Fault Identification Algorithm Research Articles

An efficient diagnostic technique for distribution systems based on under fault voltages and currents

Service continuity is one of the major aspects in the definition of the quality of the electrical energy, for this reason the research in the field of faults diagnostic for distribution systems is spreading ever more. Moreover the increasing interest around modern distribution systems automation for management purposes gives faults diagnostics more tools to detect outages precisely and in short times. In this paper, the applicability of an efficient fault location and characterization methodology within a centralized monitoring system is discussed. The methodology, appropriate for any kind of fault, is based on the use of the analytical model of the network lines and uses the fundamental components rms values taken from the transient measures of line currents and voltages at the MV/LV substations. The fault location and identification algorithm, proposed by the authors and suitably restated, has been implemented on a microprocessor-based device that can be installed at each MV/LV substation. The speed and precision of the algorithm have been tested against the errors deriving from the fundamental extraction within the prescribed fault clearing times and against the inherent precision of the electronic device used for computation. The tests have been carried out using Matlab Simulink for simulating the faulted system.

Robust Detection and Estimation in Dynamic Systems and Statistical Signal Processing: Intersections, Parallel Paths and Applications

Algorithms for Robust Fault Detection and Identification (RFDI) in dynamic systems strive to be sensitive to faults in plant components, while being insensitive to plant and noise model uncertainties. In the field of signal detection with applications such as radar and medical images, sensitivity to signals of interest is also an objective, while robustness is concerned with undesirable interferences residing in unknown or partially known subspaces, and uncertainties in the signal and learned interference models. In this review article, we present the similarities and differences in the game theoretic, stochastic, and geometric subspace formulations and solution approaches to robust detection in dynamic systems and in signal processing. We illustrate how innovative formulations are successfully addressing challenging problems by briefly describing results from two examples, one in functional Magnetic Resonance Imaging (fMRI) of the brain, and the other in non-invasive image based classification of stem cells.

FDI algorithms of abrupt faults in controlled autoregressive processes

PurposeThe purpose of this paper is to present research in detecting and identifying abrupt faults in controlled auto‐regressive (CAR) processes.Design/methodology/approachModel‐based approach is adopted in this paper. Two series of fault‐tolerant iterative estimators are set up to estimate online the coefficients in a CAR process. Based on these fault‐tolerant estimators, the detailed detecting and identifying algorithms are obtained for not only the pulse‐type faults but also the step‐type faults in CAR process.FindingsThis paper illustrates the useful information that can be obtained from residuals and that can be used to detect pulse‐type faults as well as step‐type faults. A fault‐tolerant recursive estimator for the coefficients of the CAR process is put forward. Using a simple transformation from step‐ to pulse‐type faults, all kinds of diagnosis methods to detect and identify step‐type faults can be used.Research limitations/implicationsFault‐tolerant estimators and fault detection and identification algorithms are aimed at abrupt faults in CAR processes.Practical implicationsMost of the algorithms given in this paper can be used in many different fields, such as process monitoring, safety control and change detection, etc.Originality/valueThis paper contributes to research of abrupt faults and abrupt changes in a CAR process and emphasizes identification of magnitudes of abrupt faults. The fault‐tolerant estimators are effective not only to detect faults but also to identify safely the coefficients CAR model.

FAULT DIAGNOSIS AND LOCATION ON MULTI-LOAD DISTRIBUTION NETWORK BY A CURRENT CLARKE–CONCORDIA PATTERN APPROACH

A new approach for fault diagnosis and location on electrical power multi-load distribution network is presented. Normally, the bus voltage in the distribution networks is almost constant even in presence of a line fault. Therefore, it is possible to identify and locate different types of faults only through the analysis of current lines. Under this consideration it is possible to convert this problem into a current pattern recognition analysis. Consequently, using a Clarke-Concordia transformation it is possible to obtain a different pattern for each fault and location. To identify those patterns an eigenvalue approach is used. A polynomial function is also used to establish the nonlinear relationship existing between fault location and line current patterns. Results are presented which show the effectiveness of the proposed algorithm for a correct fault identification and location on an electrical power multi-load distribution network.

A PARALLEL PROBABILISTIC SYSTEM-LEVEL FAULT DIAGNOSIS APPROACH FOR LARGE MULTIPROCESSOR SYSTEMS

In this paper, we present a system-level fault identification algorithm, using a parallel genetic algorithm, for diagnosing faulty nodes in large heterogeneous systems. The algorithm is based on a probabilistic model where individual node fails with an a priori probability p. The assumptions concerning test outcomes are the same as in the PMC model, that is, fault-free testers always give correct test outcomes and faulty testers are totally unpredictable. The parallel diagnosis algorithm was implemented and simulated on randomly generated large systems. The proposed parallelization is intended to speed up the performance of the evolutionary diagnosis approach, hence reducing the computation time by evolving various sub-populations in parallel. Simulation results are provided showing that the parallel diagnosis did improve the efficiency of the evolutionary diagnosis approach, in that it allowed faster diagnosis of faulty situations, making it a viable alternative to existing techniques of diagnosis. Moreover, the evolutionary approach still provide good results even when extreme non-diagnosable faulty situations are considered.

