Modelling Geothermal Energy Extraction from Low-Enthalpy Oil and Gas Fields Using Pump-Assisted Production: A Case Study of the Waihapa Oilfield

Abstract

As the energy sector transitions toward decarbonisation, low-to-intermediate temperature geothermal resources in sedimentary basins—particularly repurposed oil and gas fields—have emerged as promising candidates for sustainable heat and power generation. Despite their widespread availability, the development of these systems is hindered by gaps in methodology, oversimplified modelling assumptions, and a lack of integrated analyses accounting for long-term reservoir and wellbore dynamics. This study presents a detailed, simulation-based framework to evaluate geothermal energy extraction from depleted petroleum reservoirs, with a focus on low-enthalpy resources (<150 °C). By examining coupling reservoir behaviour, wellbore heat loss, reinjection cooling, and surface energy conversion, the framework provides dynamic insights into system sustainability and net energy output. Through a series of parametric analyses—including production rate, doublet spacing, reservoir temperature, and field configuration—key performance indicators such as gross power, pumping requirements, and thermal breakthrough are quantified. The findings reveal that: (1) net energy output is maximised at optimal flow rate (~70 kg/s for a 90 °C reservoir), beyond which increased pumping offsets thermal gains; (2) doublet spacing has a non-linear impact on reinjection cooling, with larger distances reducing thermal interference and pumping energy; (3) reservoirs with higher temperatures (<120°C) offer significantly better thermodynamic and hydraulic performance, enabling pump-free or low-duty operations at higher flow rates; and (4) wellbore thermal losses and reinjection effects are critical in determining long-term viability, especially in low-permeability or shallow fields. This work demonstrates the importance of a coupled, site-specific modelling in assessing the geothermal viability of petroleum fields and provides a foundation for future techno-economic and sustainability assessments. The results inform optimal design strategies and highlight scenarios where the geothermal development of oil and gas fields can be both technically and energetically viable.

1. Introduction

Over the past three decades, geothermal development has predominantly focused on high-temperature hydrothermal systems located in geologically favourable regions [1,2]. While these systems have enabled commercial-scale geothermal development, they are spatially limited and represent only a fraction of the global geothermal resource base. In contrast, low-temperature geothermal resources—estimated to account for approximately 70% of global potential—remain vastly underutilised [3]. As the global energy sector moves toward decarbonisation, there is a growing need to tap into these abundant but underexplored resources to support a cleaner, more resilient energy mix.

One particularly promising subset of low-temperature resources lies within sedimentary basins, many of which coincide with mature or depleted oil and gas fields [4,5]. These fields exist across a wide range of geological contexts, including the United States [6,7], Europe [8,9], China [10], Australia [11,12], Canada [13,14,15], and New Zealand [16,17,18,19]. They offer a unique opportunity to repurpose existing infrastructure for geothermal energy extraction, potentially enabling power generation or direct-use heating from fluids below 150 °C [20] (p. 18) [21,22,23,24,25,26].

A critical review of the literature highlights that most studies fail to address key aspects essential for long-term geothermal development in petroleum systems, especially sustainability, thermal coupling, and methodological appropriateness. Early demonstrative projects, such as that at the Naval Petroleum Reserve in Wyoming, reported favourable geothermal outcomes, largely attributable to anomalously high heat flow and shallow-depth production (~1200 m), yielding mass flow rates of approximately 83 kg/s [27]. However, such cases represent exceptions rather than norms. The majority of oil and gas reservoirs reside in purely conductive thermal regimes with markedly low flow rates and low vertical permeability, rendering conventional geothermal frameworks inapplicable [28].

One of the examples is the pilot project by Xin et al. [29], conducted in the Huabei oilfield. This field achieved a net electric output of approximately 310 kW from a low-enthalpy source (110 °C), with a production rate of 33 kg/s, supported by pumping. The reservoir, spanning 44.9 km² with a high permeability of 158 mD, allowed for reinjection of produced water at 114 °C into the reservoir at 85 °C. However, the study highlights a significant limitation—most oilfields lack similar reinjection capacity, high inter-well connectivity, or expansive lateral dimensions [30,31].

The lack of an appropriate systematic methodology when critical information was missing from petroleum dataset resulted in unexpected results or became a hurdle in conducting thorough analysis in several cases [7,14,15,31,32,33,34,35,36]. For examples, in the Australian Great Artesian Basin, preliminary tests were conducted at a demonstration site. The analysis resulted in an oversized cycle (20 kW), which consumed more power (3.5 kW) than anticipated [11]. Similarly, many investigations ended up generating overly optimistic estimates from the extrapolation of the bottom-hole temperature dataset [14,18,19].

The critical review of the methodologies in the literature revealed that they often overlooked one or more aspects of an integrated geothermal system composed of wellbore, reservoir, aquifer, and surface network. One of the most overlooked aspects is the thermal losses in wellbores models. For example, in the Williston Basin, the thermal losses resulted in a significant temperature decline in the production fluid (from 135 °C to 64 °C) due to unique low-flow rates (ranging from 0.4 kg/s to 0.6 kg/s) observed in oil and gas fields [37]. Similar findings were reported by the investigations within the Los Angeles Basin [32,38]. Although the reservoirs in these regions are generally deep (3.5 km), most have a radius of 4 km and a permeability of 0.05 md to 0.2 md (millidarcies); the analysis did not suggest a viable project. These studies argued a higher probability of success with increased production rates supported by pumps [37].

One of the other inadequately addressed components is the aquifer response. For example, Liu et al. [39] implemented a roadmap to assess the geothermal potential from coproduced fluids of a single wellbore in the Villafortuna–Trecate oilfield in Italy. The presence of a strong aquifer at 5000 m provided pressure support as well as an in situ hydrothermal resource (160°C), eliminating the need for fluid reinjection. The study concluded that the primary contribution to thermal energy harvest came from the water phase. However, the study did not account for the absence of a strong aquifer on energy extraction or impose restrictions on the volume of fluid withdrawn.

