Adaptive Threshold-based Fault Detection for Systems Exposed to Model Uncertainty and Deterministic Disturbance

Fault detection in a grid-connected photovoltaic system using adaptive thresholding method

Fault detection for linear discrete-time periodic systems

Suspension operation reliability is one of the most significant performance indexes that concerns maneuvering stability and drive safety. In this paper, in order to guarantee good suspension reliability, an adaptive fault detection and isolation scheme is proposed for quarter-car active suspension systems with parametric and nonlinear uncertainties. To realize fault diagnosis for active suspensions, an adaptive fault detection estimator and several fault isolation estimators are designed to generate state residuals. Corresponding adaptive thresholds are developed to help judge the occurrence and type of possible faults. In the process of constructing state estimators, the uncertain parameter is updated online so that good sensitivity of fault detection can be achieved. Illustrative simulation is carried out to validate the effectiveness of the fault diagnosis scheme, and results indicate that the proposed method is sensible to sudden faults and maintain robustness to model uncertainties appearing in suspension systems.

In this paper, the fault detection problem is investigated for a class of wireless networked systems with compressed measurements. A new compressed fault detection framework is proposed. The measurement data is compressed with a compressed sensing technique before being transferred to the fault detection node via wireless communication channel. Besides, the influence of compressed measurements on fault detectability is analyzed and a new fault detection method is provided based on the integration of model-based and statistical methods. The residual generator is designed by the available compressed measurements and priori knowledge about the system, and the statistic testing method is developed to determine whether there is a fault. Maintaining the acceptable fault detection performance, the compressed fault detection framework can greatly reduce the cost of data acquisition, transmission and storage. As the compressed data is utilized directly in the fault detection process, our method can also reduce computational requirements and enhance the efficiency of the whole fault detection system. Finally, a simulation study is carried out to demonstrate the effectiveness and applicability of the proposed method.

An approach to automatic fault detection in four-point system for knitted fabric with our benchmark dataset Isl-Knit

Various faults occurred in the heating, ventilation and airconditioning (HVAC) systems usually lead to more energy consumption and worse thermal comfort inevitably. This paper presents a feasible and valid solution of HVAC fault detection and diagnosis (FDD) problem based on statistical machine learning technology. It learns the consistent nature of different types of faults of HVAC operation based on support vector machine (SVM), and then identify types of fault in all subsystems using the statistical relationships between groups of measurements. In order to speed up the learning process, principle component analysis (PCA) has been applied to compress the training data. Our approach models the dynamical sub-systems and sequence data in HVAC system. The learnt models can then be used for automatic fault detection and diagnosis. The approach has been tested on commercial HVAC systems. It had successfully detected and identified a number of typical AHU faults.

In this paper, a networked observer-based fault detection (FD) system is presented. First, the influence of noisy data transmission on the FD performance is analyzed and modeled. Based on it, a design scheme is developed for the construction of observer and residual evaluation unit. The core of this scheme is the integrated design of the coding strategy and FD system. The resulting FD system ensures high fault detectability and low false alarm rate.

This paper deals with the problem of fault detection and isolation (FDI) of a networked control system (NCS). In this configuration a linear time invariant(LTI) system communicates with controller/FDI over a shared medium. The medium supports a limited number of simultaneous connections between sensors and FDI system. Furthermore, it is assumed that there is the random and unknown packet dropouts in the controller-to-actuator link. Objective is to find a strategy for residual signals generating which network effects are taken into account. In addition, by means of these residuals fault detection and isolation of system is possible. A set of structured residual based on parity relation approach is proposed. It guarantees robustness to system disturbances and packet dropout while each residual is designed to be sensitive to some sensor faults. The efficiency of the proposed approach will be illustrated with a simulation example.

A novel fault detection system taking into account uncertainties in the reconstructed signals

A fault detection and diagnosis method for the hydraulic servo system based on adaptive threshold and self-organizing map (SOM) neural network is proposed in this study. The nonlinear, time-varying, fluid-solid coupling properties of the hydraulic servo system are considered. Fault detection is realized based on a two-stage radial basis function (RBF) neural network model. The first-stage RBF neural network is adopted as a fault observer for the hydraulic servo system; the residual error signal is generated by comparing the estimated observer output with the actual measurements. To overcome the drawback of false alarms when the traditional fixed fault threshold is used, an adaptive threshold producer is established by the second-stage RBF neural network. Fault occurrence is detected by comparing the residual error signal with the adaptive threshold. When a system fault is detected, the SOM neural network is employed to implement fault classification and isolation by analyzing the features of the residual error signal. Three types of common faults are simulated to verify the performance and effectiveness of the proposed method. Experimental results demonstrate that the proposed method based on adaptive threshold and SOM neural network is effective in detecting and isolating the failure of the hydraulic servo system.

A practical fault detection approach for PV systems intended for online implementation is developed. The fault detection model here is built using artificial neural network. initially the photovoltaic system is simulated using MATLAB software and output power is collected for various combinations of irradiance and temperature. Data is first collected for normal operating condition and then four types of faults are simulated and data are collected for faulty conditions. Four faults are considered here and they are: Line to Line faults with a small voltage difference, Line to line faults with a large voltage difference, degradation fault and open-circuit fault. This data is then used to train the neural network and to develop the fault detection model. The fault detection model takes irradiance, temperature and power as the input and accurately gives the type of fault in the PV system as the output. This system is a generalized one as any PV module datasheet can be used to simulate the Photovoltaic system and also this fault detection system can be implemented online with the use of data acquisition system.

Fault detection systems are typically applied in the railway industry to examine the structural health status of the wheel/rail system. We herein propose a time-domain kurtosis beamforming technique using an array of microphones for the fault identification and localisation of the wheel/rail system under an environment with high background noise. As an acoustics-based noncontact diagnosis method, this technique overcomes the challenge of the contact between the sensors and examined structures, and it is more applicable for impulsive signals of broadband nature, such as impact noise generated from faults on the wheel surface. Moreover, the application of kurtosis enables the identification and localisation at low signal-to-noise ratio. Under such circumstance, the impulsive signals generated by faults were totally merged in rolling noise and background noise. Meanwhile, different types of faults on the wheels could be identified and localised by observing the kurtosis value on the beamforming sound map. The effectiveness of the proposed method to diagnose the type of wheel fault with low signal-to-noise ratio and moving source has been validated experimentally. This method may provide a useful tool for the routine maintenance of trains.

UKF-based fault detection for nonlinear stochastic systems with strong noise and communication protocols.

This paper investigates the problem of fault detection for discrete linear systems subjected to unknown disturbances, actuator and sensor faults. A bank of Augmented Robust Three stage Kalman filters is adapted to estimate both the state and the fault as well as to generate the residuals. Besides, this paper presents the evaluation of the residuals with Bayes test of binary hypothesis test for fault detection to adaptive threshold compared with fixed threshold. This test allow the detection of low magnitude faults as fast as possible with a minimum risk of errors, the increase of detection probability and the reduction of false alarm probability.

In this paper, the fault detection and isolation problem for multi-agent systems with directed topology is investigated. A novel fault detection and isolation strategy is proposed. Using relative information between the agent and its neighbours, observers can be designed. It is proved that the existence conditions of observers for multi-agent system are the existence conditions of unknown input observer for the single system, which can be accomplished easily. Moreover, by the generated residual signal and the corresponding algorithm, the faulty agent can be detected simply. Compared with other results, our method decreases the number of observers and reduces the computing loads to a large extent.