Technoeconomic Design of a Geothermal-Enabled Cold Climate Zero Energy Community

Design for zero energy buildings and communities focuses on identifying cost-effective packages of energy efficiency measures and solar photovoltaics (PV) such that the PV can produce as much energy as the site consumes in a typical year. Compared with moderate and hot climates, cold climate zero energy design is economically challenged by higher building heating demands and lower solar resources. There is opportunity to enhance such communities by also incorporating geothermal electric and thermal energy production, coupled with thermal energy storage (TES).

We previously developed a design methodology to evaluate technical and economic feasibility of using geothermal power generation for a zero energy community use case [1]. In that work, geothermal was able to compete with rooftop PV, but neither community-scale PV nor electric utility service. Here, we extend that the prior work by adding geothermal thermal direct use and thermal energy storage to the design methodology to examine if geothermal cogeneration (heat and electricity production, with associated energy storage) can deliver better cost-effectiveness. Thermal energy storage and distribution systems are scoped to enable increased dispatchability of geothermal electricity generation while also meeting time-varying thermal demands.

Geothermal direct-use and cogeneration has been adopted elsewhere in the world, such as Iceland where the practice is not uncommon [2]. In North America, due to several factors including low electricity prices, geothermal has not been commonly applied in this way. (Note that this work focuses on higher temperature geothermal resources useful for electricity generation—to enable zero energy construction—and not geothermal heat pumps.)

The U.S. Department of Energy (DOE) defines a zero energy community: “an energy-efficient community where, on a source energy basis, the actual annual delivered energy is less than or equal to the on-site renewable exported energy” [3]. Today’s zero energy design practice focuses on identifying cost-effective packages of energy efficiency measures and PV. Typically this practice focuses on identifying efficiency measures with a lower cost per kilowatt-hour (kWh) saved than the site’s PV cost per kWh produced. Energy analysis can then estimate the site’s annual energy demand, and PV was selected to produce an equal amount of energy through a typical year.

Reaching zero energy in cold northern climates is more economically challenging than in moderate and hot regions because building heating demands are higher, and solar resources are typically lower. Therefore, zero energy buildings require higher levels of efficiency and more renewable generation resulting in significantly higher costs for end users. Here, we present a novel approach that adds geothermal cogeneration supply to support electric and thermal energy demands.

The United States is the largest generator of geothermal electricity in the world [4,5]. In particular, geothermal electric power plants are concentrated in the western United States, with the majority residing in California, Nevada, Oregon, and Idaho. In general, the western states are a key geographical region for the development of naturally occurring, hydrothermal resources (Fig. 1). At the intersection of this region of increased attractiveness for geothermal electricity production and an area with a relatively cold climate, we focus on Pocatello, Idaho, to design and analyze a hypothetical zero energy community.

Although zero energy community design and analysis has precedence in both existing literature and practice, the consideration of geothermal energy in this context is very limited. Of primary relevance is the high capital cost of a geothermal power plant. Although highly dependent on the characteristics of the specific system, the overnight capital costs per kilowatt (kW) for a typical hydrothermal system can range from $5,000/kW to $30,000/kW. In addition, land and water use, regulatory, and technical (e.g., insufficient fluid temperature and/or flowrate) can challenge adoption of geothermal energy technologies.

To address some of these considerations, we assume the community will adopt a geothermal binary cycle power plant. Binary cycle plants are typically used for geothermal resources with temperatures below about 200 °C, as these temperatures cannot efficiently produce the large volumes of steam required for dry steam or flash steam plants. Binary cycle power plants use a heat exchanger to transfer the heat from geothermal fluids (e.g., water, steam, or both) to another liquid or “working fluid.” This working fluid then drives a generator turbine to produce electricity. These power plants typically use either organic Rankine cycle or Kalina cycles.

Recent developments in binary cycle technology have enabled exploitation of low-temperature geothermal resources. Power generation units are more modular and cost-efficient, and business models for the technology have allowed previously disregarded resources and applications to be reconsidered [7,8]. For zero energy community designs that incorporate geothermal, such technologies should help achieve techno-economic feasibility.

