近组分模型中主变量的选取问题研究

Study on Selections of Primary Variables in Kvalue-based Compositional Model

近组分模型有两种主变量选取方法,分别是选取自然变量作为主变量的方法和选取总属性变量作为主变量的方法。比较了这两种方法在数学模型和数值求解流程方面的差异,分析了在计算精度和计算效率方面产生的异同。二氧化碳驱概念模型的模拟结果表明,油气两相都存在时,虽然两种方法在一维驱替和二维驱替中的计算精度是一样的,但是选取自然变量作为主变量的方法不仅允许的最大时间步长更大,而且在相同时间步下运行速度更快。在相同时间步下,选取自然变量作为主变量的方法在二维驱替中提速效果有所降低。

This paper compares the differences between the two methods in terms of mathematical models and numerical solution processes,and analyzes the similarities and differences in terms of calculation accuracy and calculation efficiency.A CO2 flooding numerical simulations shows that,when both oil and gas phases are present,the accuracy of the two primary variable selections in one-dimensional displacement and two-dimensional displacement is the same,but the Kvalue-based compositional model with the natural variable selection not only has the larger maximum allowed time step,but also it runs faster at the same time step.At the same time step,the speed-up effect of Kvalue-based compositional model with the natural variable selection is reduced in two-dimensional displacement.