论文部分内容阅读
为了对预测控制中的模型失配问题进行性能评价与监视,基于互相关分析理论,通过引入操纵变量激励信号,分析其与预测偏差之间的互相关性来确定传递函数矩阵的失配问题,然后将模型失配问题转化为互相关系数在置信区间上的分布问题。结合各个通道互相关函数仿真图来直观监视模型失配与否,从而对其进行性能评价。将此方法成功运用到Wood-Berry塔模型上,分析了基于DMC的预测控制系统各通道模型失配情况。仿真结果表明,该方案对预测控制系统模型失配的性能分析具有一定的有效性。
In order to evaluate and monitor the model mismatch in the predictive control, based on the theory of cross-correlation analysis, the mismatch of the transfer function matrix is determined by introducing the manipulated variable excitation signal and analyzing the cross-correlation between it and the prediction error. Then the problem of model mismatch is transformed into the distribution of cross-correlation coefficient in the confidence interval. Combined with each channel cross-correlation function simulation chart to visually monitor the model mismatch or not, so as to evaluate its performance. The method is successfully applied to the Wood-Berry tower model, and the mismatch of each channel model in predictive control system based on DMC is analyzed. The simulation results show that this scheme is effective to the performance analysis of the mismatch of predictive control system.