摘要
利用径向基函数神经网络可对大型机床导轨故障进行诊断,但有收敛速度慢、易陷入局部极小值等缺陷。现有生物地理优化算法具有参数少、实现简单、收敛速度快等优点,但因有搜索能力弱会导致其应用范围受到限制。该研究将利用BBO算法来训练RBF神经网络,并将其应用于龙门机床导轨的5种状态(正常、粘结、爬行、变形、磨损)的诊断。与传统诊断结果比较,迭代次数的大幅降低证明经过BBO算法训练的RBF神经网络故障诊断克服了传统神经网络寻优收敛速度慢、易陷入局部极小值等缺陷。具有较好的收敛性、稳定性,能够大幅提高故障诊断的准确度。
Radial Basis Function （ RBF） can be used to diagnose the fault of the guide rail of large machine tools,but it has some disadvantages such as slow convergence rate and easy to fall into the local minimum. The exist-ing bio - geography - based optimization （ BB0） has many advantages, such as few parameters, simple implementa-tion and fast convergence speed. However,it has the shortcomings of weak search ability,so limited. In this paper, RBF neural network is trained by using BBOalgorithm,and radial basis function neural net-work trained by bio - geo - optimization algorithm is applied to the diagnosis of five ling, deformation and wear） . Compared with the traditional radial basis function neural network,the experimentalresults showthat the radial basis function neural network trained by the bio - geographic optimization algorithm over-comes the problem that the traditional neural network has a slow convergence rate and is easy to fall into the localminimum and otlier defects. Which has good convergence and stability,and can improve the acsis to a great extent.
出处
《电子科技》
2017年第12期43-47,共5页
Electronic Science and Technology
基金
国家自然科学基金主任基金(51245009)
关键词
故障诊断
生物地理优化算法
径向基函数
算法训练
龙门机床导轨
fault diagnosis
bio - geographic optimization algorithm
radial basis function
algorithm training,gantry machine tool guide
