层流冷却过程混合智能参数辨识

Hybrid intelligent parameter identification of the laminar cooling process

热轧层流冷却过程带钢温度难以连续检测,其换热系数具有随工况频繁变化而变化、动态特性具有时变、强非线性等综合复杂特性.正确辨识热轧层流冷却过程带钢温度的离散动态模型中的换热系数是提高模型精度的关键.本文将案例推理技术和神经网络技术相结合,提出了混合智能参数辨识方法.采用某钢铁公司热轧层流冷却过程实际运行数据对所提出的方法进行实验研究.结果表明本文提出的混合智能参数辨识方法大大提高了层流冷却过程带钢温度预报精度.

In a heat-rolling laminar cooling process, it is difficult to continuously measure the strip temperature online. The heat transfer parameters are subjected to changes due to the varying operating conditions, with time-varying and nonlinear characteristics. Its correct identification is the key factor in the determination of the discrete dynamic model for the strip temperature during the laminar cooling process. A hybrid intelligent parameter identification algorithm is developed by combining the （RBF） neural networks and case-based reasoning. Tests using real industrial data in a steel plant show that the hybrid intelligent parameter identification approach contributes great precision improvement in the prediction of the strip temperature during the laminar cooling process.