应用BP神经网络预测原油含水率问题的研究

Study of water containing forecasting of crude oil based on BP neural network

原油含水率是一个复杂的随机系统,本文根据射频原油含水分析仪的非线性问题(双值性),以及原油含水率与其影响因素之间存在的映射关系,建立了一个BP人工神经网络模型,并将其用于原油含水率的预测,给出BP算法动态演化过程的训练调试界面,并在算法上进行改进。实例表明,该模型软件预测精度高,有一定推广价值。

Containing is a complicated stochastic system.A BP neural network model was established to forecast based on relstionship mapping between water containing in crude oil and its influential factors.This also included training and interface of dynamic evolvement process of BP algorithmic.The results indicate to use back propagation neural networks to perdict water containing has a high predicable ability and credibity ,It is fessible in technology.