论文部分内容阅读
为了预测某裂缝性气藏在18年规划周期内的累计产量,建立了3种元模型,并结合数值模拟累计产量预测结果对各模型进行了验证和对比。估算了气藏的基本参数(孔隙度、渗透率及含水饱和度)并对气藏进行了数值模拟。将气藏内某井18年内的累计产量表示为各气藏参数的函数,采用储量计算经验函数排除了研究空间内的低潜力点,并基于最大熵准则选取25个高潜力点作为设计点。建立了二次模型、乘性模型和RBF模型3种元模型,基于各模型预测了25个设计点和7个试验点处的累计产量。预测结果表明:各模型均能比较准确地预测累计产量;RBF模型的累计产量预测准确性优于另外2种模型;与数值模拟方法相比,元建模方法的计算时间大幅缩短。
In order to predict the cumulative production of a fractured gas reservoir over the 18-year planning cycle, three kinds of metamodels were established, and the models were verified and compared with the numerical simulation of the cumulative production forecast results. The basic parameters (porosity, permeability and water saturation) of the gas reservoir are estimated and the gas reservoir is numerically simulated. The cumulative production of a well within 18 years of a gas reservoir is expressed as a function of each gas reservoir parameter. The low potential point in the research space is excluded by the exponential function of reserve calculation, and 25 high potential points are selected as the design points based on the maximum entropy criterion. Three models of quadratic model, multiplicative model and RBF model were established. Based on each model, the cumulative production at 25 design points and 7 test points was predicted. The prediction results show that all the models can predict the cumulative production more accurately. The accuracy of the RBF model is better than the other two models. Compared with the numerical simulation method, the computational time of the meta-modeling method is greatly shortened.