摘要
井层优选和施工方案优化是油气田重复压裂技术的核心 ,其“瓶颈”问题是“数据有限”和“模型与参数给不准”以及“许多问题的机理不清楚” ,无法获得问题的显式表达。文章尝试应用统计学来获取影响压裂效果的各项因素与压裂效果的关系模型和预测模型 ,从而优选施工井层和优化施工方案。实践证明 ,对于样本“数据有限”(小样本 )的情况下 ,支持向量机算法技术适应性强、精度高 。
Optimum seeking well and formation and program optimization is the key of repetitive fracturing technology for oil and gas fields. The bottle-neck problem is 'data limited', 'model and parameters provided not accurately'. and 'mechanism of many phenomena not clear'. So, the explicit expression can't be derived. The article tries to get the relationship model and the prediction model between the different factors that influence fracturing and the fracturing results. So that optimum seeking well and formation and program optimization can be done. Practice proves that under the conditions of 'data limited', the algorism of Support Vector Machine has strong adaptability and high accuracy. It can be applied broadly in the study field for repetitive fracturing.
出处
《天然气工业》
EI
CAS
CSCD
北大核心
2004年第3期75-77,共3页
Natural Gas Industry
基金
"十五"国家重点科技攻关项目研究成果之一 (项目编号:2 0 0 1BA6 0 5A - 0 5 - 0 2 - 0 3)
