Research On Deep Optimization Of Rolling Process Parameters Based On Rolling Mill Vibration

With the application of plate and strip in the field of high-tech engineering,the requirements for the quality of plate and strip products are also increasing,which requires the rolling mill,the key equipment for the production of plate and strip,to have higher operational stability.Rolling mill vibration is the most intuitive manifestation of unstable operation of the rolling mill,which can cause vibration marks on the surface of the strip,and in severe cases,steel-heaping,roll bursts,etc.,which reduce the quality stability and production efficiency of rolled products.Therefore,it is of great significance for the research of rolling mill vibration.Traditional rolling process optimization methods are generally based on minimum energy consumption,without considering the impact of rolling mill operation stability.This paper comprehensively considers the complex relationship between the rolling process force parameters and the rolling mill system vibration mechanism.Aiming at the problem of low matching degree between F2 rolling mill equipment status and rolling process of a 1580 mm hot strip mill in a factory,multi-point vibration testing and process parameter collection were carried out on the F2 rolling mill,and the key process parameters were optimized.The specific research content is as follows:Firstly,with the F2 rolling mill of the 1580 mm hot tandem mill as the research object,the research team collected the mill transmission system,rack,bearing seat and other positional vibration data,and extracted the real-time matching process parameter data.This paper Select the data of the F2 rolling mill with more obvious vibrations for analysis.In order to improve the accuracy of the optimization model,process parameter data and vibration data are integrated and then processed for outliers,clustering and normalization.Secondly,based on the BP neural network,RBF neural network,and Kriging model,the relationship model between the key process parameters of the rolling mill and the vibration data is established.By comparing the prediction performance of each model,it is found that the accuracy of the RBF prediction model and the Kriging prediction model is higher than that of the BP prediction model.But the prediction accuracy is still low.Therefore,the RBF-Kriging integrated model based on differential evolution algorithm is proposed,and the prediction effect is relatively ideal.Taking the RBF-Kriging integrated model as the fitness function,and taking the vibration of the three measuring points of the rolling mill as the optimization objective,the NSGA? algorithm is used to optimize the three vibration objectives,and the Pareto optimal solution set is obtained.On this basis,the minimum sum of the vibration values of the three optimization objectives is used as the measurement standard,and a set of solutions is obtained as the optimal process parameters.Finally,considering that the optimized new process cannot be debugged and verified in the actual production process of the F2 rolling mill of the 1580 mm hot tandem mill,it is also difficult to find a Chinese platform with a matching production capacity to test the new process.Based on the system dynamics mechanism of the four-high rolling mill,this paper establishes a multi-body system coupling dynamics simulation model of the strip rolling mill to compare the process parameters before and after optimization.The results show that the optimized process parameters can effectively reduce the vibration of the rolling mill.Based on the theory of machine learning,this paper conducts data mining on the key process parameters and vibration data of the rolling mill,proposes a deep optimization method of rolling process parameters based on the vibration of the rolling mill,and validates the optimization results by simulation.It provides theoretical guidance for the deep optimization of the process parameters of the strip mill.