Robust Fault Detection Observer Design using H ∞ /µ Techniques for Uncertain Flight Control Systems

Sadrnia, Mohammad Ali; Chen, J.; Patton, Ron J.

doi:10.1016/s1474-6670(17)42654-1

Cited by 4 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, equation ( 25) gives a recursive form for magnitude of the residual signal of the faulty system estimate the parameters of systems de ned by equaclosely approaches the magnitude of the residual signal tion (3). For this purpose, the system model of the healthy system.…”

Section: Sensor Fault Diagnosismentioning

confidence: 99%

“…To model the healthy system, a feedforward ANN, As a result, many methods have been developed in order which is subjected to delayed data of the inputs and to detect any unsatisfactory operation conditions within outputs of the systems, is trained by the back-propathe processes [1][2][3][4]. Such unsatisfactory operation congation algorithm to give the following model: ditions are commonly related to a malfunction of a set ŷ(t +1)=F ( f s (t)Õ 1y f (t), f s (t Õ1)Õ 1y f (t Õ1), of sensors, actuators or system parameters.…”

Section: Notationmentioning

confidence: 99%

See 1 more Smart Citation

Fault diagnosis for unknown non-linear systems via neural networks and its comparisons and combinations with recursive least-squares based techniques

Wang

Noriega

2001

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

In this paper, an algorithm for the detection and diagnosis of faults in sensors and actuators in unknown non-linear systems is presented. Firstly, the normal operation of the non-linear system is modelled via a feedforward neural network in order to produce a healthy model which is used later as a redundant relation. It is assumed that every faulty sensor or actuator can be modelled by a single parameter f. Such a parameter is de ned to be f =1 when the system is healthy and 0< f <1 otherwise. A residual signal is evaluated from the diVerence of the output of the healthy model and that of the system. A simple threshold function is used to detect the faulty behaviour of the system. The estimation of f is then computed by minimizing the diVerences between the output of the system and that of the healthy neuro model with the help of the gradient descent rule. It has been shown that the algorithm is able to estimate the fault with large magnitude accurately. However, if the deviation of f from its healthy value is small, a linearized healthy model, together with the recursive least-squares algorithm, can be used to estimate the fault. As a result, a combination of both methods will provide a powerful technique for accurate fault diagnosis. Keywords: fault detection and diagnosis, non-linear dynamic systems, actuators and sensors, neural networks, least-squares estimation, non-linear autoregressive moving-average model with exogenous input (NARMAX model ) NOTATION l r step size or the learning rates for experiment system m, n structure orders of the system a i , b j parameters of the linearized model p information vector composed of past B 1 , B 2 bias matrices of neural networks inputs and outputs e modelling error P(t ) covariance matrix e(t ) Gaussian noise sequence u(t ) input to the system E, J fa , J fs objective functions w ij weights of neural networks f fault parameter W 1 , W 2 weight matrices of neural networks f a (t) actuator fault size at sample time t x f (t ) information vector of the linearized f s (t ) sensor fault size at sample time t model f a (t) estimated actuator fault size at sample y(t ) output of the system time t ŷ estimated output of the model f s (t ) estimated sensor fault size at sample y f (t) measured output of the system time t Y 00 d.c. component of the linearized model F( .)non-linear dynamic behaviour of the system á , è learning rates for the arti cial neural F (.) estimated non-linear dynamic behaviour network weights training of the system ¢y, ¢u incremental values of output and input respectivelyThe MS was

show abstract

Section: Sensor Fault Diagnosismentioning

confidence: 99%

Section: Notationmentioning

confidence: 99%

Fault diagnosis for unknown non-linear systems via neural networks and its comparisons and combinations with recursive least-squares based techniques

Wang

Noriega

2001

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

show abstract

“…Sauter et al [1997] studied this problem where the fault sensitivity is enhanced by applying an optimal post-filter to the "primary residual". The problem of enhancing fault sensitivity while increasing robustness against disturbances and modeling errors was studied extensively by Sadrnia et al [1997aSadrnia et al [ , 1997bSadrnia et al [ , 1997c. The essential idea is to reach an acceptable compromise between disturbance robustness and fault sensitivity.…”

