A New Method to Calculate Loading Liquid Water of Gas Wells in Qinghai Oilfield

Sebei Gas Field is located in the Qaidam Basin, China, and is an unconsolidated sandstone gas reservoir with edge and bottom water layers. This special characteristics is a great challenge for gas production. As liquid column in the wellbore with a height of 1000 meters will bring 100 times of atmospheric pressure to the bottom of the well. Once the pressure of water in the wellbore is higher than the bottom hole flow pressure, the gas well will face the risk of shutdown. Actually most of the wells in Sebei, which have produced for several years, are producing water and taking drainage measures to assist gas producing. However, the exact moment of liquid loading is hard to determine, therefore the following drainage measures, also known as deliquifications, are difficult to design and optimize. Accurately and timely judging the degree and the height of liquid loading in the wellbore is one of the prerequisites for rational deliquification design. In this paper, Dr. Cao's liquid loading calculating model is applied to compute liquid loading in Sebei gas wells. The model is verified to be strong while the water production is relatively low but bring more errors for high water production. To solve the problem, a machine learning method is used to optimized and modified Cao’s model by adding a dynamic correction coefficient.