M1-1油田厚陡窄复杂油藏水平井产能预测方法研究

Study on productivity prediction method of horizontal wells in thick,steep,narrow complex reservoirs of M1-1 oilfield

针对M1-1油田次生断层发育,储层厚、陡、窄,致使水平井水平段有效长度差异较大的特点,传统的比采油指数配产法与公式法无法实现该类水平井产能的准确预测,利用灰色关联法与神经网络法相结合建立的新方法可以很好地解决这一问题。首先,采用灰色关联法筛选出影响M1-1油田水平井产能的主控因素主要为水平段长度、渗透率、原油黏度、生产压差、有效厚度,然后将所确定的主控因素作为BP神经网络的神经元,建立神经网络模型。经M1-1油田实际数据网格训练,预测结果与实际数据吻合较好,表明该方法适合M1-1油田次生断层发育的厚陡窄复杂油藏水平井产能预测,最终预测调整井的初期产能为45~75 m^3/d。

In view of the development of secondary faults in M1-1 oilfield,the reservoir is thick,steep and narrow,which makes the effective length of horizontal section of horizontal wells greatly different.The traditional production allocation method and formula method of specific oil recovery index can't realize the accurate prediction of productivity of this kind of horizontal well.A new method,which combines the grey relation method with the neural network method,can solve this problem well.First of all,the main factors affecting the productivity of horizontal wells in M1-1 oilfield are selected by using the grey correlation method,which are mainly horizontal section length,permeability,crude oil viscosity,production pressure difference and effective thickness.Then the main controlling factor is regarded as the neuron of BP neural network,and the neural network model is established.After the training of M1-1 oilfield actual data grid,the prediction results are consistent with the actual data,which shows that this method is suitable for the prediction of production capacity of complex reservoirs with thick,steep and narrow secondary faults in M1-1 oilfield.By using this method,the initial production capacity of the adjustment well was predicted to be 45~75 m^3/d.