一种水驱气藏产量预测的新方法

A New Method for Predicting the Production of Water Drive Gas Reservoirs

针对水驱气藏生产初期气井产量预测时,需要根据井口压力计算井底流压,间接预测气井产气量导致计算误差难以避免的问题,提出了基于最小二乘支持向量机的产气量预测方法。运用生产动态数据,直接预测出水驱气藏生产初期气井产气量。以某水驱气藏某区块气井为研究对象,首先从渗流机理出发,确定影响水驱气藏气井产气量的特征因素,然后用最小二乘支持向量机建立了气井产气量的预测回归模型,最后采用列队竞争算法对最小二乘支持向量机预测回归模型进行优化。结果表明:建立该模型所需数据特征少,应用性强,对水驱气藏产气量预测精确度较高。该方法为其他水驱气藏生产初期气井产气量预测提供了参考。

When the conventional binomial productivity equation predicts the production of a gas well in a water drive gas reservoir,it is necessary to calculate the bottom hole pressure according to the wellhead pressure,and indirectly predict the production,there is a certain error in the prediction;considering the defects of the above method,a least squares support vector machine production forecast method was proposed.The production forecast method uses production dynamic data to directly forecast production.Taking a gas well in a certain block of a water drive gas reservoir as the research object,firstly,starting from the seepage mechanism,two characteristic factors affecting the production of a water drive gas reservoir were determined,and then a least square support vector machine was used to establish a prediction regression model for the gas well productivity.Finally,the queue competition algorithm was used to optimize the least squares support vector machine predictive regression model.The experimental results showed that the established model required few data features,had strong applicability and high accuracy in predicting the production of water drive gas reservoirs.The method provides some reference for other water drive gas reservoir production prediction.