$$H_{2}$$-Optimization-based Fault Detection for LDTV Systems
Maiying Zhong,Ting Xue,Steven X. Ding,Donghua Zhou
DOI: https://doi.org/10.1007/978-981-19-5438-2_5
2022-01-01
Abstract:$$H_{2}$$ -optimization is a prevalent concept in control theory and is adopted, in its initial form, for optimization issues related to $$H_{2}$$ -norm of transfer functions. Thanks to its optimization interpretation in the context of the well-known linear quadratic Gaussian (LQG) control problem, $$H_{2}$$ -optimization is widely accepted as an established term for optimal control and filtering of linear systems under a quadratic cost function. In this regard, we will, in this chapter, introduce fault detection schemes for LDTV systems. Apart from presenting the standard algorithms, we will focus on gaining deep insights into optimal estimation, residual generation and optimal fault detection issues behind the optimization, and learning alternative interpretations. To this end, some basic estimation and fault detection problems will be, at first, addressed. Although these problems are formulated in the context of static processes, they are helpful to understand the centerpiece of conceptual solutions of any fault detection problems, independent of the system type under consideration. The subsequent study on fault detection in LDTV systems is in fact a natural extension and applications of these solutions, with the aid of advanced system theoretical handlings.