摘要
为实现宽频范围内对发动机振动进行积极隔振,提出了一种混合模式的磁流变悬置结构。以磁流变悬置动态性能试验结果为数据样本,利用BP神经网络分别对磁流变悬置正、逆模型进行辨识;同时,建立了三点磁流变悬置系统6自由度模型,设计了基于磁流变悬置BP神经网络正、逆模型的模糊控制器,对磁流变悬置系统进行隔振控制,仿真结果表明:BP神经网络具有较高的准确度,在磁流变悬置模型辨识方面具备优越性能;模糊控制器均能较好地在宽频范围内的衰减发动机振动,发动机稳定转速下对应的位移和加速度振动峰值明显减小。
In order to resolve noise and vibration problems generated by vehicle engine, a new mixed mode MR mount is proposed. The direct and inverse models of the MR mount were identified by dynamic performance test results with BP neural network. Meantime, a six degrees of freedom dynamic model of an in-line four-cylinder engine which has three points mounting system is derived by considering the dynamic direct and inverse models of MR mount and its state space form is established. The fuzzy control strategy was adopted for the semi-active mounting model. The results showed that the fuzzy control strategy has better performance in a wide frequency range, which, the displacement and acceleration amplitude reduced significantly. The BPNN network inverse model and the variable universe fuzzy control a strategy were correct and effective.
出处
《科技创新导报》
2016年第9期53-58,共6页
Science and Technology Innovation Herald
基金
四川省教育厅科研项目资助(项目编号:15ZB0400)
