【摘要】 针对当前连通性评价模型仅能预测产液动态数据、现有分流量方程不能准确表征整个含水变化周期内产出特征的内在缺陷性,基于油水产出动态数据,建立了一种可快速反演水驱油藏井间连通状况的连通性评价新方法.该方法将油藏离散为一系列以生产井为中心的控制单元(CRMP),通过CRMP电容模型和分段含水率表征方程进行井点液量计算和含水率追踪,获取井点处的生产数据;在此基础上,采用新近提出的随机近似梯度StoSAG优化算法实现动态数据预测值与观测值的自动拟合,快速反演连通系数、时滞常数、泄油孔隙体积等参数.应用实例表明,相比于集合卡尔曼滤波(EnKF)算法和投影梯度算法,StoSAG算法鲁棒性更强,油水动态数据拟合效果好,连通性评价结果可靠;相比于单一的Koval方程,分段含水率表征模型能够更好地反映全生命周期内的产水规律,适用范围更广.更多还原

【Abstract】 To address the issues that the existing connectivity evaluation models can only predict liquid production rate and the widely-used fractional-flow equations are not efficient to represent the water-producing characteristics in the whole water-cut variation period accurately, a novel connectivity evaluation method for waterflooding reservoir is established by using injection and production data. This developed model discretizes the reservoir into a series of production-based reservoir control volume. By coupling the producer-based capacitance-resistance model(CRMP) with a segmented water fraction-flow equation to obtain liquid production rate and water-cut, oil-water production data of each well can be further determined. Finally, the estimated parameters including connectivity, time constant and drainage volume of each producer are determined by history matching the measured production data with the newly developed Stochastic Simplex Approximate Gradient(StoSAG) optimization algorithm. Case studies show that,compared with the EnKF algorithm and the traditional projected gradient algorithm, the robustness of the ensemble-based StoSAG optimization procedure is higher. Moreover, the segmented water fractional-flow equation can achieve better dynamic matching and prediction effect of the measured production data than that of the single Koval fractional-flow model, which fully validate the reliability of the proposed interwell connectivity evaluation method and provides the theoretical basis for the actual applications in strongly-heterogeneous waterflooding reservoirs.更多还原