Robust fault estimation of discrete‐time nonlinear plants via a comprehensive partition‐based switching scheme

Abstract

The problem of robust fault estimation (FE) for a class of nonlinear systems in discrete-time domain is studied via proposing a comprehensive partition-based switching scheme. Both enhancement factors and attenuation factors are designed synchronously for accomplishing a comprehensive partition of the entire space spanned by the considered normalized fuzzy weighting functions. Thanks to the proposed comprehensive partition-based switching scheme, six subspaces would be obtained by identifying the real-time permutation order of the underlying parameters. Therefore, a novel fuzzy FE observer with six kinds of switching modes can be on-line implemented. As a result, the robust H ∞ performance index can be enhanced since much more information is employed than those reported results. Finally, simulation comparisons over the existing results are provided to show the obtained advantages of the article.

Enhanced gain‐scheduling fault estimation of discrete‐time Takagi–Sugeno nonlinear systems: A novel free matrix approach

“…▪ Remark 3. In recent literature such as, [36][37][38] the zero-order free matrices are designed in order to adjust related positive/negative margins among various monomials. In principle, all the different monomials must share the same zero-order free matrices and thus this limitation does influence the algorithm efficiency.…”

“…It does mean that the estimation accuracy has been improved in this study. Therefore, it is reasonable to infer that the developed scheme of ( 6) is superior to the reported result 38 about the following four aspects: smaller 𝛾 min , less consumed time of solving the required LMIs, smaller N online and better SUM(t).…”

“…Next, based on Theorem 1 with g = 1 and s = 2, two groups of (J m k , G m k , Q m k ) have been obtained by means of ( 11)-( 12) for 𝛾 = 2.6227: TA B L E 1 Comparisons of the algorithm efficiency for IJRNC 2020 38 Further, we employ (6) with our obtained (J m k , G m k , Q m k ) for estimating i L (t)/f (t) by setting its initial values to be v C (t) = vC (t) = 0.0V and i L (t) = î L (t) = 0.0A. The actual trajectory of i L (t) and the estimated trajectory of î L (t) have been exhibited in Figure 2 while the actual trajectory of f (t) and the estimated trajectory of f (t) have also been given in Figure 3.…”

“…Therefore, it does provide a good feasibility to obtain equal or better results with usage of much less computing resource over the recent one. 38 The effectiveness and superiority of our developed free matrix approach are well validated by using performance comparisons via implementing the benchmark case study.…”

Fault estimation for periodic piecewise T‐S fuzzy systems

“…The studies on fault diagnosis from fault detection to fault tolerance have theoretical meaning and practical significance to monitor system safety and improve system reliability 24,25 . Concerning the safety and reliability of T‐S fuzzy systems, fault detection and fault estimation play important roles in both system maintenance and fault‐tolerant control 26‐28 . Observer‐based fault detection and estimation approaches have been widely applied to fuzzy systems subject to practical limitations, such as communication constraints, 29 measurement outliers, 30 and time‐delay 31,32 .…”