| With the increase of machinery operating speed and precision degree in high-speed machinery, nonlinear problem of mechanical system which may cause abnormal state even fault directly or indirectly becomes more and more obvious. Theoretical and experimental studies focusing on this problem are very important, especially the researches of early detection of operating state and fault diagnosis which use the methods based on nonlinear dynamics.Funded by the National Natural Science Foundation of China "Theory and Methods aimed at Enhancement Detection of Transmission Train Damage of Helicopter Based on Nonlinear Dynamics" and the Excellent Youth Foundation of Hunan Scientific Committee of China "Research on Theory and Method of Early Detection and Prognosis of Fault for Mechanical Systems based on Chaos Synchronization and Array Coupled Stochastic Resonators", targeted for the early fault detection of rolling bearings, this dissertation built two kinds of weak signal detection models, include, methods based on chaos suppression in the nonresonant parametric drive and array of stochastic resonance (SR). The main contents of the dissertation include,1. With the analysis of main fault patterns and character of rolling bearing, studied the chaotic states evolution of Lorenz system, proposed the weak period signal detection method based on the sudden change of Lorenz system, when chaos suppression occurs.(1) Analyzed the evolution of chaotic states on the one orbit of parameter r of Lorenz system, and found the best chaotic state for built the detection model from the entire evolution process.(2) Built the model of detecting weak periodic signals in strong noise based on the principle of chaos suppression with nonresonant parametric drive. According to the parametric equivalent relationship obtained using averaging and renormalization method, the critical values of detection parameters are determined.(3) Discussed the influence of some factors to the model, confirmed the special means to detect weak signal, and then using this method processed the fault signal from machinery fault simulator.2. To compensate the shortage of forenamed model on the pattern of signal frequency detection, studied the enhance effect of array of SR based on the classical theory of SR, then built two kinds of array of SR.(1) Analyzed the enhance effect of SR array, based on the SR model on double-state system, built the models of coupled and uncoupled array of SR, and then made the numerical simulation, analyzed the simulated fault signal of rolling element bearing and practical fault data.(2) Compared two kinds of array of SR models with single SR oscillator. According to the results, the fact of signal-noise ratio improvement (SNRI) exceedingunity was confirmed. The enhance effect of two kinds array models is obvious. |
