海上油田多层砂岩定向井合采初期产能预测方法及应用——以渤中C油田为例
Prediction Method and Application of Initial Productivity in Combined Production of Multilayer Sandstone Directional Well in Offshore Oilfield—Taking Bozhong C Oilfield as an Example

海上油田初期产能评价通常采用基于DST测试资料的比采油指数法，但由于储层物性和流体性质的变化，以及多层合采层间干扰等因素影响，取值具有人为性，造成初期产能预测评价准确性降低。根据海上油田定向井多层合采开发的特点，通过分析储层渗透率、渗透率变异系数、流体性质和层间干扰系数等对产能的影响，建立储层物性参数、渗透率变异系数等参数和比采油指数的关系，提出了综合产能指数法和利用定向井实际投产数据样本通过机器学习法预测评价定向井初期产能。对比分析不同预测方法适用性和预测效果，为海上油田定向井多层合采初期产能预测提供可靠方法，以提高新井初期产量评价预测的准确性。实例应用表明，基于生产井大数据样本应用机器学习法评价预测定向井初期产能，比传统的比采油指数法更加便捷、准确性也较高，具有良好的应用前景。
The specific oil recovery index method based on DST test data is usually used to evaluate the ini-tial productivity of offshore oil fields. However, due to the change in reservoir physical and fluid properties, as well as the interference between layers of multi-layer production, the value is ar-tificial, resulting in a decrease in the accuracy of the initial productivity prediction and evaluation. According to the characteristics of multi-layer combined production development of directional Wells in offshore oil fields, the relationship between reservoir physical property parameters, permeability variation coefficient, fluid properties and interlayer interference coefficient and specific oil recovery index is established by analyzing the effects of reservoir permeability, permeability variation coefficient, fluid properties and interlayer interference coefficient on productivity. The comprehensive productivity index method and machine learning method are proposed to predict and evaluate the initial productivity of directional Wells by using the actual production data samples. The applicability and prediction effect of different prediction methods are compared and analyzed to provide a reliable method for the initial productivity prediction of multi-zone combined production of oriented Wells in offshore oil fields, so as to improve the accuracy of the initial production evaluation and prediction of new Wells. The example application shows that the application of the machine learning method to evaluate and predict the initial productivity of directional Wells based on big data samples of production Wells is more convenient and accurate than the traditional specific oil recovery index method, and has a good application prospect.