Volumen 28 (2018): Heft 2 (June 2018)
Advanced Diagnosis and Fault-Tolerant Control Methods (special section, pp. 233-333), Vicenç Puig, Dominique Sauter, Christophe Aubrun, Horst Schulte (Eds.)

This paper combines methods for the structural analysis of bipartite graphs with observer-based residual generation. The analysis of bipartite structure graphs leads to over-determined subsets of equations within a system model, which make it possible to compute residuals for fault detection. In observer-based diagnosis, by contrast, an observability analysis finds observable subsystems, for which residuals can be generated by state observers. This paper reveals a fundamental relationship between these two graph-theoretic approaches to diagnosability analysis and shows that for linear systems the structurally over-determined set of model equations equals the output connected part of the system. Moreover, a condition is proved which allows us to verify structural observability of a system by means of the corresponding bipartite graph. An important consequence of this result is a comprehensive approach to fault detection systems, which starts with finding the over-determined part of a given system by means of a bipartite structure graph and continues with designing an observerbased residual generator for the fault-detectable subsystem found in the first step.

This paper deals with the fault diagnosis of wind turbines and investigates viable solutions to the problem of earlier fault detection and isolation. The design of the fault indicator, i.e., the fault estimate, involves data-driven approaches, as they can represent effective tools for coping with poor analytical knowledge of the system dynamics, together with noise and disturbances. In particular, the proposed data-driven solutions rely on fuzzy systems and neural networks that are used to describe the strongly nonlinear relationships between measurement and faults. The chosen architectures rely on nonlinear autoregressive models with exogenous input, as they can represent the dynamic evolution of the system along time. The developed fault diagnosis schemes are tested by means of a high-fidelity benchmark model that simulates the normal and the faulty behaviour of a wind turbine. The achieved performances are also compared with those of other model-based strategies from the related literature. Finally, a Monte-Carlo analysis validates the robustness and the reliability of the proposed solutions against typical parameter uncertainties and disturbances.

This paper considers the problem of diagnosability analysis of discrete event systems modeled by labeled Petri nets (LPNs). We assume that the LPN can be bounded or unbounded with no deadlock after firing any fault transition. Our approach is novel and presents the on-the-fly diagnosability analysis using verifier nets. For a given LPN model, the verifier net and its reachability graph (for a bounded LPN) or coverability graph (for an unbounded LPN) are built on-the-fly and in parallel for diagnosability analysis. As soon as a diagnosability decision is established, the construction is stopped. This approach achieves a compromise between computation limitations due to efficiency and combinatorial explosion and it is useful to implement an engineering approach to the diagnosability analysis of complex systems.

The water loss detection and location problem has received great attention in recent years. In particular, data-driven methods have shown very promising results mainly because they can deal with uncertain data and the variability of models better than model-based methods. The main contribution of this work is an unsupervised approach to leak detection and location in water distribution networks. This approach is based on a zone division of the network, and it only requires data from a normal operation scenario of the pipe network. The proposition combines a periodic transformation and a data vector extension together with principal component analysis of leak detection. A reconstruction-based contribution index is used for determining the leak zone location. The Hanoi distribution network is employed as the case study for illustrating the feasibility of the proposal. Single leaks are emulated with varying outflow magnitudes at all nodes that represent less than 2.5% of the total demand of the network and between 3% and 25% of the node’s demand. All leaks can be detected within the time interval of a day, and the average classification rate obtained is 85.28% by using only data from three pressure sensors.

The paper deals with the problem of designing sensor-fault tolerant control for a class of non-linear systems. The scheme is composed of a robust state and fault estimator as well as a controller. The estimator aims at recovering the real system state irrespective of sensor faults. Subsequently, the fault-free state is used for control purposes. Also, the robust sensor fault estimator is developed in a such a way that a level of disturbances attenuation can be reached pertaining to the fault estimation error. Fault-tolerant control is designed using similar criteria. Moreover, a separation principle is proposed, which makes it possible to design the fault estimator and control separately. The final part of the paper is devoted to the comprehensive experimental study related to the application of the proposed approach to a non-linear twin-rotor system, which clearly exhibits the performance of the new strategy.

This paper presents a fault tolerant position tracking controller for a hexarotor system. The proposed controller has a cascaded structure composed of a position and an attitude control loop. The nominal controller is augmented by an adaptive control allocation which compensates for faults and failures within the propulsion system without reconfiguration of the controller. Simultaneously, it is able to implement a degraded control strategy which prioritizes specific control directions in the case of extreme degradation. The main contribution is a controller that is a step closer to application scenarios by including outdoor GPS-based flight tests, onboard computation and the handling of unknown degradation and failure of any rotor.

Two techniques for the control of a grid side converter in a wind energy conversion system. The system is composed of a fixed pitch angle wind turbine followed by a permanent magnet synchronous generator and power electronic converters AC-DC-AC. The main interest is in how to control the inverter in order to ensure the stability of the DC link voltage. Two control methods based on the fuzzy approach are applied and compared. First, a direct Mamdani fuzzy logic controller is presented. Then, a T-S fuzzy controller is elaborated based on a T-S fuzzy model. The Lyapunov theorem and H-infinity performance are exploited for stability analysis. Besides, the feedback controller gains are determined using linear matrix inequality tools. Simulation results are derived in order to prove the robustness of the proposed control algorithms and to compare their performances.

