, such as convolutional layers, fully connected layers (FC), activation layers or pooling layers.

[0059] (3) Back Propagation Neural Network (BPNN), or BP neural network, is a multi-layer feedforward neural network. Its main feature is to train the network through the back propagation algorithm, that is, to continuously adjust the network weights and thresholds so that the network output gradually approaches the expected value. The BP neural network consists of an input layer, a hidden layer (also called an intermediate layer), and an output layer. Each layer can have several nodes, and the connection status of the nodes between layers is reflected by weights. In the process of forward signal propagation, the input data is passed from the input layer to the hidden layer, and then from the hidden layer to the output layer, and finally the prediction result is obtained. In the process of back propagation of the error, the output error is calculated, and the error is back propagated back to the network layer by layer to update the weights and biases of each layer.

[0060] In this embodiment, Figure 1 As shown, a method for determining the amount of steel overload of a rough rolling work roll of a newly built hot rolling line comprises the following steps:

[0061] Step 1: Collect wear-related data of multiple hot rolling lines and steel overload data of rough rolling work rolls of each hot rolling line in multiple different online cycles;

[0062] The data that has a greater impact on the wear of the rough rolling work rolls on the hot rolling line are collected as input data for the work roll wear model to better predict the change of work roll wear on this hot rolling line with the increase of over-steel amount, so as to determine the appropriate over-steel amount.

[0063] Step 2: Collect the wear data of the rough rolling work rolls in the corresponding online cycle;

[0064] The actual rough rolling work roll wear data collected is used for model training and comparison of test results. The model is optimized based on the error between the model prediction results and the actual results to achieve better prediction results.

[0065] Step 3: Normalize the collected data and divide the collected data into training set and test set;

[0066] Data normalization is an important preprocessing technique in data analysis and machine learning. Its purpose is to convert the original data into a fixed range or make all data have the same scale, thereby eliminating the impact of dimension and order of magnitude differences and improving the training effect and generalization ability of the model.

[0067] The main function of the training set is to train the model and estimate the parameters in the model through training data. These parameters enable the model to reflect the data relationships in the real world and make predictions.

[0068] The purpose of the test set is to evaluate the performance of the model on unseen data, that is, the generalization ability of the model. The test set can be used to verify whether the model can continue to work effectively on new, unseen data. This helps to discover possible overfitting problems in the model, that is, the model performs well on the training data but poorly on new data.

[0069] Step 4: Establish a work roll wear model based on BP neural network, taking the data collected in step 1 as input and the work roll wear amount as predicted output;

[0070] Step 5: Train the work roll wear model with the training set data, compare the output value with the collected rough rolling work roll wear data, and adjust the model according to the error back propagation algorithm;

[0071] Step 6: Test the work roll wear model with the test set data until the error is less than the threshold or the maximum number of iterations is reached;

[0072] The error threshold is used to determine whether the model has converged to an acceptable level of accuracy. The maximum number of iterations refers to the maximum number of iterations allowed during the iteration process. When this number is reached, the algorithm will stop running regardless of whether the current error is small enough. This is mainly to control the running time and computing resources of the algorithm.

[0073] Step 7: Denormalize the output data and draw the relationship curve between the steel amount data of each hot rolling line and the wear amount of the working roll;

[0074] Denormalization is the process of restoring normalized data back to its original size. When normalizing, we usually scale the data using the minimum and maximum values, while denormalizing requires restoring the data to its original scale based on these parameters.

[0075] Step 8: Input the wear-related data and over-steel test data of the newly built hot rolling line into the working roll wear model, and determine the appropriate over-steel amount based on the slope data in the relationship curve.

[0076] Furthermore, the wear-related data include rolling force, total number of rolling passes, steel grade hardness, rolling speed, strip width, and strip length.

[0077] Furthermore, the normalization process uses the Min-Max normalization method to convert the original data into data within the range of 0 to 1.

[0078] The Min-Max standardization method scales the data to the interval [0, 1], where Min is the minimum value 0 and Max is the maximum value 1.

