Oscillatory Failure Case Research Articles

A sliding mode observer approach to oscillatory fault detection in commercial aircraft

The Flight Control System (FCS) is one of the most important systems in all modern aircraft. For such systems it is required to have robust Fault Detection Isolation and Reconfiguration (FDIR) functionalities with high detection performance. In this work we specifically consider the Oscillatory Failure Cases (OFC), which, if not mitigated, can cause additional structural loads for which the aircraft is not designed. A Sliding Mode Observer (SMO) based detection method is proposed for fast and consistent detection of these OFC faults. A benchmark of a generic aircraft FCS equipped with OFC simulation capabilities, as well as the presented solution for detection, have previously been presented within a competition at the 2020 IFAC World Congress.

An Approach to Oscillatory Failure Case Detection in the Aerospace Industrial Benchmark

This paper proposes an approach to detect the oscillatory failure case (OFC) and reconfigure the flight control law after the OFC detection in the Aerospace Industrial Benchmark. To cope with system nonlinearity, a square root unscented Kalman filter is utilized to generate the residual signal. An estimation of the parameter uncertainty based on the fault-free data is incorporated to reduce the influence of model uncertainty on the residual signal significantly. To evaluate the residual signal, a modified integral absolute error based index is introduced along side a check of the oscillation regularity. Our approach is capable of detecting OFCs with very low amplitudes within at most 3 oscillation periods without producing any false alarm, regardless of load factor command, turbulence level or OFC location, type and frequency. The minimum detectable OFC amplitude is 0.6 mm and 0.9 mA for liquid OFCs located at sensor signal and current input signal, respectively, and 0 mm and 0.9 mA for solid OFCs.

Aerospace industrial benchmarking and fault-tolerant control design using NARX model-based observer

This paper proposes a fault tolerant control (FTC) design with a fault detection (FD) scheme for oscillatory failure cases (OFC) within the framework of “The Joint Airbus-Stellenbosch University Industrial Benchmark.” The FD scheme consists of an observer using “Nonlinear Autoregressive models with eXogenous input” (NARX) model, a signal processing part, and an FD algorithm. The proposed FTC design smoothly switches controllers by adjusting weights according to a sigmoid function using a reconfiguration scheme. The parameters in both schemes are optimized by grid search based on the metrics set to evaluate performance. These schemes have the advantage of the ease of industrial use because they have a simple structure, and parameters can be tuned in a systematic way. Simulation results show that the proposed FD scheme's ability to detect OFC meets the benchmark's prescriptions regardless of the defined turbulence or load factor input. The robustness of the proposed FD scheme was evaluated using a Monte Carlo method by varying the parameters of the actuator model to confirm the limit value of the OFC amplitude that can be robustly detected.

Fault Diagnostics for Oscillatory Failure Case in Aircraft Elevator Servos

Oscillatory Failure Case (OFC) is a characteristic sinusoidal fault that propagates through control surfaces and actuator servo loops of commercial aircrafts that originates from spurious sensor measurements and actuator commands. The goal of this work is to detect such faults of unknown amplitudes and frequencies in the presence of turbulence. The fault detection scheme employs a graph-theoretic approach that uses a structural model of the aircraft's longitudinal dynamics and the servo actuators to determine residuals for OFC fault detection and isolation. A smooth controller reconfiguration strategy is proposed to switch between a nominal control law and a degraded alternate control law which utilizes sensor measurement feedback as well as virtual sensors from a digital model of longitudinal cruising dynamics without affecting the controller structure and feedback loops.

Detection of Oscillatory Failures in Hydraulic Actuators of Aircraft using Linear Predictive Coding and Signal Spectrum Analysis

This paper focuses on fault detection for oscillatory failures in hydraulic actuators of aircraft. These oscillatory failures, if not rectified in time, can cause severe loads on the airframe and eventually lead to structural damage. In this paper, a novel oscillatory failure case (OFC) detection algorithm is proposed which uses a nonlinear observer based on the mathematical model of the actuator in order to generate residual following a Linear Predictive Coding analysis of the residual to detect oscillatory behavior. Finally, OFC decisions are made after the quantification of the residual in frequency domain. To illustrate the effectiveness of the proposed algorithm, results are presented using a high-fidelity industrial benchmark simulation. Furthermore, a comparative study is provided against an existing technique.

