将电弧炉电极调节系统表示为线性部分时变的N—L—N模型,在此基础上提出了基于时变参数N—L—N系统的在线辨识方法,解决了因电弧的时变性而导致模型变化、预测精度下降的问题.针对在线辫识过程中出现的参数分离的难题,该方法采用依次对模型线性部分、非线性部分参数进行更新的思路,实现了系统参数的有效分离。仿真结果表明,该方法能够有效跟踪电弧特性的变化,达到实时监测系统参数的目的。
Based on the expression of electrode regulalor system as a time-variant N-L-N model, an online method is proposed to identify the electrode regulator system of electric arc furnace, which improves the model precision as time-varying of arc characteristic. To deal with the difficulty on parameter separation during the process of on-line identification, the idea that updating parameters of linear and nonlinear parts respectively is derived. Simulation results validate the effectiveness of the proposed algorithm on tracking the parameters of the model which achieves the purpose of monitoring system parameters real-time.