Research On Energy Prediction In The Aluminum Extrusion Process System Based On Data-Driven

The aluminum extrusion process is widely used in aerospace,rail transportation and other fields due to its high efficiency,no chipping,and high material utilization.However,the energy consumption cost of the hot extrusion forming process is extremely high.In order to cope with rising prices of energy and increasingly strict environmental policies,etc.,improving the energy efficiency of our country's aluminum processing industry has become an inevitable trend in aluminum processing and manufacturing.Hot extrusion is a complex thermal-mechanical coupling process,there are many parameters that affect energy consumption.It is difficult to construct a mechanism model for energy consumption in extrusion process.In most cases,the process parameters are adjusted based on experience.However,it may cause unnecessary energy consumption in the process by adjusting with experience.This thesis will combine the existing mechanism analysis,based on the data drive to study the process parameters such as the temperature change characteristics of the die bar and the extrusion speed process parameters during the aluminum profile extrusion production process,and establish the aluminum profile extrusion based on the data-driven model for predicting energy consumption during forming.This can not only avoid the inaccuracy caused by the simplified process due to the mechanism analysis alone,but also avoid the problem of difficulty in optimizing energy consumption due to the fact that the process is completely driven by data and the process is regarded as a black box function.The main work of this thesis is as follows:(1)Combined with the analysis of the energy consumption mechanism of the aluminum extrusion forming process,the key factors and process parameters affecting the energy consumption of the extrusion forming are determined.The main process parameters that specifically affect the energy consumption of the aluminum extrusion forming process include the extrusion speed and the preheating temperature of the blank,Extrusion barrel temperature,blank length,die angle,etc.,and use these process parameters as the input parameters of the energy consumption prediction model.(2)A prediction model of energy consumption in extrusion of aluminum profiles with multiple models was proposed.First,Bagging integrated learning algorithm is used to optimize the data set,and then sub-models of energy consumption prediction for aluminum extrusion forming are established based on extreme learning machine(ELM)and Gaussian process regression(GPR)respectively.Finally,entropy weight method is used to weighted the two sub-models.The effectiveness of the proposed multi-model fusion modeling method is verified.(3)Experimental data are used to verify the multi-model fused aluminum extrusion energy consumption prediction model and the data set optimization method based on Bagging integrated learning algorithm.The experiment proves the feasibility of selecting five aluminum extrusion parameters in this paper,and the multi-model fusion modeling method has a good prediction accuracy for aluminum extrusion energy consumption.