Moreover, numerical reservoir simulations have been limited in the literature, thereby resulting in the importance of thermal and hydraulic performance of the reservoir during the long term being overlooked. For example, in the North Perth Basin geothermal resource, Wellmann et al. [40] qualitatively described sustainable pumping rates for a doublet system, both with or without an advective background flow, but recommended numerical simulations to carefully examine the time it takes for reinjected cold water to reach the production wellbore, a phenomenon known as the breakthrough time [40]. Apart from key reservoir properties such as porosity and permeability distributions, the thermal breakthrough time may be influenced by hydraulic dispersion or heat conduction in the fluid phase or delayed by in situ thermal resource in the bedrocks [41]. Sanyal and Butler reported an analytical methodology for evaluating the technical potential of operating a geothermal doublet in petroleum fields [42]. According to their study, resources below 190 °C are favourable for pump-assisted production. They later extended the methodology to include the productivity index and wellbore interference in an infinite radial system [43]. However, this methodology tends to overestimate the flow potential of a fractured reservoir [32]. Furthermore, the analytical technique relies on several assumptions, such as a homogeneous, isotropic, infinite reservoir, radial heat conduction, etc. Clotworthy et al. [44,45] adopted probabilistic modelling to estimate flow from a simplified wellbore model supported by a pump with a fixed discharge head. However, their analysis did not account for breakthrough effects on the longevity of a doublet well. Hydraulic dispersion or heat conduction in the fluid phase might reduce the longevity of sustainable pumping rates. While probabilistic modelling can provide some benefits [44], it is subject to uncertainties when the field is confined, such as the presence of growth faults [44,46].

Gong et al. [47], in their numerical modelling, found that increased reinjection resulted in a greater decline in the average reservoir temperature. Their results also indicated an increase in the produced oil volume while the volume of water remained constant during increased reinjection. As a result, numerical simulations have been widely recognised as a useful tool in multiple studies [29,47,48,49,50].

While a few studies have attempted to define systematic approaches for energy extraction and assess the geothermal potential, none have fully modelled the entirety of a hydrocarbon resource in terms of geothermal energy extraction. Thus, the motivation of this study is to develop a framework that covers key components contributing to energy extraction and aims to model petroleum fields as complete geothermal system. The authors present some components of this framework and apply an integrated subsurface approach to simulate dynamically coupled heat transfer effects within the sediments.

This work answers the following research questions:

• How well can the flow from a petroleum system be predicted using a geothermal simulator?
• How important are the heat transfer effects during modelling energy extraction from oil and gas resources?
• How do you assess optimum extraction rates from a petroleum producer?

The remainder of this document is structured as follows: Section 2 describes the methodology adopted in the study, Section 3 presents the results of different cases to assess the influence of parameters on temperature, heat, and energy production, Section 4 discusses the implications of the results, and Section 5 provides the summary conclusions. Further research directions are suggested in Section 6.

2. Methodology

The proposed method in this paper combines numerical reservoir simulation, wellbore modelling, and lithology integration to determine suitable reservoir properties for geothermal development while considering wellbore heat losses, constant pumping discharge pressure, and the presence of an aquifer. The study uses the Waihapa field in New Zealand as a case study to investigate how well an oil and gas system can be modelled as a geothermal resource and to what extent thermal energy output can be predicted using a geothermal simulator, Volsung [51]. Volsung is a geothermal modelling software that streamlines the modelling process by linking reservoir, wellbore, and surface network models. It enables efficient and accurate simulation of the geothermal system. The method presented in this study can be applied to other geothermal resources in oil and gas fields worldwide.

Based on our previous work [52], this study introduces an extended model that integrates a wellbore-modelled producer with a reservoir model to analyse the geothermal potential of an oil field supported by pumping. The high-level workflow for this evaluation is presented in Figure 1.

• The goal is to create a history-matched model that provides reference values for key system parameters, i.e., wellbore deliverability curve, reservoir size, porosities and permeabilities of the reservoir, and distance between the wellbores. The history-matched model is essential for calibration and validation purposes. It begins with user input of wellhead pressure, mass flow rate, and dimensions of the wellbore obtained from well logs.
• Next, the wellbore model is used to estimate productivity index and construct a deliverability curve. In cases where some wellbore data are unavailable, an alternate strategy is adopted to address this issue.
• The wellbore model is then imported into Brynhild component of the Volsung package to integrate with the reservoir model. The reservoir model has stratigraphic information, which incorporates lithology information obtained from petroleum reports for the oilfield. This integration allows for a more realistic representation of the system.
• To complete the integrated model, an empirical heat-to-power conversion model is included, assuming a constant reinjection temperature.

The model determines various factors, including thermal and hydraulic breakthrough effects, sustainable pumping rates for the longevity of the production well, and the energy required to run pumps at a constant wellhead flow rate.

2.1. Modelling an Oil and Gas System Using a Geothermal Tool

The proposed method focuses on assessing the feasibility of energy extraction from an oil and gas system by considering the physical processes that impact fluid flow and temperature. Wellbores play a crucial role in extracting energy, and the model assumes a production–injection wellbore configuration, such as a doublet or five-spot [53]. It is recommended to adopt an exact configuration of wellbores when information is available to achieve more realistic results.

