基于Kringing近似模型的非线性预测函数控制  

Nonlinear Predictive Functional Control Based on Kringing Approximation Model

作　　者：符莎 施惠元[1] 赵民新 欧阳海鹏[1] 王尚明 

机构地区：[1]辽宁石油化工大学信息与控制工程学院

出　　处：《控制工程》2018年第3期396-401,共6页Control Engineering of China

基　　金：辽宁省自然科学基金项目(2013020024);辽宁省高等学校优秀科技人才支持计划项目(LR2015034)

摘　　要：针对复杂非线性过程，提出一种基于Kringing近似模型的预测函数控制方法。该方法首先对已知样本点进行抽样和拟合，并建立非线性系统的Kringing近似模型，利用Kringing近似模型中已知样本点的响应来构建预测函数控制的输出，将Kringing近似模型与预测函数控制（PFC）方法相结合，从而使非线性预测函数控制优化问题被转化为线性预测函数控制优化问题，解决了求控制系统输入时解非线性的问题。通过一组输入幅值不同的随机序列来检测模型的精度，在不同干扰的情况下将所提方法与一种非线性预测函数控制方法进行比较，并将该算法应用于经典的Box—Jenkins燃气炉系统。仿真结果表明，该方法具有较好的控制效果。A nonlinear predictive functional control method based on Kringing approximation model is proposed for complex nonlinear processes. The known sample points are sampled and fitted, and then the Kringing approximation model for the nonlinear system is established. The output of the predictive functional control is constructed by the response of the sample points of the Kringing approximation model. Combining the Kringing approximate model with the predictive functional control （PFC） method, the nonlinear predictive functional control optimization problem is transformed into a linear predictive functional control optimization problem to overcome the nonlinear problem of the control input, and then the optimal solution of the control system input can be deduced. A set of different amplitudes of random sequence are used to detect the precision of the model. In the case of different interferences, the control method of this paper is compared with a nonlinear predictive functional control strategy, and the algorithm is applied to the classic Box-Jenkins gas furnace system. The simulation results show that the proposed control algorithm has better control quality.

关 键 词：非线性系统 近似模型 预测函数控制 Kringing 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]