摘要
针对油田注水站机组现有机械设备维护方式经济性差、不能完全避免事故发生的问题 ,提出用具有遗传功能的神经网络对油田注水站机组运行状态进行预测 ,给出了具有反馈功能的预测神经网络结构图 ,采用一种新型的全局随机优化搜索算法的遗传算法来训练前向神经网络。根据从大庆油田采回的现场数据建立的神经网络模型所做预测结果表明 ,遗传算法是一种启发式搜索 ,易收敛于全局最优 ;收敛速度方面遗传算法明显优于BP算法 ,预测精度明显优于常规预测方法 ,遗传算法能较好地反映机组运行状态的变化趋势。
In view of the problems with the maintenance of the equipment for water injection station, it is proposed to adopt the neural network with genetic function to predict the running status of the equipment. The structure diagram of the neural network is offered. A new genetic algorithm is used for training forward neural network. A neural network model is established based on the data acquired from Daqing Oilfield, and the running status trend of the equipment is predicted according to the model. The result of the trend prediction shows that the genetic algorithm is preferable to the conventional predicting method in predicting the running status of the equipment for oilfield water injection station.
出处
《石油机械》
北大核心
2000年第9期23-25,2-1,共3页
China Petroleum Machinery
基金
原机械工业部教育司基金
关键词
油田
注水站
设备
运行状态
趋势预测
神经网络
water injection station equipment running status trend prediction genetic neural network
