电弧炉终点温度是炼钢过程中的重要指标之一,决定了钢水的质量和整体成本.电弧炉终点温度预报模型的建立是实现炼钢自动化的重要环节,为了得到高精度的终点温度预报值,提出了一种GASVR_GM的钢水终点温度预报模型.该模型以定量因素为主,采用遗传算法优化的支持向量回归机预报终点温度；再运用灰色模型进行预报误差的补偿,解决非定量因素的影响,实现滚动预报.试验仿真表明,与智能软测量方法相比,GASVR_GM预报模型具有更高的精度和鲁棒性.%The endpoint temperature of electric arc furnace ( EAF) is one of the most important indexes of the steelmaking process, it determines both the quality of molten steel and the overall cost. The prediction model establishment of the endpoint temperature is an important part for implementing automation of steelmaking. In order to obtain highly accurate predictive value of the endpoint temperature, the GASVR_ GM prediction model is proposed. This model is mainly based on quantitative factors, it predicts the endpoint temperature by adopting the support vector regression machine optimized with genetic algorithm, then compensates the predictive error by using grey model to solve the influence from non-quantitative factors for implementing rolling prediction. The experimental simulation indicates that comparing with intelligent soft sensing method, the GASVR-GM prediction model possesses higher accuracy and robustness.
展开▼
