Hopfield神经网络在磁粉制动器设计中的应用

为了优化磁粉制动器的结构,提出了Hopfield神经网络的磁粉制动器结构设计方法。方法中利用磁粉制动器的结构建立优化模型,以磁粉制动器的总体积为目标函数,推导约束条件。根据推导的目标函数和约束条件,采用外点罚函数法,构造出优化的增广目标函数,在满足约束条件下,目标函数最小,从而得到磁粉制动器结构的设计方案。实际结果表明,优化后的制动器在满足各项指标的条件下,总体积减小了23%。

Design of the magnetic power brake using hopfield neural network

To optimize the structure of the magnetic power brake, a new method is proposed using Hopfield neural network. The optimal model is constructed by the structure of the magnetic power brake, and with the volume of the magnetic power brake as the objective function, the restrained conditions are derived. The energy function is established based on the objective function and restrained conditions. With the restrained conditions satisfied, the objective function obtained is the minimum. In this way the perfect design is obtained. The experiments show that the volume is reduced by 23% with the optimized magnetic power brake meeting all conditions.