基于模糊集合理论的体积压裂水平井产量预测方法

Production prediction method of volume fracturing horizontal wells based on fuzzy set theory

赵振峰 白晓虎 陈强 苏玉亮 范理尧 王文东

ZHAO Zhenfeng BAI Xiaohu CHEN Qiang SU Yuliang FAN Liyao WANG Wendong

中国石油长庆油田分公司油气工艺研究院 中国石油大学(华东)石油工程学院

Oil and Gas Technology Institute, PetroChina Changqing Oilfield Company, Xi'an 710021, Shaanxi, China College of Petroleum Engineering, China University of Petroleum (East China), Qingdao 266580, Shandong, China

传统的统计分析方法存在散点乱、数据量大、难以分析总结出规律等问题，无法准确判定致密油水平井体积压裂主控因素。提出了一套的基于模糊集合理论水平井体积压裂效果评价方法：首先根据模糊集合理论建立了水平井产量分类方法，以鄂尔多斯盆地55口致密油水平井为例，根据产量对55口水平井进行分类；其次，通过分类数据的归一化处理得到每个因素变化区间的斜率，并据此绘制产量影响因素暴风图；最后，结合各影响因素的权重，通过多元回归方法建立了考虑多因素的水平井体积压裂产量预测模型。研究结果表明，该方法能够清晰准确地反映出各因素对体积压裂水平井开发效果的影响，并给出不同因素对体积压裂水平井产量影响程度。结合油田生产实例进行了分析，预测产量的平均相对误差仅为7.6%，为油田现场的压裂方案优化和产量预测提供了理论依据。

Traditional statistical analysis method is chaotic in scatters and huge in data bulk and can hardly analyze and summarize the laws, so it fails to accurately determine the main factors controlling the stimulated reservoir volume (SRV) of tight oil by horizontal well. A method for evaluating the effect of horizontal well SRV was proposed on the basis of fuzzy set theory. Firstly, a horizontal-well production classification method was developed according to the fuzzy set theory, and 55 tight-oil horizontal wells in the Ordos basin were classified based on the production. Then, the slope of the change interval of each factor was obtained by normalizing the classified data, and accordingly the storm chart of production influence factor was plotted. Finally, based on the weight of each influence factor, the production prediction model of horizontal well SRV including multiple factors was established by means of multiple regression method. It is indicated that this new method can reflect the effect of each factor on the development result of horizontal well SRV clearly and accurately, and provide the influential degree of each factor on the production of horizontal well SRV. What’s more, analysis was conducted on the actual case of oilfield production, and the average relative error of predicted production is only 7.6%. The research results provide the theoretical basis for the optimization of fracturing scheme and the prediction of production in the field.

10.13639/j.odpt.2019.04.019