应用人工神经网络方法预测气井积液

气井井筒积液对天然气的开采影响极大,准确地计算气井临界流量对气井开发至关重要。气井携液临界流量理论计算模型主要有液滴模型和携液率模型,然而在实际计算过程中往往会出现计算结果偏差大、不能满足工程需要等问题。文中提出一种应用人工神经网络方法预测井筒积液的新模型,该模型充分利用了气井现有的生产测试数据,简化了大量复杂的机理研究,具有更广泛的实用性。生产井的计算结果表明,应用神经网络模型预测气井积液的成功率较高,可以用来判断气井积液。

Using artificial neural network method to predict liquid loading in gas well

Liquid loading in gas well can pose a serious threat to the exploitation of natural gas.To accurately calculate the critical flow rate of gas well is vital to gas reservoir development.Currently,engineering technicians use the liquid drop model and liquid holdup model to calculate the critical flow rate for liquid loading in gas well.However,the above two old models have a significant shortcoming that the calculated result is far from the reality and can not meet the requirement of gas well development.This paper presents an artificial neural network model for predicting the minimum flow rate for continuous removal of liquids from the wellbore.The model is developed taking full advantage of the test data in gas wells,and the new model can also simplify the complex mechanism studies of liquid loading,which has a wider range of practical application.The new model has been used to calculate the actual production of gas well.The results show that the developed model can provide high accuracy in predicting liquid loading in gas well and can also determine whether there is liquid loading in gas well or not.