非常规油气开发初期改进双曲递减模型产量预测浅析

Analysis of Modified Hyperbolic Decline Model Prediction Shale Oil and Gas Well Production in Early Development

以北美地区的页岩油气生产井为研究对象,依据单井实际生产数据,使用超双曲递减模型拟合生产动态,计算递减指数,分析递减指数和生产时间的关系,并应用概率统计方法建立不同时间的概率分布。研究结果显示,超双曲递减指数和最终递减率是影响单井产量预测及可采量评估的主要因素,递减指数是随时间变化的函数,开发早期应用改进双曲递减模型时要注重递减指数的确定,较短时间生产数据拟合的产量趋势并非未来的真实反映。应尽可能建立递减指数在不同时间段的概率分布,合理预测产量和可采量。探索性地提出结合扩展指数递减模型和超双曲递减模型估算最终递减率的方法,扩展指数递减模型预测某一技术极限下井产量,用其约束超双曲递减模型预测得到最终递减率。实例分析显示,较好地解决了非常规油气开发初期单井产量预测过程中的不确定性问题。

We took the shale oil and gas wells in North America as the research object. According to the actual production data of single wells, we used the super hyperbolic decline model to fit production performance, estimated the decline exponent, analyzed the relations between the decline exponent and production time, and applied probability statistics method to establish the probability distribution at different time. Research results showed that the super hyperbolic decline index and final decline rate were main factors affecting unconventional oil and gas wells production prediction and reserves evaluation, and the decline exponent varies with time. When applying the modified hyperbolic decline model at the early stage of devel- opment, we should lay emphasis on the determination of decline exponent h, and the output trend fit by production data in a short time did not actually reflect the future. In addition, we had to set the probability distribution of decline expollents at different periods of time as far as possible to rationally predict output and exploitable yield. We put forward that to combine the expanding exponent decline model and the super hyperbolic decline model to estimate ultimate decline rate. We used the expanding exponent decline model to predict the output with a technical limit and to constrain the super hyperbolic de- cline model to obtain the ultimate decline rate. Case study indicated that the method solved uncertainties during prediction of single-well output at initial stage of unconventional oil and gas development.