气田开发方案优选的新方法及其应用

在进行油气田开发方案的优选时，为了获取充分的信息，往往选择多个特征向量进行分析。然而，特征向量个数太多会增加对问题分析的复杂性；此外，评价结果的科学性很大程度上受研究者主观倾向或经验的限制，导致优选结果不一定完全符合客观实际。文章针对气田开发方案优选的特殊性，综合主成分分析和模糊数学两类方法的优势，构建了综合特征向量分析新方法。首先运用主成分分析法对各方案特征向量进行降维，提取综合特征向量，既能表征原来的特征向量所反应的信息，又能使各个综合特征向量之间互不关联，信息不重复；然后再通过两两比较判断法比较它们的相对重要性，得出各自的权重并加权成一个比较值，使方案优劣程度量化明朗化，最后利用该值进行方案的优选。通过在四川洛带气田蓬莱镇组气藏高低压分输方案优选的实际应用表明，该方法应用效果符合客观实际，可供类似问题借鉴。

New Method and Application of Optimization of Gas Field Developing Programs

When optimizing oil/gas field developing programs, lots of characteristic vectors are usually selected to analyze for obtaining abundant data. However, it will make the analysis more complicated because of too much vectors. In addition, the scientific property of the evaluation results is restricted greatly by the researchers’ subjective tendency and experience, which leads the optimized result maybe not fit the real cases completely. Aiming at the particularity of optimization of gas field developing programs, combining the advantages of the both methods of main component analysis and fuzzy mathematics, a new method of comprehensive characteristic vectors analysis is developed. First, the main component analysis is conducted to lower the dimensions of characteristic vectors for various programs and extract comprehensive characteristic vectors, which can characterize the information reflected by the original characteristic vectors, and also make the comprehensive characteristic vectors not correlative and the information unrepeated. Then, comparing every 2 vectors to judge their relative importance, the weight of each vector is obtained and weighted into a comparable value, which makes the program evaluation quantified and clarified. At last, the value is used for optimization of programs. With the practical application for optimization of separated transmitting programs by high and low pressure in Penglaizhen group gas reservoir of Luodai gas field, Sichuan, it shows the result from the method fits the real situation. The method provides the reference for the similar issues.