[0079] Furthermore, the collected data is divided into a training set and a test set, specifically including: taking the wear-related data of a hot rolling line, the steel overload data of a machine cycle and the corresponding working roll wear data as a complete data sample, and performing data division based on the data sample.

[0080] Furthermore, the proportion of the training set to the total data is 75%.

[0081] This ratio division ensures that there is enough data to train the model while retaining a portion of the data for evaluating the model's generalization ability.

[0082] Furthermore, the BP neural network includes an input layer, a hidden layer and an output layer.

[0083] The input layer is the starting point of the entire neural network. It is responsible for receiving external input data signals. Its main function is to pass the received signals to the next layer, the hidden layer, to provide basic data for subsequent processing. The hidden layer is located between the input layer and the output layer and is the core part of the neural network. Its main function is to perform nonlinear transformation and feature extraction on the input data to adapt to complex practical problems. The hidden layer performs nonlinear transformation by weighted summation of input signals and applies activation functions to improve the network's expressive power. The output layer is the final output end of the neural network and is responsible for generating the final prediction results or classification results.

[0084] Furthermore, the learning rate of the BP neural network is set to 0.1.

[0085] The learning rate is one of the important hyperparameters for training neural networks. It represents the step size of the gradient moving toward the optimal solution of the loss function in each iteration. It is usually represented by the symbol η, and its size determines the speed of network learning. A suitable learning rate can ensure stable learning of the network while reducing training time with reasonable high-speed convergence. If the learning rate is too high, it may cause oscillation or even divergence during training; if the learning rate is too low, it may lead to slow training speed and difficulty in achieving the expected effect.

[0086] Furthermore, the number of nodes in the hidden layer of the BP neural network is determined by trial and error.

[0087] The trial and error method is used to determine the number of nodes in the hidden layer of the BP neural network, and the optimal solution is found by gradually increasing or decreasing the number of nodes and changing the error.

[0088] The number of neurons in the input layer is equal to the dimension of the input data, and the number of neurons in the output layer is also equal to the dimension of the output data.

[0089] Furthermore, the model is adjusted according to the error back propagation algorithm, specifically including adjusting the weights and biases of each layer in the model according to the error back propagation algorithm to reduce the error.

[0090] The core idea of ​​the error back propagation algorithm is to calculate the error between the actual output value and the expected output value, and propagate this error from the output layer to the input layer layer by layer, so as to adjust the weight matrix and bias vector between each layer, so as to make the network output as close to the actual value as possible; based on the chain rule, the partial derivative of the error with respect to the parameters is decomposed into a series of simple steps, starting from the output layer, and calculating the partial derivative of the error with respect to the parameters between each layer layer by layer, and finally reaching the input layer.

[0091] Furthermore, determining the appropriate amount of steel passing according to the slope data in the relationship curve specifically includes: finding a starting point on the curve with a positive slope and a maximum slope of the curve, and taking a point adjacent to this point as a point with an appropriate amount of steel passing.

[0092] The part of the curve with a positive slope and the maximum slope of the curve is the part where the wear of the rough rolling working roll increases sharply. In the absence of abnormal operation, it means that the workload of the working roll is too high, which significantly increases the wear rate. Therefore, the point where the amount of steel passing increases by 5% from the starting point of this part is selected as the appropriate point for the amount of steel passing, so as to ensure the quality of the working roll and support roll surface while taking into account the production rhythm as much as possible.

[0093] In this embodiment, Figure 2 As shown in the figure, a curve is drawn showing the relationship between the over-steel amount data during the working roll machine cycle and the work roll radius wear. The points on the curve are the points corresponding to the different over-steel amount data and the predicted wear on the newly built hot rolling line. Therefore, the wear may decrease, which is a normal error in the model prediction process.

