基于GM(1,1)的一种新型嵌套方法在气井产量预测中的应用  

A New Nesting Prediction Model Based on GM(1,1)Used in Gas Production Prediction

作　　者：郭科[1,2] 罗芳[1,2] 詹泽东[1,3] 李月林[4] 

机构地区：[1]四川省数学地质重点实验室(成都理工大学) [2]成都理工大学管理科学学院,四川成都610059 [3]中国石油化工股份有限公司西南油气分公司,四川成都610041 [4]成都理工大学地球科学学院,四川成都610059

出　　处：《国土资源科技管理》2015年第3期90-94,共5页Scientific and Technological Management of Land and Resources

基　　金：2013高等学校博士学科点专项科研联合资助项目(20135122110010)

摘　　要：针对常规气井产量预测中因忽略气井工艺措施等其它因素影响而将其处理为干扰异常点,造成与气藏排水采气的开发策略不一致的技术难题,结合GM(1,1)模型与灾变模型,提出一种新的嵌套型GM(1,1)模型。该模型通过将气井工艺作为灾变因子,可以更准确地预测气井自然生产模型下的产量变化规律,有效弥补了常规预测法的不足。研究将新模型应用于川西气藏某生产气井单井产量预测,预测结果分别与GM(1,1)预测结果和Arps产量递减预测模型结果进行比较,结果显示新预测模型比GM(1,1)及产量递减预测模型的回归斜率和相关系数更接近1且回归截距更趋近于0,表明新方法的预测精度更加精确,能准确反映气田开发中的产量变化,不失为一种可选的新型产量预测方法。Because of the ignorance of the influence of gas well process measures and other factors,which are usually treated as abnormal interference point,the conventional production prediction always varies from the development strategy of gas reservoir.In this paper,combining GM（1,1）model with catastrophe mode,a new nested type GM（1,1）model was proposed.This new model can predict more accurately the production history in natural production model by regarding gas well process measures as disaster factors.The new model is applied to gas reservoir in West Sichuan gas production prediction of a single well production.The results were compared to the results of GM（1,1）and prediction model of production decline.The results showed that SLOPE and correlation coefficient of new model were more closed to 1than the others and regression intercept of the new was more closed to 0,which indicated that prediction precision of new model is more accurate and can reflect changes of the production of gas field development precisely.

关 键 词：灰色建模 气井产量 嵌套预测 

分 类 号：O29[理学—应用数学] TE328[理学—数学]