Accurate Prediction Wellbore Transient Temperature Profile Under Multiple Temperature Gradients: Finite Difference Approach and Case History

Abstract

Abstract Failure to estimate the correct temperature during cementing operations can often lead to long delays in rig operations or can allow the flow of gas or water from shallow over pressured formations. API (American Petroleum Institute) Sub-committee 10 (Well cements) has developed methods to determine the bottom hole circulating temperature (BHCT) for the proper design of cementing slurries. These methods work adequately in most cases. But, in wells that are deviated and have multiple temperature gradients like those found offshore in deep water basins, these methods are inadequate. This paper presents a finite difference model developed to predict the wellbore and formation transient temperature behavior during fluid circulation in the wellbore. It also provides the information on temperature recovery after the cement slurry becomes static. Emphasis is placed on evaluation of wells with multiple temperature gradients, multiple fluid circulation schedule, and wellbore deviations. The effect of offshore water currents is also discussed. The predictions of the wellbore temperature profiles and return temperatures from this model are validated through actual measured wellbore temperature profile history, including offshore and onshore cases. The paper also discusses the impact of job execution parameters such as fluid density, flow rates, inlet temperature, and fluid rheology. Guidelines are provided to help the engineer design an optimal working schedule to achieve one's particular goal of controlling wellbore temperature at desired level during actual cement placement operation. Introduction The prediction of wellbore temperatures is above all other parameters; the single most influencing assumption made in the design and testing of cement slurry. Wellbore temperature affects all aspects of cement slurry design, including thickening time, compressive strength development, fluid rheology, and fluid loss. The primary chemical reactions governing the hydration process of the cement slurry are directly affected by temperature. Over or under estimation of the wellbore temperature can lead to any number of possible non-productive time events ranging from excessive Wait On Cementing (WOC) to the drilling out of cement filled casing. The effect of small changes in estimated temperature on additives used to control cement slurry properties can be significant. Besides the properties of cement slurry, the rheology of other fluids pumped into the well will also be affected by well temperatures. For decades the thickening time well simulation schedules contained in API RP10B "Recommended Practice for Testing Well Cements", have served as the fundamental basis for determining well cementing temperatures. With the increasing thermal complexity of horizontal, extended reach, and deepwater wells, the predictive limits of the schedules contained in the API document were exceeded. As such, the need for well condition specific cementing temperature determinations became evident. Over the years two approaches have been emerged to estimate the circulating fluid temperature: analytical and numerical. The analytical solutions were obtained for system geometries of lesser complexity, such as the case of a single casing string and single temperature gradient. Holmes and Swift1 obtained solution for steady-state heat transfer both in the conduits and in the formation. Arnold2 presented analytic solutions for steady heat flow in the conduits but unsteady-state flow in the formation. The coupling of steady heat flow in the wellbore with unsteady heat flow in the formation was first introduced by Ramey3, which has been successfully applied by many authors, dealing with steam injections, oil, and geothermal steam production, among others. Note that these earlier analytical models, proposed by Edwardson et al4. and modified version by Tragesser et al5., are good for a simple specific condition but unsuitable as generalized predictive tool.