While the geothermal heat pump (GHP), also known as a ground-source heat pump, is an established heating, ventilating, and air conditioning (HVAC) technology for buildings, we did not include it in the analysis for the following reasons:

1. The input data required to simulate GHP products available today (model inputs/coefficients) are lacking.

2. The ground heat exchanger models are not trusted to accurately simulate heat exchange with the shallow earth for residential situations (mathematical formulations).

3. Site data on shallow earth heat transfer parameters (model inputs/coefficients) are unavailable.

4. Regional/local costs to drill and install GHP systems (economic inputs/coefficients) are unavailable.

5. Operation and maintenance (O&M) cost data for GHP systems over their life span are unavailable.

Several analysis tools, developed by National Renewable Energy Laboratory (NREL) and others, are combined in a novel workflow to design and simulate a geothermal-enabled zero energy community. The purpose of these simulations is to evaluate the total costs associated with different technology packages, with a focus on how geothermal electricity generation might compete with PV to supply the zero energy community. We first develop a set of building models using ResStock™ [9], BEopt™ [10], and EnergyPlus™ [11]. These building models are discussed in Sec. 2.2. As a demonstration of the design methodology, the baseline design presented here is based on established practices for zero energy design: residential building energy optimized designs using BEopt and commercial designs from a Denver-area suburban development [12]. (In an actual greenfield development, the representative buildings would be substituted with actual buildings planned for the community.)

Time-series electrical and thermal loads resulting from these baseline designs are input into REopt [13], which is then used to select, size, and dispatch energy generation and storage based on the community’s annual energy demand. Geothermal power, geothermal direct use heat, solar power, and grid energy resources are considered on the supply side and are constrained such that the community achieved zero energy on an annual basis.

REopt produces the levelized cost of energy (LCOE) for the selected supply-side assets. Because these energy prices differ from the prices assumed for initial efficiency design (i.e., if prices are higher, the building’s lifecycle cost will go down if it is more efficient), the community and its buildings are re-optimized under the new costs and updated loads produced. The techno-economic sizing and dispatch of generation are then assessed again. This process iterates until it converges to a consistent load and supply design. Typically, the process converges quickly and only one or two iterations were needed.

Community electric and thermal loads are simulated using a mix of archetype models for residential detached, residential attached, and commercial buildings. Capabilities from BEopt, the ResStock Analysis tool, and the State and Local Energy Data (SLED) tool are leveraged to establish a realistic set of buildings to represent a zero energy community. Load profiles are scaled based on the size of the desired community and breakdown of building type. Table 1 details the relative mix by floor area and total electric load for each archetype.

BEopt is whole-home simulation software that includes a geometry creation tool, a library of existing options that represent typical equipment and construction practices, and the ability to perform economic analysis. BEopt is used as a front end for EnergyPlus, DOE’s flagship building simulation software. BEopt also has the capability to parametrically optimize the building design, finding the lowest-cost design subject to energy costs and solar resource availability.

ResStock characterizes and models the entire United States housing stock with the use of a sampling routine and the EnergyPlus simulation software. Drawing from numerous data sources, ResStock samples across distributions of home characteristics and geographies to generate models that account for the diversity within the US residential building stock.

SLED [14] is a data analysis tool that provides energy use and activity on the census tract level. By leveraging data from utilities, real estate information, and census survey data, SLED provides a breakdown of energy usage used to calibrate the mix of residential and commercial buildings for this case study.

The residential building models include large, midsized, and small detached homes, as well as a five-unit townhome and a duplex, as shown in Fig. 3. The baseline design for homes were created using an amended version of the 2012 International Energy Conservation Code (IECC) adopted by Idaho [15,16] and the ResStock Analysis Tool [9]. Construction and operational characteristics not outlined in the IECC were determined using the sampling capabilities of ResStock, which provided inputs for highly representative sample homes specific to new constructions in Pocatello, Idaho.

Residential homes in the all-electric scenario and the electric business-as-usual scenario use minisplit heat pumps with backup electric resistance heaters to meet building HVAC loads. The homes in the direct-use thermal scenario utilize a ducted hydronic air handling unit for space heating and indirect water heaters for water heating.

