一种提高整体拟合程度的对数分布广义翁氏模型新解法  

A new method for the logarithmically distributed generalized Weng's model to improve the degree of global fitting

作　　者：刘刚[1] 孙建博[1] 徐杰 史鹏[1] 王晖 LIU Gang;SUN Jianbo;XU Jie;SHI Peng;WANG Hui(Natural gas research Institute of Shaanxi Yanchang Petroleum(Group)Co.,Ltd,Xi'an 710065,China)

机构地区：[1]陕西延长石油(集团)有限责任公司天然气研究院,陕西西安710065

出　　处：《复杂油气藏》2024年第3期317-322,共6页Complex Hydrocarbon Reservoirs

基　　金：鄂尔多斯盆地陆相页岩气地质-工程一体化智能评价方法研究(2022JM-171);延长陆相页岩气可采储量计算可行方法探究(ycsy2022jcts-B-44)。

摘　　要：广义翁氏模型已有解法在进行参数优选线性拟合时普遍存在生产初期拟合点离散偏差性较强的问题,这些偏差点与整体拟合趋势线的偏离程度较大,且偏差点的位置相对较为集中,加大了偏差点对整体拟合程度的影响,进一步影响了模型参数的优选准确性。针对模型已有解法存在的问题,本文从模型拟合点的分布形式研究出发,提出了一种对数分布广义翁氏模型新解法,区别于已有解法以时间t作为线性拟合函数自变量,新解法以时间对数ln(t)作为线性函数的自变量,拟合点的分布形态也从均匀分布转变为对数分布。实际应用表明,新解法使生产初期参数拟合点分布相对更加分散,分布趋势也更贴近整体拟合点的趋势,参数拟合初期拟合点偏离误差平均值减少了2.1%,参数拟合曲线整体偏离误差平均值减少了1.7%,模型新解法对应的产量偏离误差较模型已有解法平均值减少2.87%,表明模型新解法相对模型已有解法更能够提高产量的整体拟合程度。The existing solution for the generalized Weng′s model commonly suffers from the problem of strong discrete deviation of the fitting points at the initial stage of production when these parameters are preferred for linear fitting,and the deviation of these deviation points from the overall fitting trend line is larger,and the location of the deviation points is relatively more centralized,which increases the influence of the deviation points on the overall fitting degree,and further affects the accuracy of the model parameter preference.To address the existing problems of the model,this paper proposes a new solution for the generalized Weng′s model with logarithmic distribution from the study of the distribution form of the model fitting points,which is different from the existing solution with time as the independent variable of the linear fitting function.The new approach employs the logarithm of the time as the independent variable of the linear function,and the distribution form of the fitting points is also changed from uniform to logarithmic distribution.Practical applications demonstrate that the new solution results in a relatively more dispersed distribution of parameter fitting points at the initial stage of production,with a trend that more closely aligns with the trend of the overall fitting points.Consequently,the initial fitting point deviation error is reduced by 2.1%on average,while the parameter fitting curve deviation error is decreased by 1.7%on average,and the overall output deviation error is diminished by 2.87%on average.These results indicate that the new solution is superior to the existing method in enhancing the overall yield fit.

关 键 词：广义翁氏模型 线性函数 拟合 偏差 对数分布 

分 类 号：TE319[石油与天然气工程—油气田开发工程]