Prediction Of Mechanical Properties Of High Co-Ni Secondary Hardened Steel Materials Based On I-ELM

Steel is an essential core material for social development.High Co-Ni secondary hardening steel is a new type of martensitic steel,it has high strength and toughness,and it is currently mainly used in aerospace equipment such as landing gear of aircraft.Nowadays,high Co-Ni secondary hardening steel needs to have higher performance in a more severe environment,so the composition and preparation process of high Co-Ni secondary hardening steel are becoming more and more complicated.Relying on traditional experimental methods for research,the experimental period is long and requires a lot of resources.At present,it has been unable to meet the industrial research and development needs of steel materials.Therefore,it is necessary to use computer-aided material experiments and performance analysis.At present,BP neural network is mainly used in the performance analysis and prediction of high Co-Ni secondary hardened steel.However,the hot processing of high Co-Ni secondary hardened steel has many parameters,and the amount of experimental data is small.Therefore,the model training with BP neural network has the problems of slow learning speed,generalized performance,and low accuracy.Extreme learning machine(ELM)is a new single hidden layer feedforward neural network algorithm,which has the advantages of fast learning speed,high model accuracy,and good generalization ability.It has been developed in many fields in recent years.Based on the incremental extreme learning machine(I-ELM)algorithm,the research learned the heat treatment process of high Co-Ni secondary hardened steel by training neural network.and trained the mathematics between trace elements,heat treatment process and alloy mechanical properties during heat treatment model.This paper compared the model based on I-ELM and BP neural network from multiple perspectives.The results show that the I-ELM model is superior to the BP neural network in speed,accuracy and generalization.Then,Based on the I-ELM model,this paper predicts the influence of trace elements C,Co,and aging temperature on the mechanical properties of high Co-Ni secondary hardened steel.Combined with the theoretical knowledge of microstructure of materials,the reliability of the prediction results is proved.The research in this paper has obtained high-precision experimental results.The conclusion proved that the extreme learning machine has a broad research space in the field of high Co-Ni secondary hardened steel.And the conclusion provided a useful reference for using ANN to study material systems with small sample data and complex mapping relationships.