摘要
为了维持油田能够长期稳定高产,必须制订科学合理的生产方案。而常规预测方法对数据依赖大,预测精度不高。为此,在灰色预测理论的基础上引入BP神经网络模型,建立了GM(1.1)和BP神经网络组合模型。此组合模型兼有灰色预测和BP神经网络预测的优点,克服了原始数据少,数据波动性大对预测精度的影响,同时也增强了预测的自适应性。最后通过实例对比分析,说明了组合模型的有效性及可应用性。
In order to maintain stable and high output in the oil field for a long period of time, a scientific and rational production plan must be made. Based on the actual analysis of crude oil and gray estimation theory, the BP neural network model is introduced to set up GM (1. 1) and BP neural network integration model. This model has combined the advantages of gray estimation and BP neural network estimation, it has overcome the influence of little raw data and high data fluctuation to precision of estimation and also it has enhanced the self-adaptability of estimation. At last, it is proved that GM (1. 1) and BP neural network integration model is effective with examples contrasted.
出处
《石油化工自动化》
CAS
2007年第6期43-45,49,共4页
Automation in Petro-chemical Industry
关键词
神经网络
灰色理论
产量预测
模型
精度
neural network
gray theory
output estimation
model
precision