A Fault Identification Algorithm for Satellite Networks Based on System Level Diagnosis

It is a challenge to support the fault tolerance for satellite networks, in which fault identification is primary. After modeling satellite networks with two-level-node graph, a fault identification algorithm based on PMC test invalidation model is presented and proved to be correct. The effectiveness of the algorithm in different types of satellite networks is compared and studied by simulations. The results of experiments illustrate that the

Fault Detection and Identification in an IM270 Gas Turbine Using Measurements for Engine Control

A unique fault detection and identification algorithm using measurements for engine control use is presented. The algorithm detects an engine fault and identifies the associated component, using a gas path analysis technique with a detailed nonlinear engine model. The algorithm is intended to detect steplike changes in component performance rather than gradual change of all components. Component performance deviation (efficiency and flow rate) is represented by a magnitude and a phase. The phase is selected to minimize the error of evaluation matrices. Then the magnitude is computed. By utilizing operational data of the IM270 engine, the compressor and the turbine performance deviation was quantitatively identified.

A Methodology for Fault Detection, Isolation, and Identification for Nonlinear Processes with Parametric Uncertainties

This paper presents a novel methodology for systematically designing a fault detection, isolation, and identification algorithm for nonlinear systems with known model structure but uncertainty in p...

In-Flight Weight and Balance Identification UsingNeural Networks

Weight and longitudinal center-of-gravity position of an aircraft are valuable information for flight monitoring, automatic flight control, and fault detection and identification algorithms. In trimmed symmetric flight (as cruise, climb or descent), the aircraft weight is counteracted mainly by the aerodynamic lift, whereas the associated couple is counterbalanced by a moment generated by an appropriate elevator deflection. Since, for given flight conditions, the aerodynamic lift and pitching moment are functions of the angle of attack and elevator deflection only, the last two variables carry the information about the aircraft weight and its longitudinal center-of-gravity position. It is demonstrated that this information can be effectively recovered in flight using non-parametric artificial neural networks (NNs), pre-trained on flight test data. To enhance the accuracy and speed-up the training process, basic flight mechanics relations are incorporated into the NN. It is also demonstrated that the NN-based lift and pitching moment functions can be used a posteriori to estimate the aircraft longitudinal neutral point.

Number of mutual connections in neighborhoods and its application to self-diagnosable systems

This paper gives a set of discriminators (sufficient conditions) to identify the possibility of embedding optimal, highly structured systems on regular graphs for the purpose of self-diagnosis. The discrimination process is only a set of straightforward matrix computations; however, the discriminator is powerful and leads to the result that most of the typical graphs representing interconnection networks such as Cayley graphs can be embedded by the highly structured systems. The highly structured system has an O(| E|) fault-identification algorithm that can diagnose each of the units independently, locally and in any order, where | E| means the cardinality of the set of edges.

Multiple-Model Based Fault Detection and Diagnosis of Internal Sensors in Mobile Robot. Detection and Identification of Sensor-Scale Failure via Kalman Filter.

This paper proposes a multi-model based approach to fault detection and diagnosis of internal sensors in mobile robot for a robust dead reckoning. A scale failure, at which the scale of sensor outputs differs from the normal, is handled as the failure type. The scale factor of the sensors as well as the robot velocity is estimated with single-model based Kalman filters, each of which is based on a model matching to a failure mode of particular sensors. The model-based estimates of the scale factors are compared with each other, and then fault decision is made. The proposed fault detection and identification algorithm is implemented on our skid-steered mobile robot with five internal sensors (four wheel-encoders and a yaw-rate gyro). Fifteen fault modes are modeled to correctly diagnose the failure of any one of the five sensors. Experimental results show the property of the fault detection and identification algorithm.

Optimal stochastic fault detection filter

A fault detection and identification algorithm, called optimal stochastic fault detection filter, is determined. The objective of the filter is to detect a single fault, called the target fault, and block other faults, called the nuisance faults, in the presence of the process and sensor noises. The filter is derived by maximizing the transmission from the target fault to the projected output error while minimizing the transmission from the nuisance faults. Therefore, the residual is affected primarily by the target fault and minimally by the nuisance faults. The transmission from the process and sensor noises is also minimized so that the filter is robust with respect to these disturbances. It is shown that the filter recovers the geometric structure of the unknown input observer in the limit where the weighting on the nuisance fault transmission goes to infinity. Further, the asymptotic behavior of the filter near the limit is determined by using a perturbation method. Filter designs can be obtained for both time-invariant and time-varying systems.

