Study Of Blind Extraction Algorithm And Experiment Of Automobile Brake

Vehicle foundation brake will produce severe vibration and noise during braking, it will influence the comfort and reduce the life of automotive components. Meanwhile, the sharp brake noise (screaming) will seriously interfere with people’s normal life.For the past few years, blind signal processing which could recovery or estimate source signals from the mixed signals without any prior knowledge provides an effective solution for brake noise signal extraction. But traditional blind signal processing algorithms are usually not suitable to solve mechanical signals. Therefore, the present study initially established frequency-domain brake unnormal signal blind processing model and algorithms based on the theory of blind signal processing, focus on extraction under practical operation condition, and provides feasible methods for extractiong brake abnormal sound features.Firstly, from the application of engineering practice, the selected topic background and research significance are introduced, the blind signal processing and the research status of brake abnormal sound signal processing are reviewed and the existing problems in application of blind signal processing are summarized. Then several research points are studied in detail, the main contents are given as follows:(1)According to the fact that all kinds of vibration and sound are mixed in industrial field, the theory frame of frequency-domain mechanical fault signal processing is established through the analysis of pretreatment methods and frequency-domain blind signal processing algorithm. Simulation verified the effectiveness.(2)Based on practical operation condition, the improved frequency blind extraction algorithm aimed at mechanical convolution signals and the mechanical signal constraint semi-blind extraction algorithm are proposed. The experiment results carried out in practical environment demonstrate the effectiveness of single fault feature extraction.(3)A mechanical signal blind extraction and fault diagnosis system is developed by using LabVIEW and Matlab mixed programing based on theoretical research. The effect of this system was verified by the rolling bearing and gearbox fault diagnosis experiments. It has positive effect for preventing components failure, improving equipment running reliability and safety, decreasing of fault loss level.