组合法在塔里木盆地石油产量预测中的应用  

Application of a Combination Forecast Model in the Trend Forecast of Oil Production in the Tarim Basin

作　　者：姚纪明[1] 于炳松[1] 车长波[2] 张道勇[2] 

机构地区：[1]中国地质大学地质过程与矿产资源国家重点实验室,北京100083 [2]国土资源部油气资源战略研究中心,北京100034

出　　处：《自然资源学报》2009年第5期907-914,共8页Journal of Natural Resources

基　　金：国家自然科学基金项目“新一轮全国油气资源评价”之“塔里木盆地常规油气资源评价”(ZP-C-05);国家973项目(2006CB202302)资助

摘　　要：影响塔里木盆地石油产量的因素是多方面的,这些因素与产量之间具有非线性关系,采用通常的数学方法很难对其进行分析,鉴于此,论文利用灰关联度对塔里木盆地石油产量的主要影响因素进行分析,从中选择关联度较大的因素作为影响元,利用GM(1,1)和BP人工神经网络建立的组合预测模型,对塔里木盆地石油产量发展趋势进行短期预测,运用翁氏旋回预测模型在短期预测的基础上对塔里木盆地石油产量进行中长期预测,其预测结果对决策者制定勘探战略、优化勘探项目、确定投资方向提供参考,对油气管理部门提供一定的科学依据。The factors that affect the oil production of the Tarim Basin are various, there exists nonlinear relation between these parameters and petroleum output. It is difficult to analyse it using the common mathematic method, so a forecasting model combining GM （ 1,1 ） with BP artificial neural network has been constructed, in which the other related factors were taken into account, and the primary factors which affect the production or reserves of oil and gas can be determined through grey correlation degree. Then the short-term trends of production of oil in the Tarim Basin have been predicted with the combined model and the long-term ones done with the Weng＇ s cycle. The forecasting result provides the scientific basis for the policy makers and the departments of oil and gas exploration.

关 键 词：石油 预测 灰关联度 GM(1 1)模型 人工神经网络 翁氏旋回 

分 类 号：TE122[石油与天然气工程—油气勘探] P618.13[天文地球—矿床学]