Research On Intelligent Diagnosis Technology Based On Vector Spectrum

State Monitor and Fault Diagnosis is an applied technology which has strong applicability. The intellectualized technology of rotating machinery can bring huge and potential economic effects and it relates to the workers' safety in working field. The sense for this research consists in debasing incidence rate and maintenance costs, avoiding over maintain, decreasing costs and adding runtime. Traditional research in rotating machinery is based on the analysis of single channel. But as a result of the whirl phenomenon, any channel can't reflect rotor's running state in the round. Vector Spectrum is an intelligence fusion technology, it fuses the information in the two channels which come from the same section. A mass of engineering practice have proved that Vector Spectrum can reflect rotor's running state truly, and its arithmetic is so simple, in the end, it increases real-timing and stability enormously.The main work is done as follows:1. Vector Spectrum which based on congenetic information fusion are expounded in this thesis, and its fast and steadiness numeric algorithms be elicited. The inevitability of monophyletic information result in miscarriage of justice are expounded and proved. And it proved the advantage of Vector Spectrum compare to the traditional methods by experiment.2. Utilizing the method of Vector Spectrum and Rough Sets Theory to optimize fault characters in four familiar faults, such as imbalance, noncentering, oil whirl and oil whip. Through the optimum assembly to assort the fault style accurately. The experiment proved that this method can increase the veracity and real-time in fault diagnosis, and lay the foundation for the subsequent section.3. Combine Vector Spectrum and Fuzzy Synthetic Decision, and advanced Vector Spectrum-Fuzzy Synthetic Decision diagnostic method, and build the mathematical model of this method. At last, through experiment to prove the method's advantage of intuitiveness and veracity. 4. Based on the theory of Vector Spectrum and Ant Colony Algorithm, the thesis brings forward a new method Vector-clustering ACO. The clustering algorithm can assort the sample data fast, in this way, we can obtain the distance of sample data to clustering center, then make the result as the ACO's initial metric, and then use ant colony algorithm. Finally, research on the result of the experiment, it proved that this arithmetic improved the calculation speed and has upper precision.