神经网络算法在弹性限位浮筏系统中的应用

Algorithm of Neural Network in Floating Raft with Elastic Limiters

具有弹性限位器的浮筏隔振系统受冲击后,在运动的过程中会发生限位器接触状态的不断变化,由于接触位形的多样化,因此寻求一种快速决策限位器接触状态的算法是非常必要的。本文在建立含有接触力的弹性限位浮筏系统的动力学方程的基础上,利用高斯原理构造目标函数,求出神经网络状态方程,给出网络参数。通过神经网络算法对含有多个限位器的弹性浮筏系统的限位器接触状态进行决策,使得接触计算得以实现,同时还可以大大地减少CPU时间。最后通过计算分析,得到对工程有指导意义的结论。

In this paper,Based on dynamic model of floating raft with elastic limiters,using the Gauss＇s least action prin- ciple,the state equations of neural network are established and the parameters of network are obtained. The artificial neural network has been development to dingtinguish the contact state of elastic limiters at each step. The example analysis shows that the algorithm of neural network for contact can shorten the time of caclaution. From the example analysis,we get some results that may be useful to the ship engineering.