A Study On Power Machinery Fault Diagnosis Based On Wavelet Analysis And Neural Networks

As modern mechanical equipment becomes larger, more complicated, more automated and more integrated with each passing day, people have realized that once the equipment breaks down, it will have a unimaginably large impact on production and people's lives as well as the safety of human life and property. Therefore, people have always sought to set up monitoring, early-warning, fault-tolerance and maintenance systems to function alongside equipment systems throughout their life spans in order to guard against and even put a stop to the occurrence and development of equipment failures that hamper the normal operation of the machinery.Wavelet packet analysis is a way of handling signals based on multi-resolution analysis, capable of distinguishing each wave frequency to the desired level of detail. It can not only resolve low-frequency signals further, it can also do the same to high-frequency signals, and so has broad application in the diagnosis of machine faults. In this paper, an improved wavelet packet threshold has been designed as a de-noising method. Employing this method, the source program can filter the noise picked up in the vibration signal, and then extract the normalization eigenvector of the machine fault diagnosis.Owing to their strong sensitivity to non-linear reflection, neural networks are particularly suited to complex pattern recognition, and have therefore become a powerful tool for recognition of the status of power machinery. In this paper, the core part of the neural network, namely the three-layer BP neural network, is used to identify and classify machine fault types. On the basis of experimental demands and problem-solving requirements, it was ultimately determined that the neural network needed 8 input neurons, 30 middle-layer neurons and 5 output neurons.A self-adaptive machine fault diagnosis system was designed on the basis of experiments on the 4100QB diesel engine produced by the Kunming Yunnei Power Co. Ltd., based on wavelet analysis and neural networks. This system picks up the vibration signal of the diesel engine cylinder head, and filters the vibration noise through wavelet analysis, effectively eliminating noise interference in the vibration signal. It extracts the eigenvector of the vibration signal that indicates the diesel engine fault, and then uses the eigenvector extracted from the vibration signal coming from the engine's cylinder as a neural network experimental sample, eventually building a self-adaptive machine fault diagnosis system. By inputting a test sample into the neural network's self-adaptive machine fault diagnosis system to conduct a verification of the system, it showed that the system can effectively identify and classify the machine fault, and ultimately achieve a fault diagnosis.In the experiment, DASP2005 professional software was used to acquire the vibration signal of the engine's cylinder head under both normal and abnormal engine operating conditions. The MATLAB wavelet analysis source program was applied to filter the vibration signal to extract the relevant eigenvector to serve as a experimental and verification sample for the neural network. By studying and identifying the BP neural network, we were finally able to classify the relevant operational status and determine the engine's corresponding machine fault. These methods and experimental data have provided a foundation for further study.The results of the experiment and analysis show that it is effective and feasible to diagnose power machinery faults by employing wavelet analysis and neural networks, providing a new way of thinking about and approaching the determination and diagnosis of power machinery faults.