基于增量分析和PSO-LSSVM的热轧辊缝调平预测模型

辊缝调平在保证热轧带钢板形质量和轧制稳定性中起着关键作用,目前以操作人员目测后经验调整为主,无法满足未来少人化、智能化轧制技术需求。为此,基于增量分析方法实现工艺参数增量因子提取,有效解决传统离散数据预测中部分信息丢失问题;同时以粒子群算法(PSO)优化最小二乘支持向量机(LSSVM)模型参数,使得参数选取更加科学。采用某钢厂1580热轧生产数据进行验证,结果表明,基于增量分析和PSO-LSSVM的预测模型能够较好地预测调平值及调平曲线趋势,精轧下游F5~F7调平预测精度在95%左右,可为现场调平策略设定提供辅助手段,也为今后无人化轧制技术的发展提供关键理论支撑。

Prediction model of hot rolling gap leveling based on incremental analysis and PSO-LSSVM

Roll gap leveling plays a key role in shape control and rolling stability of hot-rolled strip. At present, the operator's experience adjustment after visual inspection is the main method, which cannot meet the demand of less human and intelligent rolling technology in the future. Incremental factor extraction of process parameters is realized based on incremental analysis method, which effectively solves the problem of partial information loss in traditional discrete data prediction. At the same time, the particle swarm optimization algorithm is used to optimize the parameters of LSSVM model, which makes the parameter selection more scientific. The production data of a 1 580 mm hot rolling line are used for verification,results show that the prediction model based on incremental analysis and PSO-LSSVM can better predict the leveling value and leveling curve trend, and the F5-F7 leveling prediction accuracy of downstream in finishing rolling is about 95%, which provides assistance for setting of leveling strategy and provides key theoretical support for the development of unmanned rolling technology in the future.