The reservoir model incorporates the fluid drive mechanism and considers the effects of the thermophysical properties of the coproduced fluid [54,55]. In terms of the fluid drive mechanism, which is primarily pressure-driven in oil and gas reservoirs, the method suggests an additional step to calibrate the boundary conditions of the grid using a natural-state run. This step ensures that the reservoir model replicates the over-pressured state of the oil and gas fields [56]. The effects of fluid properties, particularly the differences between water-based brines typically used in geothermal tools and hydrocarbon fluids present in oil and gas fields, can be accounted for through permeability distribution [57]. These effects are more pronounced in gas fields compared with oil fields [58].

To develop a geothermal system in an oil and gas field, the wellbore model is constructed, and injection wellbores are added. The wellbore pair is given data on the total production rates and pressure depletion in the reservoir during production. This simplification helps in our goal of evaluating the field for thermal energy extraction using a doublet. This assumption allows us to assess the suitability of an oil and gas system for geothermal energy extraction [56]. Rock properties such as permeability and porosity of the field structure are adjusted for calibration purposes.

In cases where data are unavailable, the model adopts a doublet configuration of production and injection wellbores. By considering the physical processes that control wellbore flow, the model assesses the suitability of oil and gas systems to produce energy.

Simulations of energy extraction are performed using a doublet configuration following the methodology for enhanced heat extraction described in the literature [53,59], as shown in Figure 2, where the blue circles represent injection wellbores and red circles indicate production wellbores. The assumed injection–production well pair configuration is chosen for maximum heat extraction, with specific guidelines such as the well spacing distance d and length recommended to optimise heat extraction [53,59].

2.2. Heat-to-Power Conversion Model

For the purpose of this study, a simplified power conversion model is adopted, assuming a constant reinjection temperature of 50 °C and a fixed power conversion efficiency of 10% [60]. The rationale for this simplification is to primarily investigate the impact of operational parameters on energy extraction. The text explicitly acknowledges that this efficiency value is likely higher than what would be typically achievable with low-temperature geothermal resources, as indicated in the literature [61].

To enhance the rigor of future analyses, the text suggests incorporating a correlation developed by Liu et al. [39]. This correlation demonstrates a decrease in power conversion efficiency with decreasing temperatures, ranging from 14% down to 5%. Integrating such a temperature-dependent efficiency model would provide a more realistic assessment of the actual power generation potential.

The pump specifications used in the study are based on a review of industry reports and available literature [42,62]. The pump capacity chosen for analysing the upper limits of energy extraction is the highest available, which is 70 bar of pressure support [42,62]. To calculate pumping power, Volsung uses the following equation [63]. The resulting hydraulic pump power is given in Equation (1).

$$Pumping power=\left(p_{wh}+(p_{hs}+p_{fr})-\left(p_{res}-{\displaystyle \frac{\dot{Q}}{\dot{PI}}}\right)\right)\times \left({\displaystyle \frac{\dot{Q}}{\rho }}\right)$$

(1)

2.3. Use Case: Waihapa Oilfield

The Waihapa oilfield (Figure 3)—located in the Taranaki region in the southwest of New Zealand—presents a unique confluence of favourable geological, operational, and logistical factors as listed in

Table 1. Key characteristics of the Waihapa oilfield in the Taranaki region, New Zealand [64].

Criteria	Waihapa Field	Ref
Maturity of the field	Long production history, infrastructure exists	[64,65,66]
High water cut	~98% water content—ideal for coproduction	[64,65]
Reservoir temperature	89 °C—viable for low-enthalpy geothermal extraction	[64]
Natural pressure support	Strong aquifer connectivity	[64]
Permeability and porosity	High permeability due to natural fractures (5 × 10−14 m2)	[64,67]
Available data	Detailed drilling, production, lithology data accessible	[64]
Reservoir depth	2500 m—balanced for geothermal development	[64]
Infrastructure and proximity to grid	Near existing infrastructure and end-use locations	[5,68]
Unitised field	Easier to manage subsurface heat and fluid flow	[5,68]
Simple well geometry	Vertical wells (0.17 m dia), easy for simulating and modelling	[43,69,70]
Original-oil-in-place (OOIP) *	5.08 × 106 m3	[71]
Peak fluid production	40 kg/s	[64]
* 1000 US barrels of oil = 158.9873 m3

* Assumes 1000 US barrels of oil = 158.9873 m³ during conversion.

, making it an ideal representative case for evaluating the feasibility of using oil and gas fields for geothermal energy extraction. Unlike many other fields that suffer from limitations in permeability, reinjection capacity, or data availability, Waihapa provides the right conditions to rigorously test and calibrate an integrated geothermal model.

2.3.1. Waihapa Wellbore Model

The proposed method utilises Gudrun, a standalone wellbore simulator of the Volsung software package, version 2024.x [69,70], to construct a wellbore-modelled producer. The producer computes the deliverability curve by considering the feed zone input and reservoir conditions.

Unlike analytical wellbore models [42], a coupled producer in Gudrun can account for reservoir anisotropy and multi-layered reservoirs, capturing their impact on wellbore performance [72]. It also takes into account the effects of fluid composition and phase mobility during coproduction, enabling a more comprehensive understanding of the reservoir’s hydraulic behaviour.

If the productivity index value is not available, the method suggests a workflow to estimate the wellbore deliverability curve. The first step of this workflow assumes a constant mass flow rate at the feed zone. The bottom-up simulation option in the wellbore simulator then calculates the productivity index value using Equation (2):

$$\dot{w}=\dot{PI }\times (p_r-p_w)$$

(2)

In Equation (2), $\dot{w}$ is the mass flow rate, in kg/s; $\dot{PI}$ is productivity index, in (kg/s)/bar; $p_r$ is the static reservoir pressure (more commonly as average static reservoir pressure), in bars; and $p_w$ is flowing bottom hole wellbore pressure, in bars.

