Abstract: The combination of horizontal wells and hydraulic fracturing has significantly enhanced shale oil production. Fracture parameter inversion plays a pivotal role in evaluating fracture effectiveness and forecasting production capacity. However, there is currently a lack of an integrated and effective methodology for fracture parameter inversion and productivity prediction that aligns with the full fracturing-soaking-producing (FSP)workflow. To address this, a comprehensive approach is developed that incorporates fracture propagation modeling, well test analysis, and production simulation based on multi-source data integration. Numerical models for fracture propagation, soaking well testing, and dynamic production simulation are established using the displacement discontinuity method, embedded discrete fracture model (EDFM), well test analytical, and the Markov Chain Monte Carlo (MCMC)method. Fracture extension is first simulated using the ZFRAC simulator based on the displacement discontinuity method, with the construction pumping curve fitted to derive fracture parameters consistent with the actual fracturing operation. These parameters are further refined using post-fracture shut-in data. A numerical well test model is then employed to constrain the inversion results, which are verified via dynamic analysis method. Finally, the embedded discrete fracture framework is used to describe fracture geometry, and MCMC is applied for automated history matching and productivity prediction. The proposed methods are applied to a case study of fractured horizontal well H2, where post-fracture parameters are successfully evaluated and single-well productivity accurately predicted. The results confirm the reliability and practical value of the integrated fracture evaluation and production forecasting methodology.

Key words: shale oil, fracturing-soaking-production integration, productivity prediction, parameter inversion, embedded discrete fracture model, multi-data collaborative analysis

摘要： 水平井+压裂技术可显著提高页岩油井产能,而裂缝参数反演是压裂评价和产能预测的关键,但目前缺乏一套有效的"压—焖—采"一体化裂缝参数反演和产能预测方法。为此,考虑页岩油压裂—焖井—排采阶段的开发特征,利用位移不连续、嵌入式离散裂缝(EDFM)、试井分析方法和马尔可夫链蒙特卡洛方法(MCMC)等手段,建立了"压—焖—采"多数据协同分析的裂缝扩展、焖井试井和生产模拟不同数值模型,形成了页岩油井裂缝参数反演与产能预测方法,并开展了实例应用。首先,基于位移不连续法利用ZFRAC模拟器开展了裂缝扩展模拟,拟合施工泵压曲线,获取符合实际压裂过程的单段裂缝参数,利用压后停泵数据进一步校正单段裂缝参数。其次,建立焖井数值试井模型约束裂缝参数反演结果,利用产量动态分析方法验证裂缝参数反演结果。最后,基于嵌入式离散裂缝刻画裂缝形态,利用马尔可夫链蒙特卡洛方法实现自动历史拟合,以开展产能预测。采用建立的裂缝参数反演和产能评价方法对压裂水平井H2井进行实例分析,评价了压后裂缝参数,预测了单井产能,证明了压裂评价和产能预测方法的可靠性和实用性。

关键词: 页岩油, "压—焖—采"一体化, 产能预测, 参数反演, 嵌入式离散裂缝, 多数据协同分析