Fault Identification in System-Level Diagnosis: a Logic-Based Framework and an O(n 2\sqrt{τ}/{\log n}) Algorithm

Much research has been devoted to system-level diagnosis—SLD. Two issues have been addressed. The first of these is diagnosability, i.e., provide necessary and sufficient conditions for a system of n units to be diagnosable provided that the number of faulty units does not exceed τ . The second is the design of fault identification algorithms, assuming that the system being considered is diagnosable. This paper focuses on the second of these concerns, discussing several algorithms of which the most efficient runs in O(n 2.5 ) . By considering a logical framework, this paper investigates the process of fault identification and proposes a fault identification algorithm which runs in O( n 2 \sqrt τ / \sqrt log n ) , τ < n/2 .

Application of Information System Theory for Actuators Diagnosis

For actuator assembly consisting of: control valve, pneumatic linear servomotor and positioner the set of possible faults was defined. Fault related algorithms are given. For description of the faults-symptoms relations the information system theory was applied. The fault identification algorithms basing on the information system rules are given.

A generalized least-squares fault detection filter

A fault detection and identification algorithm is determined from a generalization of the least-squares derivation of the Kalman filter. The objective of the filter is to monitor a single fault called the target fault and block other faults which are called nuisance faults. The filter is derived from solving a min–max problem with a generalized least-squares cost criterion which explicitly makes the residual sensitive to the target fault, but insensitive to the nuisance faults. It is shown that this filter approximates the properties of the classical fault detection filter such that in the limit where the weighting on the nuisance faults is zero, the generalized least-squares fault detection filter becomes equivalent to the unknown input observer where there exists a reduced-order filter. Filter designs can be obtained for both linear time-invariant and time-varying systems. Copyright © 2000 John Wiley & Sons, Ltd.

Powertrain Diagnostics Using Nonlinear Sliding Mode Observer

A model based fault detection and identification (FDI) algorithm for an internal combustion engine is proposed. A low frequency dynamic engine model is developed for the model-based FDI application and this model considers the air and the fuel dynamics of an IC engine including the effect of the exhaust gas re-circulation (EGR). Nonlinear sliding mode observers are used to estimate the outputs and inputs of the air and fuel dynamics subsystems, using the measurement obtained from production sensors, including a switching oxygen sensor. The observers are emmbedded in a nonlinear parity equation generation (NPERG) algorithm for fault detection and isolation. Experimental results demonstrate the promising performance of the designed observers and the effectiveness of monitoring scheme in isolating both actuator and sensor faults.

Fault Detection and Identification Using Neural Network and Fuzzy Logic

A fault detection and identification (FDI) algorithm for dynamic systems based on neural network and fuzzy logic is proposed. The algorithm comprises two modules. The first module is a typical artificial neural network used for estimation of system variables. Fault detection is achieved by comparison between the estimated and the measured values of the variables. The second module is a fuzzy neural network used for identification of the detected fault, into which the difference between the estimated and the measured values of the variables are fed. The results of a simple case study suggest that the proposed algorithm has potential applications in the field of fault detection and identification.

Comments on "An O(n/sup 2.5/) fault identification algorithm for diagnosable systems"

For original paper see A.T. Dahbura and G.M. Masson, "An O(n/sup 2.5/) fault identification algorithm for diagnosable systems," IEEE Trans. Comput., vol. C-33, pp. 486-492, Jun. 1984. Dahbura and Masson suggested an alternative version of Hakimi-Amin's theorem about the necessary and sufficient conditions for the system, described by PMC model (1967), to be t/sub p/-diagnosable. The authors found systems with connection assignment that satisfy conditions of Dahbura-Masson's theorem but are not t/sub p/-diagnosable. In this correspondence they state the additional condition under which Dahbura-Masson's theorem will hold.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A comparison-based approach to multicomputer system diagnosis in hybrid fault situations

A comparison-based system-level fault diagnosis model is considered. Provided the number of faulty units and the number of incorrect outcomes (in the sense that a fault-free unit evaluates a faulty one to be fault-free) do not exceed given bounds, the necessary and sufficient conditions for the comparison assignment to achieve correct and complete diagnosis are determined. An O(n mod E mod ) fault identification algorithm and optimal comparison assignments, in terms of minimum time of complete fault localization, are also proposed. The proposed diagnosis strategy can be performed concurrently with application tasks execution and the suitable comparison assignment can minimize the latency of fault identification. It makes the model more attractive for real-time applications.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A parallel fault identification algorithm

Consider a system of n units, at most t < n 2 of which are faulty. Suppose each unit is capable of testing any other unit for a possible fault, but that results of tests performed by faulty units might be incorrectly or even maliciously reported. We give an adaptive algorithm which identifies the faulty units in O(1 + log [ n t ] t) rounds of testing using O( n) tests. Previous adaptive algorithms required at least t + log n rounds of testing and n + t − 1 tests. If t < n 1− e , our algorithm runs within O(1) rounds.