To construct the wellbore model for the Waihapa oilfield, drilling data, casing, open-hole, and tubing information were obtained from the Petroleum Basin Explorer (PBE) database from GNS Science [73]. The wellbore represents a typical petroleum vertical wellbore with a constant diameter of 0.17 m [43]. The location of the feed zone, where the initial reservoir pressure of 291 bar is provided, is at a depth of 2575 m [66].

Because not all information could be obtained, the peak mass flow rate value in the field was prescribed to calculate the productivity index of the wellbore using the wellbore simulator Gudrun. The simulator then generates a deliverability curve for the wellbore using Equation (2), indicating the maximum allowable mass flow rate from the wellbore, as shown in Figure 4. As the pressure declines, the maximum production rate may not be sustainable [74], so pump support is included for the production wellbore to achieve useful production rates.

2.3.2. Waihapa Reservoir Model

We built a reservoir model that includes stratigraphy, wellbore profiles, feed zone locations, etc., from the database in the Brynhild reservoir simulator of the Volsung software package [75]. The Waihapa reservoir model was generated based on reports available in the New Zealand Petroleum and Minerals Database and the GNS Science Petroleum Basin Explorer Database [73,76]. The stratigraphic distribution of the field was obtained, and a full grid of the field was created as shown in Figure 5, including the reservoir, aquifer, barrier region, subsurface lithology, and wellbore locations. The figure also shows colour-coded lithology information projected on a cut-section of the spatial interrelation between the wellbores and corresponding lithology.

A usual practice is to consider a grid larger than the field boundaries to reduce pressure boundary effects [77,78]. We model the field with a large area (3000 m × 11,000 m × 3600 m) along with an additional aquifer (6500 m wide) with water influx controlled using a barrier layer (450 m wide); see Figure 5 and Figure 6. The reservoir is modelled as a three-layered MINC dual-porosity medium [79]. The reservoir model’s dimensions (areal size and 190-m thickness) were determined from petroleum reports [80].

In the case of a petroleum reservoir, the boundary is defined by the structural trap and hydrocarbon–water contact [81]. By using a large grid to represent the total petroleum reserve and the volume of accessible fluid, pressure boundary effects can be reduced in the numerical solution [77].

Regarding discretisation, the spatial and temporal discretisation does not significantly affect the numerical solution for a homogeneous reservoir with constant porosity and permeability [77,82]. This is because hydraulic steady flow conditions are quickly achieved. However, to evaluate the thermal behaviour of the reservoir, a fine grid mesh is recommended in the area of interest, particularly close to the wellbores [77]. In contrast, a coarse grid mesh can be used in the outer regions near the reservoir boundary [77]. Trade-offs need to be performed in terms of the number of grid cells to avoid excessively long computational times.

The selected element size of 100 m in the grid independence study (

Table 2. Grid independence study.

Element Size	No. of Elements	Production Temperature	Absolute Error
m		K	%
20	277,760	353.96	0.00
60	66,000	354.72	0.21
100	34,650	354.809	0.24

) was chosen based on a balance between computational efficiency and result accuracy. While finer meshes (e.g., 20 m and 60 m) were tested, the difference in key output parameters—particularly the position and shape of the thermal front—were found to be negligible beyond the 100 m resolution. This choice is justified by the minimal variation observed in the thermal front location and shape between 100 m and 60 m where these values had stabilised.

In this setup, the model assumes no-flow boundary conditions on all four sides of the grid. Additionally, the model assumes atmospheric boundary conditions at the top of the grid, and the bottom surface acts as a hot plate to regulate reservoir temperature. The model begins with initial atmospheric temperature and pressure values, assuming the rock was saturated with water millions of years ago. The natural state run establishes the pressure and temperature distribution as observed before production starts.

Table 3. Model input values.

	Units	Reservoir Layer	Barrier Layer	Aquifer Layer
Type		Minc3D	Porous	Porous
Rock Properties
Porosity	%	0.2	0.2	0.25
Density	kg/m3	2600	2600	2600
Specific Heat	kJ/(kg · K)	1	1	1
Wet Heat Conductivity	W/(m · K)	2	2	2
Dry Heat Conductivity	W/(m · K)	2	2	2
Permeability
kx	1 × 10−14 m2	7	3	1
ky	1 × 10−14 m2	7	3	1
kz	1 × 10−14 m2	7	3	1

lists the values of each parameter used to model the reservoir, barrier, and aquifer. The rock properties, such as permeability and porosity, are set based on available petroleum reports. The dual-porosity reservoir is modelled using 3 MINC layers each with a volume fraction of 10%, 20%, and 70%, with 100 m fracture spacing in each direction, a matrix porosity of 5%, and a matrix permeability of 1 × 10⁻¹⁹ m².

2.3.3. Natural State Model

Before running the natural state modelling, the model is assumed to be fully saturated with water. It reflects a depleted oilfield with high water cut (>95%), as suggested by the conclusion of the recent study [39]. The assumption is in accordance with the findings in the literature that the primary contributor in thermal response is water fluid, and oil phase has a negligible influence on the thermal output; however, the additional economical value should be studied separately. According to this study, in a depleting field, water is the major contributor in thermal exchange in contrast to the hydrocarbon component.

The reservoir simulator, Brynhild [51] calculates the thermophysical properties of the fluid using a pure water equation of state (EOS) based on IAPWS-IF97 [83]. The natural state model is run to establish the pressure, temperature, and enthalpy distribution, replicating the processes that lead to the formation of an oilfield [57]. The natural state simulation reflects the over-pressured state of the oilfield and serves as the basis for pressure history matching with actual field data. Figure 7 demonstrates the results of the natural state modelling to calibrate temperature distribution in the field. Figure 8 presents the pressure from initial conditions from millions of years ago until 1970 before production at Waihapa commenced. The pressure reaches a steady state at the end of the simulation. The natural state run in geothermal reservoir modelling is conducted to achieve pressure and temperature distribution pre-production commencement in the field.

