基于神经网络的LPG泄散范围的预测方法

Predicting Leakage and Dispersion of LPG Based On Neural Network Method

提出了一种基于BP神经网络预测液化石油气泄漏扩散范围的分析方法。详细说明了神经网络的分析模型以及样本的选取和学习过程,将BP神经网络的预测结果和采用传统方法的预测结果进行对比分析,表明基于神经网络的分析预测方法不仅能够正确区分液相与气相泄漏,还能准确地预测扩散危害距离,是一种准确、方便、快捷的分析方法,具有一定的工程应用价值。

A quantitative analysis method based on BP neural network is proposed for analyzing the risk region of leakage and dispersion of LPG (liquefied petroleum gas). The BP neural network model, the specimen option and the training process are illustrated. The results from the BP neural network method and the traditional consequence analysis method are compared. It is shown that the quantitative analysis method based on BP neural network is accurate, convenient and fast not only to distinguish liquid phase leakage from gas phase leakage, but also to predict dispersion risk distance. It has engineering applicative value.