基于RBF神经网络的机械设备油液质量检测方法研究

Research on oil quality detection method of mechanical equipment based on RBF neural network

针对传统机械设备油液质量检测方法检测周期长的问题,设计一个基于RBF神经网络的机械设备油液质量检测方法。利用原子发射光谱测量原理对油液预处理,并根据光吸收定律求得待测油液中元素含量,在此基础上,根据RBF神经网络结构,计算油液样本密度指标,利用伪逆最小二乘法,求取RBF神经网络隐含层与输出层之间的权值,最后构建油液质量的特征向量,以特征向量判断油液质量。实验对比结果表明,在颗粒数量为20时,传统检测方法与此次设计检测方法检测周期相差5 min,但当颗粒为60时,此次设计的基于RBF神经网络的机械设备油液质量检测方法检测周期为5 min,传统检测方法检测周期为30 min,比传统方法检测周期短25 min,证明此次设计的检测方法检测周期短,能够满足机械设备油液质量检测的实时性要求。

Aiming at the long detection period of traditional oil quality detection methods for mechanical equipment,a method of oil quality detection for mechanical equipment based on RBF neural network is designed.Based on the principle of atomic emission spectrometry and the law of light absorption,the element content in the oil to be measured is obtained.On this basis,according to the structure of RBF neural network,the density index of oil samples is calculated,and the weight between the hidden layer and the output layer of RBF neural network is calculated by pseudo-inverse least square method.Finally,the structure of RBF neural network is constructed.The eigenvectors of oil quality are constructed to judge the quality of oil by eigenvectors.The experimental comparison results show that when the number of particles is 20,the detection period of the traditional method is 5 minutes different from that of the designed method,but when the particle is 60,the detection period of the mechanical equipment oil quality detection method based on RBF neural network is 5 minutes,and the detection period of the traditional method is 30 minutes,which is better than that of the traditional method.The detection period is 25 minutes short,which proves that the detection method designed in this paper has a short detection period and can meet the real-time requirements of oil quality detection of mechanical equipment.