2.3.4. Pressure-Matched Model

To obtain a history-matched model, a doublet configuration consisting of a pair of production and injection wellbores is assumed for geothermal energy generation in the oilfield. The wellbore model developed previously is imported into the reservoir model, and an injection wellbore is added. The pressure history for the field is obtained from Figure 9 to establish a match, and the operation scenario follows the total fluid production rate as shown in Figure 10. Figure 9 shows pressure depletion behaviour due to production at each well. This has been converted to an average response for the whole field as originating from a single producer and injector configuration. The average response of the system has also been shown in public reports and the initial reservoir modelling study conducted for the field [80]. Thus, the simulation used total production rates as inputs for the wellbore pairs (producer and injector) to reproduce the average pressure response from the system.

For calibration, the permeability parameter of the reservoir, barrier, and aquifer is tuned to obtain a match with public data available for the field [64]. Figure 11 shows the production and injection mass flow rates used to match the pressure with the Waihapa model. The comparison between measured and simulated pressures is shown in Figure 12, demonstrating a match with the actual field data. The adjustment of the top boundary during the natural state simulation reflects the processes leading to the formation of an oilfield.

To summarise, the history-matching exercise highlighted important aspects of developing a geothermal model for an oil and gas reservoir. It is essential to capture the driving mechanism of hydrocarbon fluid, as petroleum systems are over-pressured, and production occurs due to pressure depletion (depletion drive). Fluid property effects in a brine-focused modelling environment compared with a hydrocarbon-based simulation tool are also considered.

The calibrated pressure–history matching replicates the fluid behaviour in the reservoir during production by adjusting properties such as porosity and permeability. Another limitation of this pressure matching experiment is the calibration of the transient temperature response. Once more data are available, a more realistic model can be established. However, considering the aim of this study, pressure matching is deemed suitable to model the oilfield for extracting thermal energy. Another reason behind not including temperature match is that the conduction-dominated field setting seems to reflect liner thermal gradient throughout field production history. The challenges of modelling an oil and gas system in a geothermal modelling tool are overcome through the approaches mentioned above.

The above history-matching exercise, Figure 12, highlighted important aspects of developing a geothermal model for an oil and gas reservoir. A geothermal model in a petroleum field should capture the driving mechanism of hydrocarbon fluid. Petroleum systems are over-pressured, and production occurs due to pressure depletion also known as depletion drive, unlike in geothermal systems where boiling drives the flow. The adjustment of the top boundary resulted in representing processes leading to the formation of an oilfield during the natural state simulation.

Another aspect is to represent the fluid property effects in a brine-focussed modelling environment as compared with a hydrocarbon-based simulation tool. The calibration of pressure–history matching has obtained a replication of the fluid behaviour in the reservoir during production. Key properties adjusted during calibration are the porosity and permeability parameters of the reservoir, aquifer, and barrier regions. In terms of the case considered in this paper, the above two approaches appear to overcome the challenges of modelling an oil and gas system in a geothermal modelling tool.

2.4. Energy Extraction Simulations

2.4.1. Simulated Scenarios

The model developed serves as the basis for energy extraction simulations to assess pumping requirements and identify favourable reservoir parameters for a geothermal project. The simulations consider various scenarios by varying parameters such as mass flow rate, temperature, reservoir depth, doublet distance, wellbore productivity index, and permeability-thickness production. The simulations are conducted over a period of 35 years (from 1988 to present), with a base case being defined by common values for each parameter. The range values and parameters, along with the values for the base case, are listed in

Table 4. Range of parameters for simulating scenarios related to energy extraction methodologies.

	Mass Flow Rate	Reservoir Depth	Resource Temperature	Well Productivity Index	Permeability-Thickness Product	Doublet Separation Distance
	kg/s	m	°C	(kg/s)/bar	mD-m	m
Base Case	100	2500	90	1.5	7.5	500
Range	50–150	2500	90–150	1.5	5.0–15.0	300–900

.

This section contextualises the chosen parameters within the broader landscape of low-temperature geothermal systems and relevant demonstration projects in oil and gas fields. The specified range of values aligns with operational data observed in these real-world applications. To provide a concrete basis for comparison,

Table 5. Net installed capacity of demonstration projects in oil and gas fields [61].

S. No.	Oil and Gas Fields
	Plant Name	Reservoir Temperature, °C	Flow Rate, kg/s	Net Installed Capacity, kW
1.	Teapot Dome Oilfield, Wyoming, US (Naval Petroleum Reserve No. 3)	~110	90	250
2.	Huabei Oilfield, China (LB oil reservoir)	120	33.3	400

is referenced, which compiles key parameters such as temperature, flow rate, and installed capacity from existing projects. This table serves as a valuable benchmark against which the findings of the current case study can be evaluated, allowing for an assessment of the practical relevance and potential of the proposed energy extraction method within the context of established low-temperature geothermal and oil and gas field applications.

2.4.2. Truncated Model

To capture the thermal response of the reservoir more accurately, a refined grid near the production and injection wellbores is used in the truncated model. However, refining the full model would result in excessive computational times. Therefore, the model is truncated until the reservoir top, and a no-mass flow boundary condition is implemented at the low-permeability/cap-rock boundary [56]. Conductive heat flow is allowed in this truncated model, which still maintains a 3D representation of the reservoir to simulate the effect of rock and fluid properties through fractured pathways between the production and injection wells during energy extraction. The truncated model follows an optimal discretisation approach with a fine grid near the wellbores and a coarse grid in the outer regions of the reservoir [85]. The grid for the truncated model is shown in Figure 6, including the reservoir in brown, barrier region in pink, and aquifer in green. Note that the pumped wellbore experiments in this study do not account for heat loss. This is acknowledged as a limitation in the current research and is also a common characteristic of simulators utilised in the geothermal industry. Subsequent research endeavours will address this limitation by integrating more refined heat loss modelling, leveraging an enhanced model currently in development. The wellbore model encompasses overburden rock through the application of a coarse grid. The coupled simulation employs the Hasan and Kabir correlation to accurately represent heat transfer phenomena.