Section: Introductionmentioning

confidence: 99%

Formulating and Solving Robust Fault Diagnosis Problems Based on a H∞ Setting

Chen

2008

IFAC Proceedings Volumes

View full text Add to dashboard Cite

“…Analytical redundancy has also been used in flight control systems; the very same aircraft used to conduct research on fly-by-wire technology was also used as testbed for an analytical redundancy based fault detection algorithm ( [17] and [56]). Since then, a number of results have been published on the suitability of analytical redundancy approach for reconfiguration of flight control systems ( [59] and [10]), and for diagnosis of aircraft actuator and sensor failures ( [54] and [40]). Nevertheless, doubts remain on the possibility that analytical redundancy based solutions can meet the strict fault tolerance requirements of flight control systems [44].…”

Section: Analytical Redundancy In Flight Control Systemsmentioning

confidence: 99%

Formal specification of requirements for analytical redundancy-based fault -tolerant flight control systems

Gobbo¹,

Napolitano²

View full text Add to dashboard Cite

Formal Specification of Requirements for Analytical Redundancy based Fault Tolerant Flight Control Systems By Diego Del Gobbo Flight control systems are undergoing a rapid process of automation. The use of Fly-By-Wire digital flight control systems in commercial aviation (Airbus 320 and Boeing FBW-B777) is a clear sign of this trend. The increased automation goes in parallel with an increased complexity of flight control systems with obvious consequences on reliability and safety. Flight control systems must meet strict fault-tolerance requirements. The standard solution to achieving fault tolerance capability relies on multi-string architectures. On the other hand, multi-string architectures further increase the complexity of the system inducing a reduction of overall reliability. In the past two decades a variety of techniques based on analytical redundancy have been suggested for fault diagnosis purposes. While research on analytical redundancy has obtained desirable results, a design methodology involving requirements specification and feasibility analysis of analytical redundancy based fault tolerant flight control systems is missing. The main objective of this research work is to describe within a formal framework the implications of adopting analytical redundancy as a basis to achieve fault tolerance. The research activity involves analysis of the analytical redundancy approach, analysis of flight control system informal requirements, and re-engineering (modeling and specification) of the fault tolerance requirements. The USAF military specification MIL-F-9490D and supporting documents are adopted as source for the flight control informal requirements. The De Havilland DHC-2 general aviation aircraft equipped with standard autopilot control functions is adopted as pilot application. Relational algebra is adopted as formal framework for the specification of the requirements. The detailed analysis and formalization of the requirements resulted in a better definition of the fault tolerance problem in the framework of analytical redundancy. Fault tolerance requirements and related certification procedures turned out to be considerably more demanding than those typically adopted in the literature. Furthermore, the research work brought up to light important issues in all fields involved in the specification process, namely flight control system requirements, analytical redundancy, and requirements engineering. I would like to thank Dr. Marcello Napolitano, my advisor, for his support during this research. I am thankful to Dr. Ali Mili for having brought some light in the chaos that characterized the early phases of this work. His guidance was crucial to the successful completion of this project. I am also thankful to Dr. Francesco Nasuti for his friendship and for the numerous helpful discussions on the many faces of analytical redundancy.

show abstract

Robust Fault Detection Observer Design using H ∞ /µ Techniques for Uncertain Flight Control Systems

Cited by 4 publications

References 9 publications

Fault diagnosis for unknown non-linear systems via neural networks and its comparisons and combinations with recursive least-squares based techniques

Fault diagnosis for unknown non-linear systems via neural networks and its comparisons and combinations with recursive least-squares based techniques

Formulating and Solving Robust Fault Diagnosis Problems Based on a H∞ Setting

Formal specification of requirements for analytical redundancy-based fault -tolerant flight control systems

Contact Info

Product

Resources

About