油田产量预测系统的研制

如何有效地预测油田的未来产量一直是油藏工程研究中的一项重要内容,准确地预报油田开发过程中的动态产量,是合理调整油田规划和制定方案,实现优化开发和管理的重要依据。本文研究了几种新的预测方法(基于人工神经网络的稠油预测模型、灰色预测模型、CAR预测模型等),并首先将其应用于油田产量预测。本文开发的预测系统集成了20多种常规的和新型的预测方法,对油田产量的预测及管理都能够给出有价值的数据。在充分并准确采集油田各单位产量信息及建立切合实际的数据模型的基础上,将油田产量信息采集与处理、预测与管理等环节,有机地结合成为一个整体。实际预测结果表明,该方法可以取得较好的预测效果。

Development of forecasting system for output oil quantity in crude oil production

It is an important research topic in the oil reservoir engineering that explores how to forecast the future oil output effectively. To forecast the dynamic oil output exactly is the crucial basis of adjusting the development plans reasonably, optimizing mining, establishing the scheme. A few of forecasting methods, such as neural networks based forecasting model for dense oil, the gray forecasting model and CAR forecasting model, were researched in this article. The new forecasting models have been first applied in our forecasting system. The forecasting system developed in this paper integrated more than 20 traditional and advanced output-predicting methods, the system provides valuable data for the yield predicting and management of the oilfield. On the basis of acquiring the production information of every department of the oilfield fully and accurately and setting up the practical data models, it integrates the acquiring, handling, prediction and managing of crude oil production information into an organic entity. The real predicting experiment results using the method show that perfect predictive result can be obtained.