论文部分内容阅读
提出了一种基于径向基函数的非线性过程预测控制策略。首先,开发过程径基函数网络模型:根据过程特性,选择模型阶次、径基函数类型,用K均值法确定基函数中心位置,统计F检验确定基函数中心的数目,迭代最小二乘法确定径基函数网络权系数。然后,利用网络模型抽取非线性预测控制器(NLPC)特征样本训练构造径基函数网络预测控制器(RBFPC)。仿真结果表明,与NLPC比较,由于RBFPC不必在线解非线性最优化问题,易于在线快速实施;与PI控制器比较,RBFPC具有更好的跟踪设定值性能和抗干扰性能。
A nonlinear process predictive control strategy based on radial basis function is proposed. Firstly, we develop the RBF network model: According to the process characteristics, we select the model order and the type of RBF. The K-means method is used to determine the center of the basis function. The F-test is used to determine the number of basis functions and the iterative least- Basis function network weight coefficient. Then, the network model is used to extract the non-linear predictive controller (NLPC) feature samples to train RBFPC. Simulation results show that compared with NLPC, RBFPC is easy to implement online quickly because it does not need to solve the nonlinear optimization problem online. Compared with PI controller, RBFPC has better tracking setpoint performance and anti-jamming performance.