3. Results

The results of the energy extraction simulations are presented in this section. The cases considered focus on low-enthalpy energy extraction, where the production well is pumped over the simulation time. The power output is calculated based on full reinjection at a constant temperature of 50 °C and a power cycle efficiency of 10%.

3.1. Energy Extraction

Figure 13 shows a plot of the average gross power, required pumping power, and the remaining net power over the entire simulation time for a range of production rates. As anticipated, gross power output increases linearly with production rate under the condition of stable wellhead temperature. However, achieving higher mass flow rates necessitates active pumping, which introduces a corresponding rise in energy consumption for fluid lift and circulation. This additional energy demand manifests as a non-linear increase in pumping power with respect to mass flow rate, primarily due to escalating pressure losses in the wellbore and reservoir system.

The net power output initially improves with increasing flow rate, indicating that moderate pumping efforts are offset by gains in thermal energy recovery. This trend continues up to a production rate of approximately 70 kg/s, beyond which the system enters a regime of diminishing returns. At higher rates, the pumping energy required increases disproportionately, ultimately reducing the net power output. Notably, at flow rates exceeding 110 kg/s, the simulations reveal that the pressure support required surpasses 70 bar—the practical upper limit for commercially available pumps—rendering such operational conditions technically infeasible.

This performance inflection point confirms a critical design threshold in geothermal recovery systems: while elevated production rates can enhance thermal output, they also intensify parasitic energy demands in a non-linear manner. This underscores the necessity of optimising production flow to balance thermal energy gains against mechanical energy losses, a trade-off that defines the economic and operational viability of geothermal exploitation in petroleum reservoirs.

3.2. Doublet Spacing

Figure 14 presents the relationship between average power output and doublet spacing—defined as the horizontal separation between the injection and production wells—under a constant production rate of 100 kg/s. The results reveal a pronounced sensitivity of system performance to well spacing, with both net power output and pumping efficiency exhibiting strong non-linear responses. As the doublet spacing increases from 300 m to 900 m, the average pumping power requirement declines significantly from 504.87 kW to 421.35 kW—a 16.5% reduction—while the net power output more than doubles, rising from 423.21 kW to 789.73 kW.

This performance enhancement is primarily attributed to a reduction in reinjection-induced thermal breakthrough and improved hydraulic gradients across the reservoir. Greater well spacing mitigates rapid cooling near the production wellbore, thereby preserving higher fluid temperatures over time and reducing the viscosity-driven pressure losses that elevate pumping demands. Although the decline in pumping power is modest in absolute terms, the resulting increase in net power is substantial, demonstrating that even incremental increases in doublet spacing can yield disproportionately large gains in energy recovery.

The gross power output for a doublet spacing of 500 m ranges from 884.56 kW to 2504.7 kW, with pumping power requirements spanning 461.36 kW to 182.67 kW, depending on the evolving thermal conditions within the reservoir. These findings reinforce the importance of optimising well placement during the design phase, as appropriate spacing not only improves long-term system efficiency but also delays thermal breakthrough, extending the productive life of the geothermal system.

To isolate the respective contributions of pressure gradients and thermal effects on pumping requirements, supplementary simulations were performed using a constant mass flow rate of 50 kg/s and a short well spacing of 100 m. In one scenario, the reinjection temperature matched the reservoir temperature, while in the other, a colder reinjection fluid was used. In the former case, pump power demand remained stable throughout the simulation, confirming that thermal conditions did not alter hydraulic resistance. In contrast, the colder reinjection case showed a progressive increase in pumping power as the thermal front advanced toward the production wellbore. This clearly demonstrates that cooling effects—not just pressure differentials—substantially affect long-term pump duty, ultimately reducing net energy recovery over time.

Figure 15 illustrates the final position of the thermal front for various production rates at a reservoir temperature of 150 °C and a doublet spacing of 900 m. At higher production rates, such as 125 kg/s, the thermal front approaches the production wellbore significantly faster, accelerating reservoir cooling and decreasing fluid temperatures at the surface. Conversely, at lower flow rates like 50 kg/s, thermal front propagation is slower, resulting in more thermally stable production conditions. These findings confirm that production duration and flow rate are critical parameters influencing surface temperatures and associated pumping energy demands.

Figure 16 substantiates this by depicting the temporal evolution of pumping power at different flow rates. For all cases, once initial pressure transients subside, a steady increase in pumping power is observed, which aligns with the progression of the thermal front and resulting viscosity changes in the production fluid. Notably, moving from 50 kg/s to 75 kg/s results in a 4.8-fold increase in pumping power at the end of the simulation, while the increase is 2.3-fold between 75 kg/s and 100 kg/s—highlighting a diminishing rate of increase due to non-linear thermal behaviour.

To explain these trends, Figure 17 presents the thermal decline in production fluid temperature over time. For flow rates of 50 kg/s, 75 kg/s, and 100 kg/s, the end-of-simulation temperatures are 76.65 °C, 72.37 °C, and 69.4 °C, respectively. These temperatures reflect significant reservoir cooling and are particularly relevant when assessed against the minimum temperature thresholds for viable geothermal power production—e.g., 78 °C at Chena Hot Springs [61,86] and 70 °C in the LOW-BIN project [87,88]. Flow regimes that cause production temperatures to fall below these limits may preclude electricity generation, though they may still support direct-use applications.

