PNN在船用污水处理装置状态诊断中的仿真研究

Simulation research of PNN on state diagnosis of ship sewage treatment equipment

概率神经网络(PNN)是基于贝叶斯分类规则和Parzen窗函数的前向型自监督神经网络模型,具有强大的非线性处理能力,快速的收敛速度和准确的分类效果。在分析PNN基本结构和原理的基础上,采用PNN结构针对船用污水处理装置的状态诊断进行仿真,以提高船用污水处理装置状态诊断正确率。通过Matlab采用实际监测的数据实现了PNN诊断处理。理论分析和仿真结果表明:利用PNN对船用污水处理装置的状态诊断的方法可行和有效,有很高的实用价值。

Probabilistic neural network（PNN）is a feedforward self-supervisory neural network model based on Bayes classifying rules and Parzen window function and possesses powerful nonlinear processing ability,fast convergence,and accurate sorting effect.On the basis of analyzing the fundamental structure and principle of PNN,PNN structure is adopted to simulate the state diagnosis of ship sewage treatment equipment to improve the state diagnosis validity.With Matlab,the PNN diagnosis is realized utilizing the practical monitoring data from a ship.The theory analysis and simulation results indicate that using PNN on ship sewage treatment equipment for the state diagnosis is feasible,effective and with high practical value.