[0094] Depend on Figure 2 It can be seen that when the amount of steel overrun in the newly built hot rolling line is between 44,000 tons and 52,000 tons, the slope of the curve is positive and is the maximum slope part of the curve. Therefore, in this embodiment, the starting point of 44,000 tons and the amount of steel overrun of 46,200 tons increased by 5% are selected as the appropriate amount of steel overrun for the current newly built hot rolling line, that is, the appropriate amount of steel overrun for a single machine cycle of the roughing working roll. While ensuring that the roughing working roll has a certain amount of steel overrun and avoiding frequent replacement of the roughing working roll to affect the production rhythm, it also avoids excessive wear of the roughing working roll causing peeling accidents of the roughing support roll surface, thereby ensuring the safe use of the roughing support roll.

[0095] It should be noted that the present application is not limited to the above-mentioned embodiments. The above-mentioned embodiments are only examples, and the embodiments having the same structure as the technical idea and exerting the same effect within the scope of the technical solution of the present application are all included in the technical scope of the present application. In addition, without departing from the scope of the main purpose of the present application, various modifications that can be thought of by those skilled in the art to the embodiments and other methods of combining some of the constituent elements in the embodiments are also included in the scope of the present application.

Claims

1. A method for determining the amount of steel overload of the rough rolling work roll of a newly built hot rolling line. It is characterized in that The following steps are involved: Step 1: Collect wear-related data of multiple hot rolling lines and steel overload data of rough rolling work rolls of each hot rolling line in multiple different online cycles; Step 2: Collect the wear data of the rough rolling work rolls in the corresponding online cycle; Step 3: Normalize the collected data and divide the collected data into training set and test set; Step 4: Establish a work roll wear model based on BP neural network, taking the data collected in step 1 as input and the work roll wear amount as predicted output; Step 5: Train the work roll wear model with the training set data, compare the output value with the collected rough rolling work roll wear data, and adjust the model according to the error back propagation algorithm; Step 6: Test the work roll wear model with the test set data until the error is less than the threshold or the maximum number of iterations is reached; Step 7: Denormalize the output data and draw the relationship curve between the steel amount data of each hot rolling line and the wear amount of the working roll; Step 8: Input the wear-related data and over-steel test data of the newly built hot rolling line into the working roll wear model, and determine the appropriate over-steel amount based on the slope data in the relationship curve.

2. According to the method for determining the amount of steel overrun of the rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that The wear-related data include rolling force, total number of rolling passes, steel grade hardness, rolling speed, strip width, and strip length.

3. According to the method for determining the amount of steel overrun of the rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that The normalization process uses the Min-Max standardization method to convert the original data into data within the range of 0 to 1.

4. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that The collected data is divided into a training set and a test set, specifically including: taking the wear-related data of a hot rolling line, the steel passing data of a machine cycle and the corresponding working roll wear data as a complete data sample, and performing data division based on the data sample.

5. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 4, It is characterized in that The training set accounts for 70% to 80% of the total data.

6. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that The BP neural network includes an input layer, a hidden layer and an output layer.

7. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 6, It is characterized in that The learning rate of the BP neural network is set to 0.

1.

8. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 6, It is characterized in that The number of nodes in the hidden layer of the BP neural network is determined by trial and error.

9. A method for determining the amount of steel overrun of a rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that The adjusting the model according to the error back propagation algorithm specifically includes adjusting the weights and biases of each layer in the model according to the error back propagation algorithm to reduce the error.

10. A method for determining the amount of steel overrun in a rough rolling work roll of a newly built hot rolling line according to claim 1, It is characterized in that Determining the appropriate amount of steel passing according to the slope data in the relationship curve specifically includes: finding a starting point on the curve with a positive slope and a maximum slope of the curve, and taking a point adjacent to this point as a point with an appropriate amount of steel passing.

Citation Information

Patent Citations

  • Wear compensation and self-adaption method for hot continuous rolling backup roller

    CN104942020A

  • Steel plate roller abrasion loss prediction method and system

    CN112037209A

  • Rolling process roller wear prediction method based on data driving

    CN117875137A

  • Roughly-rolled intermediate billet convexity prediction method based on actually-measured wear of working roll

    CN118045872A

  • Reconstruction method of plate and strip continuous casting and rolling production line

    CN118080566A