论文部分内容阅读
应用多类判别方法进行油气藏描述的关键技术是如何确定标准的判别模型，并且保证多类判别过程中的稳定性。本文针对上述问题，应用非线性映照方法，对判别模型进行有效性检验，以确保模型建立的正确性；并提出了“数据高精度截位”法，以消除多类判别过程中的不稳定性。此法已在工作站上全面开发和完善，定名为多元统计分析系统（简称MSAS）。用该软件系统对胜利油田潍北021井区一号及二号目的层进行了油气藏综合描述，取得了明显的效果。
The key technique of applying multi-class discrimination method to describe oil and gas reservoirs is how to determine the standard discriminant model and ensure the stability of multi-class discrimination. In order to solve the above problems, this paper applies the non-linear mapping method to test the validity of the discriminant model to ensure the correctness of the model. The method of “high-precision data truncation” is proposed to eliminate the instability of multiple discriminative processes . This method has been fully developed and improved on the workstations, named multivariate statistical analysis system (referred to as MSAS). The software system is used to describe the oil and gas reservoirs of the No. 1 and No. 2 0 wells in Wei Bei, Shengli Oilfield well, and achieved remarkable results.