考虑参数相关性的蒙特卡洛模拟方法在储量计算中的应用

Application of Monte Carlo Simulation Method Considering Parameter Correlation in Reserves Calculation

针对使用容积法计算储量时参数选取存在的不确定性和多解性,以及参数之间的相关性问题,介绍了利用蒙特卡洛方法计算储量的数学原理与相关性理论,以昌吉油田吉17、吉37井区块芦草沟组为例,借助水晶球软件,建立了储量计算参数的数学模型,以含油面积为均匀分布、油层厚度为正态分布、孔隙度为对数正态分布、含油饱和度为极小值分布计算出了储量的概率分布曲线以及累积概率分布曲线。考虑了孔隙度与含油饱和度的相关性,计算两者之间的相关系数为0.633,在水晶球软件中设置相关系数重新进行了储量模拟,计算结果显示,储量概率分布曲线符合对数正态分布,储量均值约为2830万吨。与未考虑相关性计算的储量进行了对比,两者总体趋势一致但有差别,考虑相关性计算的储量更符合客观实际,更具可靠性。

Aiming at the uncertainty and multiple solutions of parameter selection and the correlation between parameters when using volume method to calculate reserves,the mathematical principle and correlation theory of using Monte Carlo method to calculate reserves are introduced.Taking Lucaogou Formation of Ji17 and Ji37 well blocks in Changji Oilfield as an example,a mathematical model for calculating reserves is established by Crystal Ball Software.The probability distribution curve and cumulative probability distribution curve of reserves are obtained when the oil-bearing area is evenly distributed,the thickness of reservoir is normal distribution,the porosity is lognormal distribution,and the oil saturation is minimum distribution.The correlation coefficient between porosity and oil saturation is 0.633.Reserves is calculated again by setting the correlation coefficient between porosity and oil saturation in Crystal Ball Software.The result shows that the probability distribution curve of reserves is consistent with the lognormal distribution,the average reserves is about 28.3 million tons.The calculated reserves is compared with the reserves calculated without considering the correlationcoefficient.The general trend of the two is the same but there is difference between them.The reserves calculated by considering the correlation are more consistent with the objective reality and more reliable.