Monitoring and diagnosis of rotating machines has become an effective and indispensable tool for the efficient and timely detection of defects, avoiding then incidents that can have serious economic and human consequences. Bearings are the most sensitive parts of these machines; this is why special attention must be paid to its monitoring. This paper presents a methodology for diagnosing an aircraft air compressor bearing using neural networks that have been optimized by genetic algorithms. We used in our study a database of vibratory signals that were recorded on a test bench from bearings with different defects. The faults features are extracted from these noisy signals using the estimate of the spectral density. The diagnostic capacity of obtained model has been demonstrated by a comparative study with two other automatic classifiers, which are discriminant analysis and neural networks whose training has been done with the Back-Propagation algorithm. This approach has the advantage of simultaneously ensuring the optimal structure of the neural network and accomplishing its learning. The importance of this study is the construction of a diagnostic tool that is characterized by efficiency, speed of decision making and ease of implementation not only on the computers on the ground, but also on the mounted calculators on aircraft.

In this paper, a sensor fault diagnosis system is proposed for an aerial vehicle (UAV) quadrotor. Given the nonlinearity of this system and the inaccuracies of modeling, adapted tools have been adopted to ensure control and diagnosis. After synthesizing control laws required for quadrotor control using the Sliding Mode Control method, a Hammerstein-Wiener model has been developed. The goal is to estimate the states of the quadrotor system and build a set of residuals to detect sensor faults. The advantage of this solution is that it does not require prior knowledge of the model and can be easily generalized to other types of vehicles. Then, to ensure decision making for fault isolation, the neural network has been combined with the diagnostic system. With judicious choice of configuration, it can efficiently classify defects from residuals. Finally, typical sensor failures have been injected during simulations. The results of the diagnosis have been very satisfactory. The model has been validated by the test data.

In this paper, we implemented a diagnostic system for vibration faults that occur on the PUMA helicopter gearbox. We used an approach based on the joint use of the Order Tracking signal analysis and the Genetic Algorithm. To achieve this goal, we first collected a database of vibration signals measured during periodic inspections. The available vibration signals are acquired under a time-varying operating conditions. Therefore, we used the Order Tracking method, which is more accurate in extracting faults features. This technique was performed by resampling the vibration data and then applying the Short Time Fourier Transform. To enable efficient and continuous monitoring of gearbox vibration faults from features, we used Genetic Algorithm to build a rules-based diagnostic model. Genetic operators have been adapted to the specificity of the problem to optimize the parameters of this model. This approach is successfully applied to the diagnosis of vibration defects of helicopter gearboxes. The results have been validated effectively with test data. The diagnostic model can therefore be implemented on helicopter computers to detect faults in flight or on the ground. This approach has been used for the first time in the field of helicopter gearbox vibration fault diagnosis.