基于Hammerstein模型的非线性自适应预测函数控制

Nonlinear adaptive predictive functional control using Hammerstein models

传统的预测函数控制对非线性和多变性的实际对象控制效果不佳。采用基于Hammerstein模型的预测函数控制(PFC)内部模型辨识方法来拟合实际对象的非线性。针对实际对象的多变性,不断辨识对象,校正内部模型参数,增强控制系统适应对象变化的能力。选取正弦多项式形式的基函数并介绍了滚动优化的办法。优化性能指标选取双值指标,计算量小,精度高。整个算法线性寻优,不需要非线性模型预测控制的非线性滚动优化,输出可以应用模型参数直接计算。实验表明该方法比PID控制效果更好而且对模型失配具有很好的自适应性能。

The common predictive function control （PFC） can＇t control the nonlinear and changeful object effectively. An inner model of PFC based on Hammerstein model is applied to fit the nonlinear character of actual objects. To fit the changeful character of actual objects, the inner model parameters are continuously adjus- ted, and the ability of the system is strengthened to control the changeful objects. The sine polynomial basis function is chosen for PFC and the way of optimization is presented. Two series of numbers are chosen for the optimization index sign, which the calculation quantity is small but the accuracy is high. The whole algorithm only needs linear optimizations and does not need nonlinear optimization. The output is calculated by model parameters. The simulation results of pH neutralization process show that this method is more effective than PID and is robust to model mismatch.