针对双辊铸轧过程中凝固终点位置这一关键参数,基于贝叶斯方法的神经网络和理论模型,根据经验模型及熔池断面几何关系建立凝固终点位置数学模型.在化学成分和工艺约束已知的条件下,采用粒子群优化算法针对凝固终点位置这一铸轧过程中的关键因素进行相应的工艺参数的优化计算.铸轧实验结果验证了优化结果的可行性,从而为提高双辊铸轧板形和板厚的控制精度,改善铸带表面质量提供可能.
In order to improve the flat,gauge control precision and surface quality of casting strips,processing parameters were optimized based on a kiss point position calculation model(combining a Bayesian neural network and a theoretical model) and using a particle swarm optimization method under given chemical composition and process constraint conditions.Optimized and experimental results showed good agreement.