3.3. Field Size

In the context of field size, Figure 18 illustrates the pumping rates over production time for a range of distances between the doublets (production and injection wellbores). The figure shows that pumping power decreases as the injection wellbore moves farther away from the producer. Additionally, the maximum pumping power is required when the injection wellbore is closest to the production wellbore. The least pumping power (522.5 kW at the end of the production period) corresponds to the largest distance between the two wellbores. The power needed to maintain production flow increases non-linearly with decreasing well separation distance. The dominant factor influencing the required pumping power is the fluid property effect.

Figure 19 provides information about the variation in bottom-hole temperature over the simulation period, indicating the impact of reinjected cold water on the temperature at the feed zone of the production well. The plot suggests a temperature drop across the range of spacing values between the doublets. The temperature drop is maximum in the case of closely located wells, with a drop of 63.2 °C at the end of the production lifetime for a spacing of 300 m. Only the case with wellbores located farthest from each other shows a temperature drop of less than 10% at the end of the production period. This result indicates the impact of field size on the extracted energy, where a smaller field area experiences a greater temperature drop, leading to an increase in the required pump duty and consequently, a lower net power output.

3.4. Pumping Power Correlation

In terms of pumping power correlation, based on optimal values for the set of parameters, a relationship can be established between the required pumping power and the production mass flow rate. Figure 20 shows a derived correlation between the pumping power requirement and the production rate from the doublet system considered in the study. A polynomial function with a degree of two matches the data obtained from energy extraction simulations. The figure demonstrates that pumping power quadratically increases with a linear increase in production rate. A similar correlation can be obtained for an optimised value of production rate in the field as a function of the spacing between the wellbores. This correlation corresponds to simulation values for the parameters in the base case.

3.5. Hotter Reservoirs

The influence of reservoir temperature on geothermal system performance was further evaluated by comparing the energy extraction trends at resource temperatures of 120 °C and 150 °C, as shown in Figure 21 and Figure 22. While both temperature scenarios follow a similar qualitative trend in gross power, net power, and pumping power as previously observed for 90 °C, key differences underscore the critical role of fluid properties at elevated temperatures.

At 120 °C, a production flow rate of 50 kg/s does not require active pumping, indicating not only delayed thermal breakthrough but also more favourable fluid dynamics that facilitate self-driven flow. This is in sharp contrast to colder reservoirs where pumping is essential, even at lower flow rates. The observed reduction in pumping requirement can be attributed to the decrease in fluid viscosity and density with increasing temperature. Specifically, at 291 bars, water at 90 °C exhibits a viscosity of 3.22 × 10⁻⁴ Pa·s and density of 978.08 kg/m³, while at 120 °C, the viscosity drops to 2.39 × 10⁻⁴ Pa·s and density to 956.94 kg/m³. This results in a more than 30% increase in the density-to-viscosity ratio, a key parameter governing flow resistance and pumping demand. These findings align with prior studies highlighting the dependence of pump energy on fluid properties [89,90].

The performance advantages of even hotter reservoirs are further demonstrated at 150 °C. In this case, a higher flow rate of 140 kg/s can be sustained with a lower average pumping power of 638.13 kW, compared with 688 kW being required to sustain 130 kg/s at 120 °C. This not only validates the thermophysical benefits of high-temperature systems but also highlights their superior net power generation potential. Notably, the decline in net power output begins at higher flow rates for hotter reservoirs, suggesting a wider operational envelope before encountering diminishing returns due to rising pump energy demands.

Geothermal energy from oil and gas fields at higher temperatures (>120 °C) can play a crucial role in decarbonising the energy sector and addressing one of the main challenges of the future [25,91]. In New Zealand, this can enable the energy sector to achieve its targets set in the Emissions Reduction Plan of decarbonising process heat and reducing industrial emissions to move toward a low-emission climate-resilient future [92,93,94]. Figure 23 demonstrates how reservoir temperature affects the average gross power, net power, and power required to run the pump for the base case in the study. These results correspond to a constant reinjection temperature (50 °C) and power conversion efficiency (10%). The figure shows that pumping power decreases with increasing reservoir temperature, and the useful net output approaches the obtained gross power at the surface.

The trend in average energy output from a reservoir at a specific temperature strongly depends on the fluid properties. Figure 24 displays the average power supplied to the pump for producing a range of mass flow rates from reservoirs at different temperatures. The results exhibit a non-linear trend in the amount of pumping power required to produce fluid, which increases with increasing flow rate. For instance, at a mass flow rate of 110 kg/s, a relatively colder reservoir (90 °C) would require more power (602 kW) compared with a hotter reservoir (150 °C) requiring 272.7 kW to produce the same mass flow rate. In other words, a higher production mass flow rate can be obtained from a hotter reservoir by installing a smaller duty pump.

3.6. Reservoir Location

The permeability-height (kh) of a reservoir represents its flow potential, and the pumping power required is a non-linear function of this flow potential. Figure 25 illustrates that pumping power decreases with an increasing value of the permeability-height product. For example, considering a permeability-height product magnitude of 10 × 10⁻¹² m³, a reservoir at a lower temperature (90 °C) would require more pumping power (349 kW) compared with a reservoir at a higher temperature (150 °C) requiring less pumping power (112.8 kW). These results have implications for the location of the reservoir below ground.

