在油气藏描述中判别模型的有效性及判别结果的稳定性

Effectiveness of discriminant model and stability of discriminant result in hydrocarbon reservoir description

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 to hydrocarbon reservoir description using multi--category discriminant lies in how to determine standard discriminant model and how to guarantee the stability in multi--category discriminant course. To cope with the above problem, we make nonlinear mapping to examine discriminant--model effectiveness so as to guarantee a correct model, and take more digits after the decimal point of data to eliminate the unstability in multi--category discriminant course. The method, named'multivariate statistical analysis system (MSAS) ', has been fully tested and perfected on workstation. We applied the softwares to comprehensive hydrocarbon reservoir description of No. 1 and 2 objective intervals in the Weibei C321-borehole area in Shengli Oil Field, bringing satisfactory result.