最小方差调节律参数灵敏度分析

Analysis on the sensitivity of minimal variance regularity

探讨了最小方差调节规律参数灵敏度问题 .给出了最小方差调节律参数灵敏度的精确描述 ,推导出标么化参数灵敏度的计算公式并进行了分析 .结果表明 :方差目标函数 (ISE)对控制器参数在最优参数附近宽范围内变化不敏感 ,为最小方差调节器的降阶设计提供了理论依据 ;适当配置系统闭环极点 ,能使系统近似ISE最优并具有较强的扰动抑制能力 ;仿真实验证明 ,适当配置了闭环极点的系统在噪声和扰动的双重作用下 ,性能良好 .

This paper is concerned about the sensitivity of integral of error squared (ISE) criteria to the parameters of minimal variance regulator (MVR). A precise definition of sensitivity of ISE criteria to MVR parameters is given, and a formulation to calculate the unitary sensitivity to parameters is deduced. Both theoretic analysis and emulation prove the conclusion that the ISE criterion is insensitive to the variety of MVR parameters in a wide neighborhood of the regulator′s parameters which make ISE criterion optimized. This conclusion provides the basis for design of the reduce order MVR. As an application of this theory, a piece of improved control regularity is developed. With closed loop system′s poles being correctly assigned, closed loop system can restraint certain disturbance effectively while the ability to restraint stochastic noise isn′t lowered. The emulation shows that this kind of approximately minimal variance regulation can offer more excellent performance when the system is under the action of both disturbance and noise. [