Time Stacking Decision Tree for Oscillatory Failure Case Detection in a Flight Control System

Flight control systems are an essential part of aircraft, consisting of all elements between the pilot and the movable parts, such as actuators, controllers, and sensors, to control the aircraft's attitude, trajectory and speed. In this context, oscillatory failure cases are problems that can occur in any part of the flight control system, causing undesired oscillatory behavior that affects the aircraft's structural load. On the one hand, correctly detecting such failures enables weight saving to improve aircraft performance in different aspects; on the other hand, such detection enables reconfiguring the control law to keep the aircraft in its path. This work describes a data-driven oscillatory fault detection model and the control law reconfiguration for a fight control system for the Aerospace Industrial Benchmark. Preliminary results indicate that the developed system can accurately detect faults in several scenarios and properly reconfigure the control loop.

Flight Control Reconfiguration in Oscillatory Failure Case Scenarios

A fault detection system together with a flight control reconfiguration algorithm is proposed to tackle the problem of oscillatory failure case scenarios in hydraulically actuated flight control loops. The fault detection system employs a set of residual filters designed using a model-based approach via command-input decoupling and free assignment of the filter dynamics via dynamic inversion. As underlying design model, an optimized, multi-model hydraulic actuator approximation covering the actuator dynamics together with stochastically distributed, uncertain parameters is derived. A flight control reconfiguration algorithm is presented which smoothly transitions the control law from its normal configuration to the alternate law after detection of a fault. The performance of the transition is analyzed using advanced linear time-varying analysis methods based on an extension of the well-known Bounded Real Lemma. The achieved detection and reconfiguration performance is verified via non-linear simulations.

Oscillatory Failure Case detection in flight control systems via wavelets decomposition

A fault detection design is proposed for addressing the Oscillatory Failure Case (OFC) detection problem, introduced in the joint Airbus-Stellenbosch university aerospace industrial-benchmark competition called at the IFAC 2020 World Congress11https://www.ifac2020.org/program/competitions/aerospace-industrial-fault-detection.. The detection scheme is comprised of an output estimator, a wavelet decomposition and an energy-based denoising method, and the residual evaluation unit. The detection problem of wide frequency range OFCs is also addressed. According to the achieved simulation results, the proposed fault detection method is able to satisfy the competition prescriptions in the frequency range [1 10] Hz for those OFC’s having an amplitude greater than 2.3 mm for OFCs at rod position sensor, or 1.4 mA for OFCs at servo input current, regardless of disturbances level, uncertainties and load factor control input. In other cases, faults are detected slightly after the prescribed detection limit, with some interesting exceptions.

A Data-Driven Approach For Actuator Servo Loop Failure Detection

This paper studies a data-driven approach to detect faults in flight control systems of civil aircraft. A particular class of failures, referred to as Oscillatory Failure Cases (OFC), impacting the actuator servo loop has motivated the authors to consider a data-driven approach based on distance and correlation measures (see reference [Goupil et al.(2016). A data-driven approach to detect faults in the Airbus flight control system. IFAC-PapersOnLine, 49(17), 52-57] of this paper) leading to promising results compared to the state-of-the-art methods based on analytical redundancy. The present paper extends the formulation and the results of the considered OFC detection approach investigating Support Vector Machine (SVM) techniques to define a more accurate detector based on distance and correlation measures.

A Data-Driven Approach to Detect Faults in the Airbus Flight Control System

This paper presents a data-driven strategy for the detection of failures impacting the flight control system. Early and robust detection of Oscillatory Failure Case (OFC) allows the aircraft structural design to be optimized, which in turn helps improve the aircraft environmental footprint thanks to weight saving. Compared to existing model-based techniques already used on in-service Airbus aircraft, this paper studies a novel signal processing approach based on distance and correlation. It is shown that a mixed similarity index between Euclidean distance and logarithmic invariant divergence gives promising detection results. This paper details the proposed approach by insisting on practical constraints due to implementation in embedded real-time systems such as the flight control computer. Preliminary results obtained from a Verification & Validation (V&V) on-going campaign are presented.

