The Collapse Deformation Prediction Model of Wide 7075 Al-Alloy Intermediate Slabs Based on Particle Swarm Optimization and Support Vector Regression During the Hot Rolling Process

Aiming at the collapse deformation problem of ultra-wide Al-alloy intermediate slabs on the finishing mill run-in table during the hot rolling process, the collapse deformation prediction model of ultra-wide 7075 Al-alloy intermediate slabs based on particle swarm optimization and support vector regression (PSO-SVR) was proposed. The 7075 Aluminum alloy (Al-alloy) constitutive relationship in a temperature range of 330-370 degrees C was established by the Gleeble thermal compression experiment. According to the actual transporting process, the thermomechanical coupling finite element model (FEM) of 7075 Al-alloy intermediate slabs was established. It was demonstrated that width, temperature, and transportation time were positively related to contact length, while thickness was negatively related to contact length. The PSO-SVR collapse deformation prediction model was proposed by combining simulation data and the PSO-SVR algorithm. The PSO-SVR prediction model constructed was applied and verified by production test in the "1 + 3" hot rolling production line of an Al-alloy production line. The results show that the prediction model can achieve 100% prediction accuracy to predict whether the slab can be produced normally, and the maximum error of the contact length was 10.5%, it can provide guidance for the high-quality and stable production of ultra-wide Al-alloy.