摘要
提出一种α稳定分布基函数概率神经网络(A lpha-PNN)结构,该网络隐层的神经元激活函数采用了对称α稳定分布的概率密度函数,和常规的高斯分布函数相比,其具有更好的可变性和延展性,从而使隐层的神经元在函数近似上具有更高的适应性,同时也克服了概率神经网络对输入数据的独立同分布假设,提高了神经网络对局部脉冲突变的近似能力。在此基础之上,提出了一种新的根据系统输入输出数据实现的故障诊断算法,并将其应用到轴承的故障诊断中。仿真结果表明,在有色噪声的背景下,该算法仍然能够实现较好的识别效果,故障的误报率低于概率神经网络方法。
A modified probabilistic neural network named alpha-stable distributions basis function probabilistic neural network（Alpha-PNN） is proposed.The activation functions of network hidden neurons adopt probability density function of symmetric alpha-stable distributions.Compared with routine gauss distribution function,it has better variability and tractility,so hidden neurons have high adaptability in terms of function approximation.Meanwhile,it overcomes the assumption that input data is independent and identically distributed.And it also improves neural network approximation ability of partial pulse burst.A fault diagnosis algorithm based on input data and output data is proposed,which is applied to the bearing fault diagnosis.The simulation results indicate that this algorithm can achieve good recognizing effects and have low false positive ratio than PNN with the assumption of colored noise.
出处
《控制工程》
CSCD
北大核心
2010年第6期824-827,840,共5页
Control Engineering of China
基金
中国博士后基金资助项目(2008043167)
江苏省博士后基金资助项目(07C2008)
关键词
Α稳定分布
基函数
概率神经网络
故障诊断
有色噪声
alpha-stable distributions
basis function
probabilistic neural network
fault diagnosis
colored noise
