No description available.

This work presents a new approach for detecting and isolating faults in nonlinear processes using independent neural network models. In this approach, an independent neural network is used to model the proton exchange membrane fuel cell nonlinear systems using a multi-input multi-output structure. This research proposed the use of radial basis function network and multilayer perceptron network to perform fault detection. After training, the neural network models can give accurate prediction of the system outputs, based on the system inputs. Using the residual generation concept developed in the model-based diagnosis, the difference between the actual and estimated outputs are used as residuals to detect faults. When the magnitude of these residuals exceed a predefined threshold, it is likely that the system is faulty. In order to isolate faults in the system, a second neural network is used to examine features in the residual. A specific feature would correspond to a specific fault. Based on features extracted and classification principles, the second neural network can isolate faults reliably and correctly. The developed method is applied to a benchmark simulation model of the proton exchange membrane fuel cell stacks developed at Michigan University. One component fault, one actuator fault and three sensor faults were simulated on the benchmark model. The simulation results show that the developed approach is able to detect and isolate the faults to a fault size of ±10% of nominal values. These results are promising and indicate the potential of the method to be applied to the real world of fuel cell stacks for dynamic monitoring and reliable operations.

No description available.

Electrical machines are subject to different types of failures. Early detection of the incipient faults and fast maintenance may prevent costly consequences. Fault diagnosis of wind turbine is especially important because they are situated at extremely high towers and therefore inaccessible. For offshore plants, bad weather can prevent any repair actions for several weeks. In some of the new wind turbines synchronous generators are used and directly connected to the grid without the need of power converters. Despite intensive research efforts directed at rotor fault diagnosis in induction machines, the research work pertinent to damper winding failure of synchronous machines is very limited. This dissertation is concerned with the in-depth study of damper winding failure and its traceable symptoms in different machine signals and parameters. First, a model of a synchronous machine with damper winding based on the winding function approach is presented. Next, simulation and experimental results are presented and discussed. A specially designed inside-out synchronous machine with a damper winding is employed for the experimental setup. Finally, a novel analytical method is developed to predict the behavior of the left sideband amplitude for different numbers and locations of the broken bars. This analysis is based on the magnetic field theory and the unbalanced multiphase circuits. It is found that due to the asymmetrical structure of damper winding, the left sideband component in the stator current spectrum of the synchronous machine during steady state asynchronous operation is not similar to that of the induction machine with broken bars. As a result, the motor current signature analysis (MCSA) for detection rotor failures in the induction machine is usable to detect broken damper bars in synchronous machines. However, a novel intelligent-systems based approach is developed that can identify the severity of the damper winding failure. This approach potentially can be used in a non-invasive condition monitoring system to monitor the deterioration of a synchronous motor damper winding as the number of broken bars increase over time. Some other informative features such as speed spectrum, transient time, torque-speed curve and rotor slip are also found for damper winding diagnosis.

The purpose of this research is to address the problem of fault diagnosis of sensors which measure a set of direct redundant variables. This study proposes: 1. A method for linear senor fault diagnosis 2. An analysis of isolability and detectability of sensor faults 3. A stochastic method for the decision process 4. A nonlinear approach to sensor fault diagnosis. In this study, first a geometrical approach to sensor fault detection is proposed. The sensor fault is isolated based on the direction of residuals found from a residual generator. This residual generator can be constructed from an input-output model in model based methods or from a Principal Component Analysis (PCA) based model in data driven methods. Using this residual generator and the assumption of white Gaussian noise, the effect of noise on the isolability is studied, and the minimum magnitude of isolable fault in each sensor is found based on the distribution of noise in the measurement system. Next, for the decision process a probabilistic approach to sensor fault diagnosis is presented. Unlike most existing probabilistic approaches to fault diagnosis, which are based on Bayesian Belief Networks, in this approach the probabilistic model is directly extracted from a parity equation. The relevant parity equation can be found using a model of the system or through PCA analysis of data measured from the system. In addition, a sensor detectability index is introduced that specifies the level of detectability of sensor faults in a set of redundant sensors. This index depends only on the internal relationships of the variables of the system and noise level. Finally, the proposed linear sensor fault diagnosis approach has been extended to nonlinear method by separating the space of measurements into several local linear regions. This classification has been performed by application of Mixture of Probabilistic PCA (MPPCA). The proposed linear and nonlinear methods are tested on three different systems. The linear method is applied to sensor fault diagnosis in a smart structure and to the Tennessee Eastman process model, and the nonlinear method is applied to a data set collected from a fully instrumented HVAC system.

Abstract A powerful tool for bearing time series feature extraction and classification is introduced that is computationally inexpensive, easy to implement and suitable for real-time applications. In this paper the proposed technique is applied to two rolling element bearing time series classification problems and shown that in some cases no data pre-processing, artificial neural network or nearest neighbour approaches are required. From the results obtained it is clear that for the specific applications considered, the proposed method performed as well as or better than alternative approaches based on conventional feature extraction.

Fault detection and diagnosis is quite important in engineering systems, and deserves further attention in view of the increasing complexity of modern machinery. Traditional single-sensor methods of fault monitoring and diagnosis may find it difficult to meet modern industrial requirements because there is usually no direct way to measure and accurately correlate a machine fault to a single sensor output. Fusion of information from multiple sensors can overcome this shortcoming. In this thesis, a neural-fuzzy approach of multi-sensor fusion is developed for a network-enabled remote fault diagnosis system. The approach is validated by applying it to an industrial machine called the Iron Butcher, which is a machine used in the fish processing industry for the removal of the head in fish prior to further processing for canning. An important characteristic of the fault diagnosis approach developed in this thesis is to make an accurate decision of the machine condition by fusing information from different sensors. First, sound, vibration and vision signals are acquired from the machine using a microphone, an accelerometer and a digital CCD camera, respectively. Second, the sound and vibration signals are transformed into the frequency domain using fast Fourier transformation (FFT). A feature vector from the FFT frequency spectra is defined and extracted from the acquired information. Also, a feature based vision tracking approach—the Scale Invariant Feature Transform (SIFT)—is applied to the vision data to track the object of interest (fish) in a robust manner. Third, Sound, vibration and vision feature vectors are provided as inputs to a neuro-fuzzy network for fault detection and diagnosis. A four-layer neural network including a fuzzy hidden layer is developed in the thesis to analyze and diagnose existing faults. By training the neural network with sample data for typical faults, faults of five crucial components in the fish cutting machine are detected with high reliability and robustness. Alarms to warn about impending faults may be generated as well during the machine operation. A network-based remote monitoring architecture is developed as well in the thesis, which will facilitate engineers to monitor the machine condition in a more flexible manner from a remote site. Developed multi-sensor approaches are validated using computer simulations and physical experimentation with the industrial machine, and compared with a single-sensor approach.

No description available.

No description available.

No description available.