In this study, modeling and fault detection of a novel faulty quadrotor is presented. It is assumed that a quadrotor vehicle has encountered a fault during a flight accident, and as a result, one of the rotors does not operate vertically. Although the rotor's rotational axis has deviated from the vertical direction, the amount of produced thrust remains constant. Detecting this fault along with utilizing a proper controlling approach can reduce the risk of failure in the vehicle. Based on this statement, the procedure of this study has been developed in three main stages. First, the kinematic and dynamic equations governing the faulty system are driven using Newton's second law and Euler's principle. Then, equations governing the faulty system and the Thau observer are employed to calculate the residual value. This parameter is calculated based on the differences between states’,measurement and estimation. Eventually, by comparing the computed residual value with the assumed threshold, thrust deviation in the shortest possible time has been detected.

The problem of optimal sensor location for fault detection in Linear Time-Invariant (LTI) systems is considered. A novel characterization of undetectable faults is shown to give rise to a new optimal sensor location problem for fault detection in the presence of L2 disturbances. An algorithm is presented whereby suboptimal solutions to this problem can be determined. A numerical example is presented to illustrate the theory and execution of the algorithm.

System faults, which usually lead to changes in critical system parameters or even system dynamics, may lead to degraded performance and unsafe operating conditions. fault detection plays an important role in ensuring system safety and reliability for unmanned aerial vehicles. Artificial neural networks have an excellent potential to detect and isolate faults in complex processes. In this paper, an observer based on an adaptive neural network is presented. In this study, the adaptive neural network is designed as an intelligent learning system to detect and isolate sensor and actuator faults in a nonlinear dynamic model of an unmanned aerial vehicle. Due to the system's nonlinearity, the neural network's weighting parameters are updated using the extended Kalman filter, which increases the convergence rate of the neural network. A set of abrupt, intermittent and incipient faults are applied to a nonlinear dynamic model of a tilt rotor to evaluate the method. Due to the high update rate of neural network weighting, the proposed method can detect abrupt, intermittent and incipient faults accurately and quickly. Numerical simulation results are also given to show the performance of the proposed method, which shows the proper performance of this design.

Vehicle driveline system and its working accuracy play an important role in the performance of car. The purpose of this study is to provide an appropriate mechanism for investigating, identifying and determining the position and size of defects in the vehicle power transmission system. This is based on the patterns of the residual signal, obtained from a simulated model of the system. Neuro-fuzzy networks have been used in diagnosis of defects because of its specific advantages and capabilities in pattern recognition. Simulation results demonstrate that the resulting fault detection system is able to properly locate the fault types under all test conditions, and is sensitive also to fault size. Test and simulation results using MATLAB software is given at the end.

Bearings are the most important and most used components in different industries. Early bearing fault diagnosis can prevent human and financial losses. One of the best methods for fault diagnosis of these elements is via vibration analysis. In this paper Empirical Mode Decomposition (EMD) which is a fairly new signal processing method of nonlinear and nonstationary signals is used for analyzing vibration signals extracted from bearings. This method was proposed by Huang in 1998. In this research, extracted signal from healthy and faulty bearings are decomposed in to empirical modes. By analyzing different empirical modes from 8 derived empirical modes for healthy and faulty bearings under different load conditions from zero to three horsepower, the first mode has the most information to classify bearing condition. From the first empirical mode six features in time domain were calculated for healthy bearing, bearing with inner race fault, bearing with outer race fault and bearing with ball fault. These eight features were used as input vector to a designed ANFIS network for bearing condition classification. The ANFIS network was able to detect different condition of bearing with 100% precession.