Design of a non-homogeneous differentiator for actuator oscillatory failure case reconstruction in noisy environment

This work addresses the problem of oscillatory failure case detection in the electrical flight control system of a generic commercial airplane. A non-homogeneous differentiator is first used to provide accurate derivatives in noisy environment and fast convergence time. In this study case, fault detection is addressed in the unknown input estimation issue for fault reconstruction with the same evaluation techniques currently employed in Airbus A380 airplanes. Based on the non-homogeneous differentiator theory, an algorithm for differentiator tuning is provided to manage a trade-off between the accuracy of fault reconstruction and the convergence time that is compliant with specified detection time requirements. Performance and robustness of the developed monitoring strategy are assessed using a high-fidelity Airbus benchmark and a parametric test campaign for the flight scenarios defined in the EU-FP7 ADDSAFE project.

From theory to flight tests: Airbus Flight Control System TRL5 achievements

In this article, concrete aerospace industrial achievements corresponding to a level 5 on the Technological Readiness Level (TRL) scale are presented. Starting from basic research levels up to industrial validation, it is shown how it is possible to bridge the gap between advanced methods advocated by the academia and the more and more demanding industrial needs. To illustrate this position, a focus is made on real-time on-board Fault Detection and Diagnosis (FDD) for the upcoming and future innovative Flight Control System (FCS) of civil aircraft. This is exemplified by three concrete cases presenting the research and development process, from theory up to simulator and flight tests: 1) detection of Oscillatory Failure Cases in FCS thanks to a derivative-free variant of an extended Kalman filter; 2) detection of control surface lock-in-place failure (a.k.a. jamming) via recursive parametric modeling; 3) flight parameter Data Fusion with dedicated fuzzy logic.

Second order sliding mode observers for the ADDSAFE actuator benchmark problem

This paper presents the evaluation process and results associated with two different fault detection and diagnosis (FDD) schemes applied to two different aircraft actuator fault benchmark problems. Although the schemes are different and bespoke for the problem being addressed, both are based on the concept of a second order sliding mode. Furthermore both designs are considered as ‘local’ in the sense that a localized actuator model is used together with local sensor measurements. The schemes do not involve the global aircraft equations of motion, and therefore have low order. The first FDD scheme is associated with the detection of oscillatory failure cases (OFC), while the second scheme is aimed at the detection of actuator jams/runaways. For the OFC benchmark problem, the idea is to estimate the OFC using a mathematical model of the actuator in which the rod speed is estimated using an adaptive second order exact differentiator. For the jam/runaway actuator benchmark problem, a more classical sliding mode observer based FDD scheme is considered in which the fault reconstruction is obtained from the equivalent output error injection signals associated with a second order sliding mode structure. The results presented in this paper summarize the design process from tuning, testing and finally industrial evaluation as part of the ADDSAFE project.

New sequential probability ratio test: Validation on A380 flight data

A new residual analysis approach is proposed in order to improve fault detection performance and robustness. The proposed methodology is based on the Wald's sequential test which assumes that residual signal distribution is Gaussian. The new sequential test overcomes this limitation and allows handling a signal with a Laplace distribution. The proposed approach is applied to the detection of oscillatory failure cases on Airbus A380 elevator actuator. The sequential test is assessed on flight data supplied by Airbus. First results are promising as there is already a noticeable improvement over results obtained with the industrial state-of-practice technique implemented on A380.

Actuator fault detection in aircraft systems: Oscillatory failure case study

This paper is devoted to the detection of Oscillatory Failure Case (OFC) in the Electrical Flight Control System (EFCS). Such failures lead to a strong interaction between loads and aeroelasticity and must be quickly detected and passivated. This paper proposes a hybrid monitoring scheme for robust and early detection of such unauthorized oscillatory events. The proposed technique has been developed within ADDSAFE project (a collaborative project supported by the European Seventh Framework Program: Advanced Fault Diagnosis for Sustainable Flight Guidance and Control). A robust finite-time differentiator is used to estimate derivatives in a noisy environment. Fault reconstruction is next achieved by solving on-line a nonlinear equation using a gradient descent method. Finally, fault detection and confirmation stage is based on the decision making rules currently used for in-service Airbus A380 airplane. Robustness and performance of the proposed fault detection scheme are tested using a high fidelity benchmark and intensive Monte Carlo simulations for several flight scenarios as specified within ADDSAFE project.

