Research

Thermal Energy Conversion and Storage

A marginal well is a valuable subsurface asset precisely because much of the required infrastructure is already in place. It still has an operator, maintenance and integrity records, installed casing, and a characterized geology; most importantly, the drilling — the most capital-intensive component of a subsurface energy system — has already been paid for. Once the well is plugged and abandoned, much of that value is lost. More than two million wells in the United States have already reached that stage, often at substantial abandonment cost and with associated risks of methane leakage and groundwater contamination. The opportunity, therefore, is to intervene while a well is still marginal rather than after it becomes an orphaned liability.

The best use of these wells depends strongly on location. Geothermal gradient, formation thermal properties, completion depth, and local energy demand vary considerably across U.S. basins and may favor different applications: low-enthalpy geothermal heat for industrial processes; thermal energy storage charged from surplus generation and discharged through a power cycle; or, particularly in hot and humid climates, thermally driven cooling and dehumidification. Determining which application a given population of wells can support is itself an important part of the research.

In each case, the central engineering problem is one of heat transfer: how much energy the formation can accept or deliver, at what rate, how much can be recovered, and how repeated thermal cycling affects well integrity.

Our approach combines analytical and numerical methods to model heat transfer through the formation and wellbore components together with the thermal-hydraulic behavior of the circulating fluid. Each subproblem is treated using the simplest formulation that retains the required physics. Radial heat conduction into the formation is modeled semi-analytically, while casing, cement, annular fluid, and rock are retained as distinct thermal-capacitance layers. This captures the transient response without the computational cost of a fully discretized formation model and is sufficiently fast for large parametric studies and well screening. At the surface, cycles and components are selected, sized, and integrated for each application, allowing the surface plant and borehole to be designed as a coupled system rather than sequentially.

Technoeconomic assessment is conducted alongside the physical modeling, including capital and operating costs, levelized cost, and comparison with permanent plugging and abandonment. Cycle life is particularly important because the usable temperature range is constrained by the thermal limits of the casing and cement. Both performance and useful life are therefore predicted rather than assumed, providing a quantitative basis for screening candidate wells. Regulatory well records are then used to identify and geospatially cluster suitable candidates, allowing neighboring wells to share surface infrastructure and aggregate to useful scale. The objective is not necessarily to replace incumbent technologies entirely. The greatest value may instead occur during periods of highest demand, when existing systems are most constrained and most expensive to operate.

Associated publications

  • Y. Ko, J. R. Barbosa Jr., “Thermal analysis of a power-to-heat-to-power system using repurposed oil wells for energy storage.” Proceedings of the 18th International Heat Transfer Conference (IHTC-18), Rio de Janeiro, Brazil, August 2026, Paper IHTC18-0178.
  • J. R. Barbosa Jr., “Repurposing of legacy oil and gas infrastructure for the energy transition: a new frontier in wellbore heat transfer modeling.” Proceedings of the 11th World Congress on Momentum, Heat and Mass Transfer (MHMT 2026), Paris, France, April 2026, Paper ICMFHT 004. Keynote lecture.

Multiphysics Wellbore and Flow Systems

Wells undergo continuous thermal and mechanical loading throughout their operating life. Production and injection change the wellbore temperature field, causing expansion and contraction of casing and tubing, loading of the cement sheath, and pressure changes in trapped annular fluids. In deepwater completions, thermal expansion of fluids in sealed annuli can generate pressures sufficient to collapse or burst casing. In creep-active formations such as salt, deformation of the surrounding rock imposes additional loads on casings and annuli over months to years. Because the produced or injected stream is often multiphase and transient, flow, pressure drop, and heat transfer must be resolved together with the thermal and structural response of the well.

We address these problems using coupled thermal–structural models of the wellbore, flowline, and surrounding formation. Transient heat transfer between the flowing fluid, wellbore, and formation is treated numerically or semi-analytically, providing the thermal field used to predict casing and tubing deformation, annular fluid expansion, pressure changes in sealed annuli, and time-dependent formation creep. The models are formulated for computational efficiency, allowing parametric analysis over injection rates, fluid temperatures, mitigation volumes, and formation properties rather than isolated design cases. Applications include prediction of annular pressure during water injection, sizing of nitrogen cushions in partially filled annuli, and definition of safe operating envelopes. The same framework extends to CO2 injection wells, where thermal transients can be more severe and containment must be maintained over long periods. Multiphase flow assurance is treated within this framework as well, including benchmarking commercial simulators against offshore field conditions to determine the range over which their closure relations remain valid.

For thermal–structural wellbore integrity, model development is complemented by experiments conducted in collaboration with the Federal University of Santa Catarina using a laboratory-scale multi-string wellbore simulator. Four concentric tubulars form three independently filled and sealed annuli. Electrical heating and temperature-controlled circulation impose both monotonic heating representative of production start-up and repeated heating–cooling cycles representative of shut-in and restart, while distributed pressure and temperature measurements capture the transient response. Such datasets remain scarce in the open literature and provide the experimental basis needed to validate transient models and extend their application beyond the conditions for which they were developed.

