A novel model for predicting the temperature profile in gas lift wells

Abstract One of the most common methods for calculating the production oil rate in a gas lift well is nodal analysis. This manner is an accurate one, but unfortunately it is very time consuming and slow. In some modern studies in petroleum engineering such as close loop control of the wells this slowness makes it impossible to have an online optimization. In fact, before the end of the optimization the input parameters have changed. Thus having a faster model is necessary specially in some of the new studies. One of the sources of slowness of the nodal analysis is the temperature profile estimation of the wells. There are two general approaches for temperature profile estimation, some like heat balance are accurate but slow. Others, similar to linear profile assumption are fast but inaccurate and usually are not used commonly. Here, as a new approach, a combination model of heat balance and linear temperature profile estimation has represented which makes the nodal analysis three times faster and it is as accurate as heat balance calculations. To create this, two points (gas injection point and end of tubing) are selected, then using heat balance equations the temperature of those two points are calculated. In normal nodal analysis the temperature of each wanted point in the well is estimated by heat balance and it is the source of slowness but here just two points are calculated using those complex equations. It seems that between these points assuming a linear temperature profile is reasonable because the parameters of the well and production such as physical tubing, and casing shape and properties and gas oil ratio are constants. But of course, it still has some deviation from the complete method of heat balance which using regression and assigning a coefficient to the model even this much of the deviation could be overcame. Finally, the model was tested in various wells and it was compared with the normal nodal analysis with complete heat balance models. Results showed that the new model is as accurate as normal heat balance but three times faster.