As we proceed deeper into the ground, the permeability generally decreases, but the temperature increases due to a conductive thermal gradient. This means that as the reservoir depth increases, the pumping power required decreases, and the overall amount of useful energy extracted increases. For instance, a reservoir at 90 °C located at shallower depths corresponding to a higher permeability-height product value (15 × 10⁻¹² m³) would require more pumping power (238 kW) compared with a deeper reservoir corresponding to a higher temperature (150 °C) and a lower permeability-height product value (7.5 × 10⁻¹² m³). Therefore, the location of the reservoir below ground is an important factor to consider as it affects the flow potential and the pumping power required for energy extraction.

4. Summary

This study demonstrates the technical feasibility and optimisation potential of extracting geothermal energy from depleted petroleum reservoirs, particularly emphasising the complex interplay between production rate, doublet spacing, reservoir temperature, and pumping power requirements. A key finding is the non-linear relationship between net power output and production rate, where increasing mass flow initially enhances energy yield, but beyond 70 kg/s, the rising pumping demand begins to erode the net power benefit. This inflection point is critical for operational planning, as production rates beyond 110 kg/s are unsustainable without exceeding the capabilities of commercially available pumps.

Doublet spacing emerges as a dominant factor influencing long-term performance. Increasing the distance between injector and producer wells delays thermal breakthrough and significantly reduces pumping requirements. Despite only a 16.5% reduction in pumping power from 300 m to 900 m spacing, the net power output nearly doubles, indicating a disproportionate gain in system efficiency. The thermal front simulations corroborate that slower cooling near the production well preserves surface temperature and pump efficiency over time.

The duration of energy extraction introduces further complexity, as the progressive decline in wellhead temperature due to reinjection cooling increases the pumping burden. Over the 35-year simulation, the required pump power increases up to five-fold for some scenarios, reinforcing the importance of long-term thermal management and well placement strategies.

From a design standpoint, reservoir temperature has a substantial impact on the pumping power needed for a given mass flow rate. At 150 °C, the same flow rate requires less than half the pumping power compared with a 90 °C reservoir. This is attributed to more favourable fluid properties—specifically, a higher density-to-viscosity ratio—at elevated temperatures. Importantly, hotter reservoirs may enable economically viable production at higher flow rates using lower-duty pumps.

Finally, reservoir depth and location, as characterised by the permeability-height (kh) product, influence flow potential and energy extraction efficiency. Shallower, cooler reservoirs may have higher kh values but require more pumping energy, while deeper, hotter reservoirs can yield greater net energy despite reduced permeability.

In terms of environmental impact, geothermal systems repurposing petroleum fields provide a low-carbon alternative for energy production. By leveraging existing infrastructure and extracting residual thermal energy, such systems minimise surface disruption, reduce emissions associated with fossil fuel extraction, and align with national and global decarbonisation goals.

Overall, the results emphasise that site-specific optimisation—balancing flow rate, reservoir temperature, spacing, and operational lifespan—is essential for achieving sustainable, efficient, and economically viable geothermal energy production. This integrated approach enhances resource longevity, reduces environmental impact, and ensures optimal use of both geological and engineered system parameters.

5. Conclusions

This study presents a detailed assessment of geothermal energy extraction from depleted petroleum reservoirs, offering new insights into how operational parameters—such as production rate, well spacing, reservoir temperature, and subsurface flow characteristics—affect system performance. One of the key contributions of this work is the identification of a non-linear trade-off between increased production rates and rising pumping power requirements, which results in a distinct threshold beyond which net power output begins to decline. This finding is critical for determining optimal flow conditions that maximise energy recovery while minimising energy losses to pumping.

Another significant outcome is the demonstration that doublet spacing plays a pivotal role in system efficiency. Increasing the distance between injection and production wells reduces thermal breakthrough and pumping demand, resulting in a substantial improvement in net power output. Importantly, these efficiency gains are achieved with only modest increases in well spacing, making it a practical and impactful design consideration.

This study also highlights how long-term cooling effects around the production wellbore progressively increase pumping requirements, especially in systems with shorter well spacing or colder reinjection temperatures. This underscores the importance of considering the thermal evolution of the reservoir over time, not just initial performance.

Additionally, the work introduces empirical correlations that link pumping power requirements to production flow rates and reservoir temperatures, offering valuable tools for early-stage system design. The analysis shows that hotter reservoirs support higher flow rates with lower pumping energy due to more favourable fluid properties, making them more suitable for economically viable geothermal development.

Finally, the influence of reservoir depth and the permeability-height (kh) product is shown to be significant. Deeper, hotter reservoirs—despite having lower permeability—require less pumping energy and offer higher net energy potential, reinforcing the importance of subsurface characterisation in site selection.

In summary, this study advances the understanding of geothermal energy recovery from oil and gas fields by establishing key design thresholds, identifying efficiency trade-offs, and providing practical insights to optimise long-term energy extraction. These findings contribute to the broader effort of integrating geothermal solutions into sustainable energy strategies, particularly for decarbonising industrial heat demand.

6. Future Work

• In future work, the focus could be shifted towards evaluating the geothermal energy extraction potential of gas fields, which are more abundant in New Zealand compared with oil fields. Gas fields have their own unique characteristics and challenges, and studying the feasibility and methodology for extracting geothermal energy from such fields would be valuable. This could involve investigating enhanced geothermal system (EGS) techniques that do not require pre-stimulation of the reservoir.
• Another avenue for future research is exploring the use of supercritical carbon dioxide (sCO2) as a heat transfer fluid in geothermal energy extraction from shallower reservoirs in oil and gas fields. While shallower reservoirs may have higher permeability, they typically have lower resource temperatures. sCO2 possesses properties that make it an attractive fluid for geothermal energy extraction, such as high heat-carrying capacity and a flowing potential like water. Additionally, sCO2 can exhibit a thermo-syphon effect, which means that no pumping power is required under certain conditions. However, the success of this approach may depend on ambient temperature conditions and the ability to cool the CO2 below its critical temperature.