Associated publications

  • I. O. Martins, A. K. da Silva, E. C. C. M. Silva, A. R. Hasan, J. R. Barbosa Jr., “Predicting the annular pressure behavior during water injection in offshore wells with a transient, multiphysics model.” Journal of Petroleum Science and Engineering 218 (2022)
  • A. P. da Veiga, I. O. Martins, J. G. A. Barcelos, M. V. D. Ferreira, E. B. D. M. Alves, A. K. da Silva, A. R. Hasan, J. R. Barbosa Jr., “Predicting thermal expansion pressure buildup in a deepwater oil well with an annulus partially filled with nitrogen.” Journal of Petroleum Science and Engineering 208 (2022) 109275.
  • E. B. D. M. Alves, A. P. da Veiga, E. A. Fancello, J. R. Barbosa Jr., “A coupled model of wellbore-formation thermal phenomena and salt creep in offshore wells.” Geoenergy Science and Engineering 233 (2024) 212548.
  • I. O. Martins, A. K. da Silva, E. C. C. M. da Silva, A. R. Hasan, J. R. Barbosa Jr., “Parametric analysis of water injection in an offshore well using a semi-transient thermal-structural model.” Geoenergy Science and Engineering 229 (2023) 212055.
  • E. B. D. M. Alves, J. R. Barbosa Jr., “Laplace transform-based modeling of heat transfer in producing wellbores.” Journal of the Brazilian Society of Mechanical Sciences and Engineering 45 (2023) 652.
  • J. L. L. de Almeida, P. M. de Oliveira, T. R. Gessner, J. R. Barbosa Jr., “Benchmarking commercial multiphase flow simulators for flow assurance in pre-salt offshore facilities.” Journal of the Brazilian Society of Mechanical Sciences and Engineering 48 (2026) 460.

Thermodynamics and Transport of CO2 in Complex Multiphase Fluids

Carbon capture and storage and CO2-based enhanced oil recovery are increasingly part of an integrated approach to subsurface carbon management. As a result, drilling, well construction, and intervention will increasingly occur in CO2-rich environments, while geothermal drilling raises related fluid-compatibility questions at higher temperatures. Conventional drilling and packer fluids were not designed for these conditions.

CO2 is considerably more soluble than methane in both water- and oil-based drilling fluids, with solubility strongly dependent on pressure, temperature, and fluid composition. Dissolved CO2 can alter density, rheology, pH, and phase behavior. One important consequence is the potential for a concealed kick: CO2 entering the well can dissolve in the drilling fluid at depth, produce little or no immediate surface indication, and come out of solution as pressure decreases during circulation to the surface. Changes in density and rheology can also promote barite sag and impair cuttings transport, while desorption can produce substantial Joule–Thomson cooling and CO2 contamination increases corrosion risk. Packer fluids present an additional challenge because their stability depends on emulsion and colloidal behavior that may also be affected by CO2 dissolution and the associated temperature changes.

We address these problems through complementary experimental and modeling work, with emphasis on characterizing the fluid and its interaction with CO2 rather than on continuous transport along the wellbore or pipeline. Solubility and diffusivity are measured using the pressure-decay method, in which a known quantity of CO2 is contacted with a fluid layer in a closed, thermostatted cell and the resulting pressure transient is interpreted through a coupled diffusion–swelling model. Measurements begin with characterization of the uncontaminated fluid, allowing changes caused by CO2 exposure to be distinguished from the fluid’s baseline behavior. Part of this work is conducted in collaboration with the Federal University of Santa Catarina, whose phase-equilibrium, transport-property, and calorimetric measurements provide the thermophysical data required for model closure and validation.

Absorption and desorption are modeled as transient diffusion with a moving interface and coupled to an equation of state capable of representing molecular association and specific interactions at the pressures and temperatures of interest. The resulting models provide a basis for defining operating envelopes, developing early kick-detection criteria, and guiding fluid selection and formulation for CCUS well construction. Transient absorption and desorption kinetics, particularly for multicomponent acid-gas mixtures, remain among the least characterized aspects of the problem. The properties and closure relations established here also provide direct inputs to the wellbore and flow models developed in the preceding thrust.

Technoeconomic Analysis of Emerging Energy Systems

Technology choices in energy systems are ultimately governed by thermodynamics and cost. Whether CO2 can be transported economically through repurposed pipelines, whether a repurposed well can store energy at competitive cost, or whether a produced-water treatment system justifies its capital investment depends on identifying where energy and exergy are lost and what those losses cost. The configuration with the highest efficiency is rarely the one with the lowest cost; the engineering decision lies in the tradeoff between the two.

System-level analyses, however, often simplify the component-level physics that determine this tradeoff. In CO2 conditioning, for example, intercooler pressure drop is commonly prescribed rather than calculated from local flow and thermodynamic conditions. Such assumptions are not neutral. Resolving pressure drop at the component level can change the optimum number of compression stages and reduce predicted exergetic efficiency by several percentage points relative to an idealized zero-pressure-drop case.

Our approach is therefore to build integrated, component-resolved models of complete energy chains and optimize them simultaneously for cost and exergetic performance. Multi-objective genetic algorithms are used to construct the Pareto front rather than identify a single optimum. Costs are developed component by component from capacity-scaled equipment correlations, updated to current dollars, with installed costs estimated using detailed factor methods rather than a single system-level multiplier. This allows the economic result to be traced directly to the physical assumptions and component performance that produced it. Sensitivity analysis over quantities such as energy prices and discount rates then establishes which assumptions control the result and where the conclusions remain robust.

The framework has been applied to source-to-sink CO2 conditioning and transport using repurposed pipeline infrastructure and to thermal energy storage in repurposed wells, where it provides metrics such as levelized cost of storage and round-trip efficiency. Produced-water management is a natural extension: it represents the largest waste stream associated with oil and gas production and presents the same fundamental optimization problem, with energy-intensive treatment and separation competing against disposal cost, water reuse opportunities, and increasingly stringent regulatory requirements.

Associated publications

  • A. Falcão, M. V. P. Carneiro, A. de Souza, A. K. da Silva, J. R. Barbosa Jr., “Thermo-economic assessment of the Cabiúnas–São Tomé CCS pilot project.” SPE/AAPG/SEG Carbon Capture, Utilization and Storage Conference (CCUS 2026), The Woodlands, Texas, 30 March – 1 April 2026.