Multi-gain coefficient switching method for strip steel deviation rectification and edge alignment system

Abstract

The invention belongs to the field of metallurgical industry automatic control, and provides a multi-gain coefficient switching method for a strip steel deviation rectification and edge alignment system, which comprises the following steps: by introducing a multi-gain coefficient library and an automatic switching mechanism, analyzing acquired data to determine gain coefficients, and selecting a deviation rectification coefficient and a control algorithm in the multi-gain coefficient library, so as to realize the automatic switching of the deviation rectification and edge alignment system. Meanwhile, a machine learning model technology is adopted, in the whole production process, data collection and analysis are continuously carried out, and correction fitting and continuous optimization are carried out on a gain coefficient, so that the strip steel deviation correction and edge alignment system can flexibly adapt to different incoming material deviation conditions, the deviation correction precision and the production efficiency are improved, the production cost is reduced, and the production efficiency is improved. The method has wide application prospects and remarkable economic benefits.

Description

Technical Field

[0001] The invention belongs to the field of automatic control of metallurgical industry, and relates to a strip steel deviation correction and edge alignment system applied to a coiled plate and strip production line of a metallurgical enterprise, and specifically to a multi-gain coefficient switching method of the strip steel deviation correction and edge alignment system. Background Art

[0002] In the prior art, the strip steel deviation correction and edge alignment system of the strip steel production line usually adopts a single gain coefficient deviation correction control algorithm. However, in the face of different incoming material conditions, the adaptability of the single gain coefficient algorithm is limited, which often leads to unsatisfactory deviation correction effects and even repeated adjustments, affecting the production rhythm and product quality. In particular, with the continuous improvement of modern industry's requirements for the quality of strip steel products and the need to improve production efficiency, the traditional single gain coefficient deviation correction control method has become increasingly insufficient. Specifically, in the actual operation of the strip steel production line, the strip steel often deviates due to different factors, such as different coil states of raw materials, differences in specifications and varieties, and changes in the coil diameter of the strip coil on the reel. However, the traditional deviation correction control system cannot adapt to these changes in a timely manner, and often requires manual intervention, which increases the complexity of operation and production costs, and reduces production efficiency. Summary of the invention

[0003] The present invention aims to solve the problems of poor adaptability and unsatisfactory correction effect of the single correction control algorithm in the prior art. Its purpose is to provide a method for switching multiple gain coefficients of the strip correction edge system. Through the automatic switching of multiple gain coefficients, it can adapt to different deviation conditions of incoming materials, thereby achieving more efficient and stable correction control and improving the overall performance of the strip production line.

[0004] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: A method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system comprises the following steps: S1. Establish a multi-gain coefficient library; Through data collection and analysis, the relationship between gain control coefficient and production parameters is fitted; S2. Setting an edge data acquisition device, setting an edge detector in front of the coiler to detect edge data; the detection data signal output end of the edge detector is connected to the gain control system; S3, the gain control system reads the steel coil information from the production line L1 system through high-speed Ethernet. The steel coil information includes width, thickness, and coil diameter, and the gain control coefficient is obtained by calculating the edge deviation; S4, the gain control coefficient enters the multi-gain coefficient library, and according to the pre-set multi-gain coefficient library, the correction coefficient and control algorithm suitable for the current situation are selected, provided to the correction control system, and switched in real time; S5. The correction control system adjusts the action of the correction mechanism according to the instructions, and continuously adjusts the correction force through the feedback mechanism to ensure that the strip remains on the set track.

[0005] Furthermore, the multi-gain coefficient library adopts machine learning model technology to continuously collect and analyze data throughout the entire production process, correct and fit the gain coefficients, and continuously optimize them.

[0006] Furthermore, the machine learning model adopts variational autoencoders, deep neural networks and a hybrid algorithm of the two, and continuously optimizes the multi-gain coefficient library and control algorithm through analysis of historical data and actual production conditions.

[0007] Furthermore, the data correction hybrid algorithm of the multi-gain database is a deep neural network that takes the residuals of the variational autoencoder as the target, learns the nonlinear parts of these residuals through training, and finally corrects the data through special weighted mixing, continuously optimizing the multi-gain coefficient library and control algorithm.

