Study On Productivity Prediction For Single Well Of Oil Production Plant Completion

A work is particularly important in the engineering development of oil field that is productivity prediction, formulating, designing and adjusting the development plan is based on an important content that is the result of the productivity prediction. Thus it can be seen, productivity prediction plays an important role in the process of oil field production, so it must have strict requirements for the results of productivity prediction.At present, there exist some problems in productivity prediction:first, empirical formula is adopted to predict by the traditional methods, but the error is big, because under the premise of setting that these formulas are obtained. At the same time flexibility is poor of most formular, because these formula for some kind of special reservoir. Second, productivity is an indicator of reservoir characteristics, then there are many factors that have impact on the reservoir production capacity, so the productivity reflect the complex relationship of the various influence factors, increase the difficulty of prediction; Third, artificial prediction mainly through expert experience, subjectivity is big and the efficiency is low.To solve the above problems, this paper combined with the actual oil field production situation and did in-depth study on productivity prediction method, the main research contents are as follows:1.The paper proposes a method of production parameters extraction based on the grey relational analysis. Through the improved grey relational analysis method for the production parameters to analysis the important degree, extracting the effectively production parameters. At the same time, in order to demonstrate the feasibility and effectiveness of the improved method, the information amount theory, B’s correlation degree, Deng’s correlation degree are used to analyze the improved methods, after the experimental demonstration,the improved method can effectively extract the parameters that have great effect on productivity according to their importance. Eventually,the productivity parameters that are optimized can be used as the basis of productivity prediction,it is an effective guarantee for the accuracy and efficiency of the productivity prediction.2.The paper proposes a method that is“BP fusion algorithm based on ant colony system and improve activation function”. In this paper, the improved activation function is applied to make up for slow convergent speed of BP algorithm; the ant colony algorithm initial weights is applied to make up for easy to fall into local minimum value of BP algorithm, and these two ways are combined to solve the problem of the BP algorithm. According to the actual situation of oil field, the accuracy and speed of productivity prediction for single well can be effectively improved by using the improved BP fusion algorithm.