Determine The Method Of Nonlinear Vibration Isolation System Parameters Based On Neural Networks

Vibrations always keep company with us and bring human and machinery a lot of trouble. In engineering of boosters, researchers have been puzzled by vibrations at all time. For passenger transports, how to control the vibration that is in sensitively human frequency scope preferably is the key of improving ride comfort. On the condition that sleeper passenger vehicles have been employed broadly, ride comfort of passengers has a great improvement, but the driver's one hasn't. So the project, which is equipping a proper anti-vibration component in driver's chairs of long-distance passenger vehicles to control the vibration characters of sensitively human frequency scopes and to improve driver's ride comfort, has a great practicability and feasibility.At present the research about single-frequency or multi-frequencies anti-vibration, which is about single-degree of freedom or two-degrees of freedom of the linear system, has become perfect. But in nonlinear anti-vibration domain, there are a lot of questions that have not been solved, especially about the modeling of the nonlinear anti-vibration system. Therefore it has active significance to explore the effective method of modeling of the nonlinear anti-vibration system to control the vibration characters of sensitively human frequency scopes.Based on the above reasons, a research about modeling of the anti-vibration system used by BP (back-propagation) Neural Network is carried out. During the research, firstly, testing samples and learning samples of Neural Network have been obtained by numerical simulations and experiments respectively, and then the samples were trained by BPNN to model the anti-vibration system. The following results are got: 1) An sensible evaluation about the connection between efficiency of linear or nonlinear anti-vibration systems and variation rules of system parameters is obtained by analyzing the results of simulations and experiments. 2) A relatively ideal result is obtained by using BPNN to model the anti-vibration system.The results show that the research conclusions can be used direct the designs of relative anti-vibration systems including the components of driver chairs.