[0008] Furthermore, the correction control system includes a sensor data acquisition unit, a control algorithm execution unit and a correction mechanism. The correction control system can perform feedback control, collect real-time position information based on the sensor data acquisition unit, respond quickly through the control algorithm, and adjust the correction action through the execution unit and the correction mechanism to ensure the stability and accuracy of the correction effect.

[0009] The beneficial effects of the present invention are: 1. Improve the accuracy of correction: Through the automatic switching of multiple gain coefficients, the correction control is more flexible and accurate, which can adapt to different deviation conditions of incoming materials and reduce quality problems caused by deviation. The improvement of correction accuracy can not only improve the quality consistency of products, but also reduce the scrap rate caused by deviation and improve production efficiency. High-precision correction control can significantly improve the dimensional accuracy and surface quality of strip steel products and meet customers' demand for high-quality strip steel products; 2. Improve production efficiency: reduce downtime and adjustment time caused by deviation, and improve the continuous production capacity of the strip production line. The automated deviation correction control system can greatly reduce manual intervention and shorten the adjustment time of the production line, thereby improving production efficiency. The improvement of production efficiency can not only increase the production capacity of the production line, but also reduce production losses caused by downtime and adjustment, and improve the economic benefits of the production line; 3. Reduce production costs: Reduce material waste and equipment loss caused by deviation, and reduce production costs. A stable correction control system can reduce equipment wear and maintenance costs, while reducing the waste of raw materials and reducing production costs. By reducing material waste and equipment loss, the operating costs of the production line can be significantly reduced, and the economic benefits of the enterprise can be improved; 4. Enhance system reliability: Through multiple gain coefficient libraries and automatic switching mechanisms, the system can maintain stable operation under different production conditions, enhancing the reliability and adaptability of the system. Ensure the continuity and stability of the production process and reduce downtime losses caused by equipment failure. The highly reliable correction control system can significantly improve the operational stability of the production line and reduce production interruptions and losses caused by equipment failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a schematic diagram of establishing, feeding back and optimizing a multi-gain coefficient library in the method of the present invention. DETAILED DESCRIPTION

[0011] The method of the present invention is further described below in conjunction with embodiments: Example

[0012] The method of the present invention is applied to the 2250 re-coiling unit of the hot rolling mill, and is aimed at the re-coiling production of 430 stainless steel before annealing. The method for switching the multi-gain coefficients of the strip steel deviation correction and edge alignment system comprises the following steps: S1. Establish a multi-gain coefficient library; Through data collection and analysis, the relationship between gain control coefficient and production parameters is fitted; S2. An edge detector is set in front of the coiler to detect edge data; the detection data signal output end of the edge detector is connected to the gain control system; S3, the gain control system reads the steel coil information from the production line L1 system through high-speed Ethernet. The steel coil information includes width, thickness, and coil diameter, and the gain control coefficient is obtained by calculating the edge deviation; S4. The gain control coefficient enters the multi-gain coefficient library. According to the pre-set multi-gain coefficient library, the correction coefficient and control algorithm suitable for the current situation are selected and provided to the correction control system, and switched in real time. The selection algorithm here includes data analysis and parameter optimization, which can dynamically adjust the correction parameters according to real-time data to ensure the stability and accuracy of the correction effect; S5. The correction control system adjusts the action of the correction mechanism according to the instructions, and continuously adjusts the correction force through the feedback mechanism to ensure that the strip steel remains on the set track. The correction control system includes a sensor data acquisition unit, a control algorithm execution unit and a correction mechanism. The correction control system can perform feedback control, collect real-time position information according to the sensor data acquisition unit, respond quickly through the control algorithm, and adjust the correction action through the execution unit and the correction mechanism to ensure the stability and accuracy of the correction effect; the feedback mechanism of the correction control system adopts an advanced control algorithm, which can quickly respond and adjust the correction action according to the real-time position information.

