基于常规测井参数的产能预测模型研究及方法优选

Study on the Methods Optimization and Productivity Prediction Model Based on the Conventional Logging Parameters

目前油田产能预测的方法较多且各具优缺点,本文以柴达木盆地英东油田为例在分析储层特征、开展储层分类及确定产能影响因素基础之上,对多元线性回归与BP神经网络两种产能预测方法进行研究并建立相应模型。分析表明当多元回归模型预测结果误差较大时,神经网络可以寻找更合适的多元非线性模型来进行预测,是产能预测的最佳方法。

At present, there are several methods for oilfield production prediction. Each of them has its advantages and disadvantages. In the paper,taking Yingdong oilfield of Qaidam Basin as an example, two kinds of prediction models of multiple linear regression and BP neural network were researched and the appropriate models were established based on reservoir characteristics, reservoir classification and affecting factors of productivity. The analysis showed that the neural network can find a more appropriate multiple nonlinear model,which is the best way for production forecasting when the prediction errors of multiple linear regression is bigger.