摘要
针对现有工业系统运行中对非高斯噪声的影响处理不彻底的问题,设计一个基于动态数据校正的非高斯系统性能评价系统,其核心是通过动态数据校正来减少非高斯噪声对整个系统数据的影响。对于单输入单输出系统设计最小方差控制器,构建基于最小方差的控制性能评价系统,然后通过动态数据校正(dynamic data reconciliation,DDR)技术分析噪声对评价的影响,构建校正评价标准。由于方差仅测量随机变量与均值的偏离程度,不能全面刻画非高斯随机特征,故引入最小有理熵(minimum rational entropy,MRE),构建DDR-MRE控制器,通过计算DDR-MRE的控制性能评价(control performance assessment,CPA)基准来获取非高斯系统评价标准,从而使控制系统长期保持在最佳状态运行。试验结果表明,该系统技术核算的数据误差率在可接受范围内。
Aiming at the problem that the influence of non-Gaussian noise in the operation of existing industrial systems is not thoroughly dealt with,a performance evaluation system of non-Gaussian system based on dynamic data correction is designed.For single-input single-output system,a minimum variance controller is designed,and a control performance evaluation system based on minimum variance is constructed.Then,the influence of noise on the evaluation is analyzed by dynamic data reconciliation(DDR)technology,and the calibration evaluation criteria is constructed.Since variance only measures the degree of deviation between random variables and the mean value,non-Gaussian random features cannot be fully described.Therefore,minimum rational entropy(MRE)is introduced to construct a DDR-MRE controller.The control performance assessment(CPA)benchmark of DDR-MRE is calculated to obtain the non-Gaussian system evaluation standard,so as to keep the control system running in the best state for a long time.The test results show that the data error rate of this system is within the acceptable range.
作者
左霞
ZUO Xia(General Education Center,Quanzhou University of Information Engineering,Quanzhou 362000,China)
出处
《系统仿真技术》
2024年第1期85-90,共6页
System Simulation Technology
