采油厂完井单井产能预测分析研究

发布时间：2018-05-30 20:03

本文选题：单井产能预测 + 改进灰色关联分析　；参考：《东北石油大学》2015年硕士论文

【摘要】：在油田工程开发当中尤为重要的一项工作便是产能预测,产能预测结果是编制开发规划、设计开发方案以及调整开发方案的重要依据。由此可见产能预测在油田生产过程中的重要作用,因此对于产能预测结果的准确度就必须有严格要求。目前,产能预测主要存在以下几个问题:其一,传统方法大多采用经验公式的方式进行预测,而这些公式是在作了许多设定的前提下得出的,误差很大,同时公式法大多针对某类特殊油藏,其灵活性较差;其二,产能是表述储层特征的一项指标,有很多因素都对储层生产能力有影响,因此产能是各类影响因素复杂关系的反映,导致预测难度增加;其三,人工预测主要通过专家经验,主观性较强,效率较低。针对以上问题,本文结合油田实际生产情况,对产能预测方法进行深入研究,主要研究内容如下:1、以优选产能影响参数为目标,本文提出了“基于改进灰色关联分析的产能参数提取方法”。通过改进的灰色关联分析方法对产能有关参数进行重要度分析,提取出有效的影响参数。同时为了论证改进的灰色关联分析方法在产能参数筛选方面是行之有效的,应用了信息量理论分析方法、邓氏关联模型、B型关联模型两个传统灰色关联分析方法对改进方法进行检验分析,经过实验论证,该改进方法能够有效的提取对产能影响度高的参数,最终将优选的参数作为产能预测的基础,是产能预测结果准确度以及效率的有效保证。2、以产能预测为目标,本文提出了“基于蚁群系统与改进激励函数的BP融合算法”。本文分别应用改进的激励函数弥补BP算法收敛速度慢的不足;应用蚁群系统处理初始权值来弥补BP算法易陷入局部极小值的不足,并将两者结合解决BP算法存在的问题。结合油田实际情况,将改进后的BP融合算法应用到单井产能预测当中,有效提高预测准确度以及预测速度。
[Abstract]:Productivity prediction is one of the most important tasks in oilfield engineering development. The result of productivity prediction is an important basis for compiling development plan, designing development scheme and adjusting development plan. It can be seen that productivity prediction plays an important role in the production process of oil field, so the accuracy of productivity prediction must be strictly required. At present, the main problems of productivity prediction are as follows: first, most of the traditional methods use empirical formulas to predict, and these formulas are obtained under the premise of a lot of settings, and the error is very large. At the same time, the formula method is mostly aimed at a certain kind of special reservoir, its flexibility is poor; second, productivity is an index to express reservoir characteristics, and many factors have an effect on reservoir productivity, so productivity is a reflection of the complex relationship of all kinds of influencing factors. Third, artificial prediction is mainly based on expert experience, which is subjective and inefficient. In view of the above problems, this paper makes an in-depth study on the productivity prediction method combined with the actual production situation of the oil field. The main research contents are as follows: 1. The aim is to select the optimal productivity influence parameters. In this paper, a method of productivity parameter extraction based on improved grey correlation analysis is proposed. The important degree of productivity parameters is analyzed by the improved grey correlation analysis method, and the effective influence parameters are extracted. At the same time, in order to prove that the improved grey correlation analysis method is effective in the selection of productivity parameters, the information quantity theory analysis method is applied. The improved method is tested and analyzed by two traditional grey correlation analysis methods of Dunn's correlation model and B type correlation model. The experimental results show that the improved method can effectively extract the parameters which have a high degree of impact on productivity. Finally, the optimal parameters are taken as the basis of productivity prediction, which is the effective guarantee of accuracy and efficiency of productivity prediction. Aiming at productivity prediction, a BP fusion algorithm based on ant colony system and improved incentive function is proposed in this paper. In this paper, the improved excitation function is used to compensate for the slow convergence speed of BP algorithm, and the ant colony system is used to deal with the initial weight value to make up for the deficiency of BP algorithm which is prone to fall into local minimum value, and to solve the problem of BP algorithm by combining the two methods. Combined with the actual situation of oil field, the improved BP fusion algorithm is applied to single well productivity prediction, which effectively improves the prediction accuracy and prediction speed.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE328

【参考文献】