Predicting Temperature Profiles in a Flowing Well

Abstract

Summary A simple model suitable for hand calculations is presented to predicttemperature profiles in two-phase flowing wells. The model, developed withmeasured temperature data from 392 wells, assumes that the heat transfer withinthe wellbore is steady-state. Comparisons between the model's predictions andfield data indicate that the model is highly accurate within its range ofapplication. Introduction Predicting accurate temperature profiles in flowing wells can Predictingaccurate temperature profiles in flowing wells can greatly improve the designof production facilities in petroleum engineering. Temperature profiles helpcalculate accurate two-phase-flow pressure-drop predictions, which in turn canimprove an artificial-lift system design. Gas-lift design can be enhanced bymore accurate prediction of temperature at valve depth. In this way, thevalve's dome pressure can be set more accurately, thereby improving thepredictability of valve throughput. Existing temperature correlations are ofteninaccurate because they do not consider the effects of different fluids in theannulus and the cooling and heating of the fluid resulting from phase change. Rigorous theoretical models are often complex and inconvenient. They depend onmany variables and require information about fluid composition. This paperdescribes two methods for predicting the temperature profile in a flowing well. The first is a model derived from the steady-state energy equation thatconsiders the heat-transfer mechanisms found in a wellbore. The second is asimplified version of the model intended for hand calculations. An extensivedata bank of temperature profiles from 392 wells was used in itsdevelopment. Literature Review One of the earliest works on predicting temperature profiles in a flowingwell was presented by Kirkpatrick. He presented a simpleflowing-temperature-gradient chart that can be used to predict gas-lift valvetemperatures at the injection depth. Much of the classic work in this area wasdeveloped by Ramey, who presented approximate methods for predicting thetemperature of either a single-phase incompressible liquid or a single-phaseideal gas flowing in injection and production wells. Satter later improved Ramey's method by considering phase changes that occur within steam-injectionprojects. Shiu and Beggs simplified Ramey's method by correlating for aspecific coefficient in Ramey's equation. Willhite gave a detailed discussionof the overall heat-transfer mechanism in an injection well, and Coulter and Eardon developed a method for predicting temperatures in gas transmissionlines. Complex theoretical models, such as those by Zelic and modified Ramey'smethods, can be used to predict temperature profiles in flowing wells. Allthese methods require additional information about the fluid mixturecomposition. In addition, these methods are computationally complex and requirethe use of a computer. Such models are ideal for predicting temperatureprofiles associated with more difficult problems-e.g., a well flowing retrygrade condensates. The model developed in this paper is based on the Coulter-Bardon equation and incorporates Ramey's and Willhite's heat-transfermechanisms in a wellbore.