Study On High Temperature Wear Behavior Of Hot Forging Die And Wear Model Modifying Based On Digital Recognition Of Wear Debris

Wear is the main type of failure for hot forging dies.The key to improve the lifespan of hot forging die is to improve the wear resistance.Wear is a very complex dynamic process for die steel in hot forging.The present study on the wear behavior and mechanism of die steel is insufficient,especially the research on the mechanism of the high temperature wear is not deep enough.At the same time,it is an urgent problem to establish an accurate prediction model of wear failure.This paper studied the friction and wear behavior of typical mold material H13 steel under different working conditions,and the morphologies of worn surface and wear debris were observed by SEM.The results indicated that the main wear mechanism of H13 steel is abrasive wear,oxidative wear and adhesive wear mechanism,and the three main wear mechanisms coexist during the wear process,Therefore,it is essential to couple three wear models to establish a new wear model.At the same time,Further discussion on high-temperature oxidative wear mechanism were conducted,the results showed that oxide film and spalling pit is the main worn surface morphology during the high-temperature,and the oxygen content of oxide particles is higher than the whole worn surface;The oxide film on worn s urface exists in two forms,oxide film on matrix surface and multilayer oxide film inside the matrix;The spalling types of oxide film is mainly the cracking at the interface,in the film and in the matrix.The former two wear rate is lower than the last.The XRD results show that the maximum diffraction peak of the worn surface is a-Fe at different temperatures,and with the increasing of temperature,the Fe2O3 phase and Fe3O4 phase gradually appear on the worn surface.Wear debris image is an important information carrier to reflect the wear condition.Digital recognition method based on morphology characteristics of debris was adopted to recognize the three types of wear debris by MATLAB software.The wear debris classifier based on BP neural network is designed,and the training and testing results show that the classifier has good classification ability.The proportion of three kinds of wear debris under different working conditions is calculated by BP neural network.A new wear prediction model was established considering the transition of wear mechanisms.Wear tests were performed to verify the wear prediction accuracy,which shows a favorable agreement with the prediction values.