HOT DEFORMATION BEHAVIOUR AND PROCESSING MAPS OF HIGH CARBON LOW ALLOY STEEL

Abstract

Manufacturers always try to reduce the cost of raw material and production without the loss of quality. Thermomechanical processing and deformation study at elevated temperatures can control the process parameters which results in improved mechanical properties of material. Review of literature shows that there is a need to perform more hot deformation studies in widest range of processing parameters for high carbon low alloy steels which find large number of applications in automobile industry, springs used for shock absorber, tyre cords, bolts, ball bearings, bridge cables, piano wire etc. In the present study hot deformation behaviour of high carbon low alloy steel has been studied using thermomechanical simulator Gleeble® 3800.The tests were carried out in temperature range from 850°C-1100°C and strain rate 0.01 s-1 to 10 s-1. Constitutive analysis is done to model the flow behaviour by calculating Zener Hollomon parameter and flow stress values are predicted. Microstructure evolution after the hot deformation at different temperatures and strain rate is studied. Processing maps are drawn based on different models, flow localization parameters and constitutive equations. Different processing maps based on different instability criteria were plotted and compared to find the optimum workability region. A multilayer Artificial Neural Network (ANN) with feed forward back propagation and having ten neurons in hidden layer is trained to predict the flow stress. A comparative study is done between constitutive equations modelling and ANN modelling. Further, initiation and evolution of dynamic recrystallization is studied from the experimentally obtained flow curves. Strain for onset of dynamic recrystallisation i.e. critical strain was derived from strain hardening rate vs stress curve. The variation of critical strain with peak strain was established. The extent of dynamic recrystallisation was found using the Avrami model and it was found that this model shows good agreement with experimental data.