基于GA的BP神经网络在本安参数评定中的应用

摘要

本质安全参数评定是本质安全防爆技术的核心内容.针对传统的本质安全参数评定方法存在较多局限性的现状,提出了一种基于遗传算法(GA)的BP神经网络本质安全参数评定模型.讨论了影响火花点燃能力的主要因素,建立了BP神经网络模型.该模型采用GA对神经网络的初始权值和阈值进行优化,以避免可能的局部搜索最小现象.仿真结果表明,该方法可有效预测本质安全参数.%The parameter assessment of intrinsic safety is the kernel part of intrinsic safety explosive protection technology.The traditional assessment method features many limitations, thus the parameter assessment model based on GA-BP is proposed.The main factors that impact the capability of spark ignition are discussed, and the BP neural network model is established.With this model, the initial weights and threshold values of the neural network are optimized using GA to avoid the possibility of local search minimum.The result of simulation indicates that this method can effectively predict intrinsic safety parameters.