Positive linear continuous-time systems are analyzed via conformable fractional calculus. A solution to a fractional linear system is derived. Necessary and sufficient conditions for the positivity of linear systems are established. Necessary and sufficient conditions for the asymptotic stability of positive linear systems are also given. The solutions of positive fractional linear systems based on the Caputo and conformable definitions are compared.

In the context of gas transportation, analytical solutions are helpful for the understanding of the underlying dynamics governed by a system of partial differential equations. We derive traveling wave solutions for the one-dimensional isothermal Euler equations, where an affine linear compressibility factor is used to describe the correlation between density and pressure. We show that, for this compressibility factor model, traveling wave solutions blow up in finite time. We then extend our analysis to networks under appropriate coupling conditions and derive compatibility conditions for the network nodes such that the traveling waves can travel through the nodes. Our result allows us to obtain an explicit solution for a certain optimal boundary control problem for the pipeline flow.

This paper deals with the problem of robust fault detection (FD) for an unmanned aerial vehicle (UAV) flight control system (FCS). A nonlinear model to describe the UAV longitudinal motions is introduced, in which multiple sources of disturbances include wind effects, modeling errors and sensor noises are classified into groups. Then the FD problem is formulated as fault detection filter (FDF) design for a kind of nonlinear discrete time varying systems subject to multiple disturbances. In order to achieve robust FD performance against multiple disturbances, simultaneous disturbance compensation and H₁/H_∞ optimization are carried out in designing the FDF. The optimality of the proposed FDF is shown in detail. Finally, both simulations and real flight data are applied to validate the proposed method. An improvement of FD performance is achieved compared with the conventional H₁/H_∞-FDF.

The main impedance control schemes in the task space require accurate knowledge of the kinematics and dynamics of the robotic system to be controlled. In order to eliminate this dependence and preserve the structure of this kind of algorithms, this paper presents an adaptive impedance control approach to robot manipulators with kinematic and dynamic parametric uncertainty. The proposed scheme is an inverse dynamics control law that leads to the closed-loop system having a PD structure whose equilibrium point converges asymptotically to zero according to the formal stability analysis in the Lyapunov sense. In addition, the general structure of the scheme is composed of continuous functions and includes the modeling of most of the physical phenomena present in the dynamics of the robotic system. The main feature of this control scheme is that it allows precise path tracking in both free and constrained spaces (if the robot is in contact with the environment). The proper behavior of the closed-loop system is validated using a two degree-of-freedom robotic arm. For this benchmark good results were obtained and the control objective was achieved despite neglecting non modeled dynamics, such as viscous and Coulomb friction.

This paper investigates an N-policy GI/M/1 queue in a multi-phase service environment with disasters, where the system tends to suffer from disastrous failures while it is in operative service environments, making all present customers leave the system simultaneously and the server stop working completely. As soon as the number of customers in the queue reaches a threshold value, the server resumes its service and moves to the appropriate operative service environment immediately with some probability. We derive the stationary queue length distribution, which is then used for the computation of the Laplace-Stieltjes transform of the sojourn time of an arbitrary customer and the server’s working time in a cycle. In addition, some numerical examples are provided to illustrate the impact of several model parameters on the performance measures.

This work extends the dynamic programming approach to calculation of an elastic metric between two curves to finding paths in pairs of graph drawings that are closest under this metric. The new algorithm effectively solves this problem when all paths between two given nodes in one of these graphs have the same length. It is then applied to the problem of pattern recognition constrained by a superpixel segmentation. Segmentations of test images, obtained without statistical modeling given two shape endpoints, have good accuracy.

In recent years, research in automated facial expression recognition has attained significant attention for its potential applicability in human-computer interaction, surveillance systems, animation, and consumer electronics. However, recognition in uncontrolled environments under the presence of illumination and pose variations, low-resolution video, occlusion, and random noise is still a challenging research problem. In this paper, we investigate recognition of facial expression in difficult conditions by means of an effective facial feature descriptor, namely the directional ternary pattern (DTP). Given a face image, the DTP operator describes the facial feature by quantizing the eight-directional edge response values, capturing essential texture properties, such as presence of edges, corners, points, lines, etc. We also present an enhancement of the basic DTP encoding method, namely the compressed DTP (cDTP) that can describe the local texture more effectively with fewer features. The recognition performances of the proposed DTP and cDTP descriptors are evaluated using the Cohn-Kanade (CK) and the Japanese female facial expression (JAFFE) database. In our experiments, we simulate difficult conditions using original database images with lighting variations, low-resolution images obtained by down-sampling the original, and images corrupted with Gaussian noise. In all cases, the proposed method outperforms some of the well-known face feature descriptors.

In a general regression neural network (GRNN), the number of neurons in the pattern layer is proportional to the number of training samples in the dataset. The use of a GRNN in applications that have relatively large datasets becomes troublesome due to the architecture and speed required. The great number of neurons in the pattern layer requires a substantial increase in memory usage and causes a substantial decrease in calculation speed. Therefore, there is a strong need for pattern layer size reduction. In this study, a self-organizing map (SOM) structure is introduced as a pre-processor for the GRNN. First, an SOM is generated for the training dataset. Second, each training record is labelled with the most similar map unit. Lastly, when a new test record is applied to the network, the most similar map units are detected, and the training data that have the same labels as the detected units are fed into the network instead of the entire training dataset. This scheme enables a considerable reduction in the pattern layer size. The proposed hybrid model was evaluated by using fifteen benchmark test functions and eight different UCI datasets. According to the simulation results, the proposed model significantly simplifies the GRNN’s structure without any performance loss.