Development and evaluation of an integral sliding mode fault‐tolerant control scheme on the RECONFIGURE benchmark

Abstract

Summary This paper describes the development, application, and evaluation of a linear parameter‐varying integral sliding mode control allocation scheme to the Reconfiguration of Control in Flight for Integral Global Upset Recovery benchmark model to deal with an actuator failure/fault scenario. The proposed scheme has the capability to maintain close to nominal (fault free) load factor control performance in the face of elevator failures/faults by including a retrofitted integral sliding mode term and then rerouting (via control allocation) the augmented control signal to healthy elevators without reconfiguring the baseline controller. In order to mitigate any chattering appearing in the elevator demands, the retrofitted signal is based on a super‐twisting sliding mode structure. This produces a control signal that is continuous and does not have the discontinuous switching nature of traditional sliding mode schemes. The scheme is evaluated using an industrial Functional Engineering Simulator developed as part of the Reconfiguration of Control in Flight for Integral Global Upset Recovery project. Monte Carlo campaign results are shown to demonstrate the performance of the proposed scheme.

Integral sliding mode fault‐tolerant control allocation for a class of affine nonlinear system

“…In this paper, in the case study, the discontinuous control signal is approximated by a continuous signal by approximating the unit vector with a sigmoid function . An alternative way of developing a continuous controller straightforwardly is to use higher‐order sliding modes (eg, Chen et al and Rios et al proposed a continuous ISM FTC allocation scheme).…”

“…To demonstrate the efficacy of the scheme, the retrofitted ISM/CA procedure will be applied to a longitudinal rigid-body aircraft model and evaluated using the RECOVER benchmark. 41 The main contributions of this paper are as follows: (i) Fault-tolerant ISM schemes developed for linear time-invariant 38,42 and linear parameter-varying (LPV) representations 43,44 are extended into a more general nonlinear affine framework; (ii) compared with the work of Rubagotti et al, 10 a CA component has been embedded to ensure fault tolerance and the dimension of the switching function reduced; (iii) one of the two design schemes developed in this paper introduces fault tolerance by retrofitting an ISM control signal to an existing baseline controller in the presence of actuator faults/failures. The structure of this paper is as follows.…”

Interval sliding mode observer-based fault accommodation for non-minimum phase LPV systems with online control allocation

“…In recent decades, the interest in FA for LPV systems has increased and many results have been published, for example Rotondo et al (2014), Chen et al (2017) and Sloth et al (2010). But generally speaking, because conventional observers do not work well for estimating states and faults for general LPV systems, almost all of these results are developed for specific LPV systems under structure restrictions such as polytopic ones in Rotondo et al (2014) or with measurable scheduling parameter (see e.g.…”

“…But generally speaking, because conventional observers do not work well for estimating states and faults for general LPV systems, almost all of these results are developed for specific LPV systems under structure restrictions such as polytopic ones in Rotondo et al (2014) or with measurable scheduling parameter (see e.g. Chen et al (2017) and Sloth et al (2010)). Recently, an effective methodology for controlling LPV systems with unknown scheduling parameters has been proposed based on interval observers in Efimov et al (2013), which has subsequently been applied to FA for faulty LPV systems in Lamouchi et al (2018).…”

“…Except for Efimov et al (2012), in Oubabas et al (2018), an ISMO is constructed via a convex sum of an upper estimator and a lower estimator, while in , an ISMO is proposed based on regular form to detect incipient faults. On the other hand, sliding mode techniques have been used for FTC extensively because of their inherent robustness to matched uncertainties, for example, sliding mode FTC with online CA , and Chen et al (2017), and sliding mode observer based FTC Li et al (2014) and Shi et al (2015). However, to the best of the authors' knowledge, ISMOs have not been applied before to FA for LPV systems.…”