摘要
利用遗传算法(GA)-BP神经网络的模型辨识方法建立了磁流变悬置的正、逆模型;并设计了变论域模糊控制器,对发动机磁流变悬置系统进行半主动控制的仿真。结果表明:变论域模糊控制比传统模糊控制具有更好的宽频隔振效果,发动机转速对应的二阶主频位移和加速度振动峰值明显减小,验证了GA-BP网络逆模型及其变论域模糊控制算法的正确性和有效性。
The direct and inverse models for magneto-rheological (MR) mount are created by using model identification technique with genetic algorithm (GA) BP neural network, a variable universe fuzzy controller is designed, and a simulation on the semi-active control of engine MR mount system is conducted. The results show that the variable universe fuzzy control has better vibration isolation performance in a wide frequency range compared with conventional fuzzy control, and the peak amplitudes of displacement and acceleration reduce significantly at the second dominant frequency corresponding to engine speed, validating the correctness and effectiveness of GA-BP neural network inverse model and its variable universe fuzzy algorithm.
出处
《汽车工程》
EI
CSCD
北大核心
2013年第11期1023-1029,共7页
Automotive Engineering
