Study Of Identifying Vehicle Noise Source With Partial Coherence Function

Now, noise is one of the social effects of pollution. Controlling vehicle noise has become the problem of increasingly attentive. How to reduce vehicle noise becomes a pop topic of automobile industry. At the same time, Magnitude of vehicle noise is importance index of automobile quality level. Controlling vehicle noise is an importance problem of automobile industry in the world.Through study of controlling vehicle's noise and vibration, it's important to select an effective noise testing method. This thesis analyze noise source and principle of vehicle, then deeply study partial coherence function to identify vehicle noise, basing on the analysis of the primary methods in identifying vehicle noise home and abroad.This thesis particularly analyzes the principle of coherence technology and the coherence analysis of the system based on the definition of coherence function. Then using power chart and conditional power chart, the formula of partial coherence function in identifying noise sources is obtained.In analysis of partial coherence function, using partial coherence function to identify noise sources is discussed, and gives detailed frame chart of partialcoherence analysis. Then the observing problem is analyzed, in order to improve calculational precision and avoid ignoring the important effect noise source.In this thesis, an identifying vehicle noise source model that use partial coherence function is founded. Because of the complexity of vehicle noise, noise of automobile's mostly phonation parts qua input, vehicle noise qua output, and then found a multi-input/single-output system model.At the same time, a simulating system of identifying noise source is founded by using MATLAB to validate the research of partial coherence function in identifying noise source theoretically. Using this system, high efficiency can be achieved;the experiment result can be predicted;parameters of measure system can be determined before experiments. In this thesis, the model with four-inputs/single-output is detailed analyzed by this simulating system. The primary noise source of the model is confirmed, and the contribution of each noise source is compared and analyzed.In this thesis, it is proved theoretically that using the method of partial coherence function to identify noise source is effective and feasible.