概率法在煤层气储量估算中的应用

煤层气田在勘探评价初期,储量估算参数往往存在一定的分布区间,采用体积法估算储量难以给出确定结果.针对此类问题,以M煤层气藏为例,利用概率法估算储量,围绕储量估算参数(储层面积、厚度、含气量),结合地质分析和储量参数样本分别建立理论分布模型(三角分布、正态分布、对数正态分布、威布尔分布、伽马分布)和任意分布模型;利用Kolmogorov-Smirnov(K-S)检验和卡方检验对相关理论分布模型进行拟合度优选;确立储层面积采用三角分布和正态分布模型,储层厚度采用威布尔分布和任意分布模型,储层含气量采用伽马分布和任意分布模型,并形成4组不同模型组合方案,再利用蒙特卡罗方法对储量进行估算.结果表明:任意分布模型服从原始样本分布,置信区间较窄,理论分布模型则体现过滤作用,两者对储量估算结果有一定影响;储层含气量对储量估算结果影响较大;概率法可以排除人为因素,估算出不同概率级别下的储量,从而起到降低决策风险的作用.

Application of probability method in coalbed methane reserves estimation

In the early stage of exploration and evaluation of coalbed methane field, it is difficult to estimate reserves accurately with the volume method due to the widely distribution of the calculation parameters of reserves. Aiming at these problems, probability method was used to estimate reserves by taking M coalbed methane field as an example. Firstly, probability distribution models of theoretical distribution (triangular, normal, lognormal, Weibull, gamma distribution) and arbitrary distribution were built for calculation parameters of reserves (reservoir area, thickness and gas content) based on the geological analysis and reserve calculation parameter samples. Secondly, some theoretical distribution models with good fit were selected by the use of K-S test and chi-square test. Thirdly, the triangle distribution and normal distribution model were adopted for reservoir area, Weibull distribution and arbitrary distribution model were used for reservoir thickness, and gamma distribution and arbitrary distribution model were used for reservoir gas content. Therefore we can use Monte Carlo method to estimate reserves based on four combinations by different models for each parameter. The results show that the arbitrary distribution model obeys the original sample distribution, the confidence interval is narrow, and the theoretical distribution model reflects the filtering effect, both of which have a certain impact on the reserves estimation results. In addition, the reservoir gas content has a greater impact than other calculation parameters on the reserve estimation results. The probability method can eliminate human factors and estimate reserves under different probability levels, thus reducing the decision risk.