An adaptive sliding mode differentiator for actuator oscillatory failure case reconstruction

This paper proposes an adaptive sliding mode super-twisting differentiator which allows the gains to adapt based on the ‘quality’ of the sliding motion. A Lyapunov based analysis for the adaptive super-twisting scheme is presented to demonstrate its properties. As an example, the adaptive differentiator proposed in this paper has been used as part of a nonlinear FDI scheme for an Oscillatory Failure Case (OFC) in an actuator. The FDI scheme requires an estimate of the rod speed which is provided by the adaptive super-twisting differentiator. Due to the conditions in which the actuator operates, normally the differentiator gains are initialised at low values to ensure good rod speed estimation in fault free conditions. However for large amplitude/frequency OFCs, the gains must adapt in order to maintain sliding and provide a good estimation. Simulations on a high fidelity nonlinear aircraft benchmark model have been carried out for both liquid and solid OFCs.

Fault diagnosis and monitoring of oscillatory failure case in aircraft inertial system

A method for oscillatory fault detection and isolation is presented and used to detect oscillatory failures of redundant aircraft sensors involved in the computation of flight control laws. The objective is to switch off the erroneous sensor and to compute a consolidated parameter using data from valid sensors, in order to eliminate any anomaly before propagation in the control loop. The benefit of the presented method is to improve the consolidation process with a fault detection and isolation approach when only few sources (less than three) are valid. Different techniques are compared to accurately detect any behavioral change of the sensor outputs. The approach is validated on a normalized real flight data set.

LPV-model based identification approach of oscillatory failure cases

The reliable early detection of oscillatory failure cases (OFC) for modern fly-by-wire controlled civil aircraft is an important aspect in optimizing the structural design objectives for reducing the environmental footprint of the aircraft. We propose a complete methodology for the design of a model based fault detection and diagnosis (FDD) system, which allows the reliable early detection of OFC characterized by low amplitude oscillations in a given frequency range. The main factors for the achieved enhanced performance are: the accurate modelling of the surface actuator via linear parameter-varying (LPV) models, an improved oscillation detection by employing the recursive discrete Fourier transform, and the integrated tuning of the free parameters of the FDD system using multi-objective optimization techniques.

INPUT ESTIMATION VIA SLIDING-MODE DIFFERENTIATION FOR EARLY OFC DETECTION

Robust and early detection of Oscillatory Failure Case (OFC) in the Electrical Flight Control System (EFCS) of new generation aircraft appears to be a challenging problem. OFC leads to strong interactions with loads and aeroelasticity and consequently must be detected in time. This paper proposes a hybrid unknown input observer for robust and early detection of such unauthorized oscillatory events. The technique has been developed during ADDSAFE project. The monitoring strategy is associated with the decision making rules as currently used for in-service Airbus A380 airplane. The observer has two subsystems operating in continuous and in discrete time. The continuous part is based on a high-order sliding-mode differentiator. The discrete part is used to solve on-line a nonlinear equation using a gradient descent method. Robustness and performance of the proposed fault detection scheme are tested using a high fidelity benchmark and intensive Monte Carlo simulations based on several flight scenarios as specified in ADDSAFE project. The performance indicators of the proposed scheme are measured by its detection delay, its propensity to issue false alarms and whether it permits a failure to go undetected.

Optimization of fault detection performance for a class of nonlinear systems

SUMMARYThe problem of observer design for fault detection in a class of nonlinear systems subject to parametric and signal uncertainties is studied. The design procedure includes formalized optimization of observer‐free parameters in terms of trade‐offs for fault detection performance and robustness to external disturbances and model uncertainties. The technique makes use of some monotonicity conditions imposed on the estimation error dynamics. Efficiency of the proposed approach is demonstrated through the oscillatory failure case in aircraft control surface servo‐loops. Copyright © 2011 John Wiley & Sons, Ltd.