[0013] The multi-gain coefficient library uses machine learning model technology to continuously collect and analyze data throughout the production process, correct and fit the gain coefficients, and continuously optimize them; Furthermore, the machine learning model uses a variational autoencoder, a deep neural network, and a hybrid algorithm of the two to continuously optimize the multi-gain coefficient library and the control algorithm through analysis of historical data and actual production conditions; The deep neural network is specifically used to fit the relationship between the gain coefficient and the production parameters (diameter x1, width x2, speed x3). The formula used is: (1) Input layer to hidden layer h h=f(w1*x+b1) x=[x1,x2,x3], forming a vector W1 is the weight matrix from the input layer to the hidden layer, with a dimension of h*3 (assuming there are h hidden neurons).

[0014] b1 is the bias vector of the hidden layer.

[0015] f is the activation function.

[0016] (2) From hidden layer to output layer, that is, gain coefficient y Y=w2*h+b2 W2 is the weight matrix from the hidden layer to the output layer, with a dimension of 1×h.

[0017] b2 is the bias vector of the output layer.

[0018] The data correction hybrid algorithm of the multi-gain database takes the residual of the variational autoencoder as the target for the deep neural network, learns the nonlinear part of these residuals through training, and finally corrects the data through special weighted mixing.

[0019] The method of the present invention makes the deviation correction control more flexible and accurate through the automatic switching of multiple gain coefficients, can adapt to different deviation situations of incoming materials, and reduce quality problems caused by deviation. Improve the dimensional accuracy and surface quality of the strip steel to meet customer demand for high-quality strip steel products; reduce downtime and adjustment time caused by deviation, and improve the continuous production capacity of the strip steel production line. Through the automated deviation correction control system, reduce manual intervention, improve the operating efficiency of the production line, and reduce production costs. Through the multiple gain coefficient library and automatic switching mechanism, the system can maintain stable operation under different production conditions, enhancing the reliability and adaptability of the system. Ensure the continuity and stability of the production process and reduce downtime losses caused by equipment failure; advanced data processing and automatic control technology improve the intelligence level of the strip steel production line, which is in line with the development trend of modern industrial automation. By introducing artificial intelligence and machine learning technology, the deviation correction control algorithm is continuously optimized to improve the system's adaptability and intelligence level.

Claims

1. A method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system, characterized in that: The steps include: S1. Establish a multi-gain coefficient library, and fit the relationship between the gain control coefficient and the production parameters through data collection and analysis; S2. Setting an edge data acquisition device, setting an edge detector in front of the coiler to detect edge data; the detection data signal output end of the edge detector is connected to the gain control system; S3, the gain control system reads the steel coil information from the production line L1 system through high-speed Ethernet. The steel coil information includes width, thickness, and coil diameter, and the gain control coefficient is obtained by calculating the edge deviation; S4, the gain control coefficient enters the multi-gain coefficient library, and according to the pre-set multi-gain coefficient library, the correction coefficient and control algorithm suitable for the current situation are selected, provided to the correction control system, and switched in real time; S5. The correction control system adjusts the action of the correction mechanism according to the instructions, and continuously adjusts the correction force through the feedback mechanism to ensure that the strip remains on the set track.

2. The method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system according to claim 1 is characterized in that: The multi-gain coefficient library adopts machine learning model technology to continuously collect and analyze data during the entire production process, correct and fit the gain coefficients, and continuously optimize them.

3. The method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system according to claim 2 is characterized in that: The machine learning model uses variational autoencoders, deep neural networks, and a hybrid algorithm of the two, and continuously optimizes the multi-gain coefficient library and control algorithm through analysis of historical data and actual production conditions.

4. The method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system according to claim 3 is characterized in that: The data correction hybrid algorithm of the multi-gain database is that the deep neural network takes the residual of the variational autoencoder as the target, learns the nonlinear part of these residuals through training, and finally corrects the data through special weighted mixing, continuously optimizing the multi-gain coefficient library and control algorithm.

5. The method for switching multiple gain coefficients of a strip steel deviation correction and edge alignment system according to claim 1 is characterized in that: The correction control system includes a sensor data acquisition unit, a control algorithm execution unit and a correction mechanism. The correction control system can perform feedback control, collect real-time position information based on the sensor data acquisition unit, respond quickly through the control algorithm, and adjust the correction action through the execution unit and the correction mechanism to ensure the correction effect.

Images