To establish cost-optimal efficiency options for the zero energy residences, we used BEopt’s optimization capability to minimize annualized energy costs and optimize discrete energy efficiency options in each residential building. A package of energy efficiency options was selected at the cost-optimal point, which often coincided with the point in which PV is more cost-effective than additional energy efficiency options. These optimization runs provide zero energy ready building models while accounting for future energy costs to the homeowner. Optimization variables included improvements in insulation, air leakage, water heating, lighting, appliances, and windows. The average electricity rate in Idaho, $0.0952/kWh, was used for the initial optimization, but was updated after each iteration of sizing and dispatch optimization, as described in Sec. 2.1.

The four commercial buildings are patterned after an existing suite of building models from a zero energy community under development in Denver, Colorado [12], shown in Fig. 4. These buildings have been designed to meet zero energy when paired with PV and in a climate similar to Pocatello, Idaho. These models simulate diverse energy demands specific to building uses such as a community center, grocery stores, retail shops, or restaurants.

The community in this study is suburban by design, composed of 90% residential buildings and 10% commercial buildings by area, as presented in Table 1. The mix of residential building types is chosen based on potential for energy efficiency, and therefore half of all dwellings are attached (duplex or townhome). The commercial buildings are chosen as reasonable building types for a suburban community, while allowing the energy demand to be primarily driven by residential buildings. SLED [14] helps to calibrate the mix of residential and commercial buildings. SLED was queried to find the ratio of residential to commercial electric and gas use for Pocatello, Idaho, and three other similar cities in the area. These ratios were averaged to find an appropriate mix of residential and commercial buildings to calibrate the community profile. Because the load of the existing building stock is composed of a mix of electric and gas loads, we individually scale nine load profiles from mixed-fuel residential and commercial buildings to establish the floor area fractions, presented in Table 1. This mix of buildings is applied to the all-electric and district heating scenarios. Although this process does not provide the granularity to capture and calibrate hourly energy peaks, the community design is in line with standard suburban communities in Idaho for yearly energy use.

Each building model is simulated using EnergyPlus at 15-min time-steps and the Typical Meteorological Year, version 3 (TMY3) weather file for Pocatello, Idaho. An hourly load profile is generated for a year, which is used for the entire 30-year analysis discussed in Sec. 2.3.

Following simulation of all building models, postprocessing steps are taken to maintain smooth load profiles when scaling to the community level. This is necessary to be able to represent the actual diversity in energy use between buildings. Water heating loads are of particular concern when scaling from a small number of residential buildings because of the sudden and intermittent hot water draws in these buildings. We remedy this by generating hourly weighting factors using an average water heater profile in Pocatello, Idaho [17], and then multiplying these by the total yearly water heater energy for each building. The average water heater load profile is generated with models based on Building America House Simulation Protocol [18] and the Residential Energy Consumption Survey [19]. In addition to water heater loads, whole building loads were smoothed by staggering each profile up to 2 h forward and backward, then summing. By taking these steps, we maintain a more realistic load profile for a community with many buildings.

The load fractions for each building type are presented in Table 1. The space and water heating equipment for the direct-use scenario uses heat exchangers to provide heat from the geothermal plant, while the all-electric case uses electric equipment to meet thermal demand. The differing equipment and energy rates for the direct-use case impact the BEopt optimization and the overall energy requirements of the community, resulting in a different set of building models than the all-electric case. The community thermal load is primarily a function of the residential building demand, which accounts for 83% of the load in the direct-use thermal case. With demand for the two scenarios modeled, the yearly load profiles can be further scaled based on the economics of the power plant, as discussed in Sec. 3.2.

We performed analysis to develop energy supply requirements, interconnection, sizing, and operation for geothermal and solar energy generation, storage, and distribution of electrical and thermal energy.

To understand the local geothermal resource in Pocatello, a full-scale resource assessment would typically be necessary at the techno-economic modeling stage. For this preliminary study, a literature search was instead conducted to determine the subsurface parameters, such as depth of resource, temperature, and flowrate. Reference [20] examined the Tyhee area of Idaho, located directly north of Pocatello. The authors conducted gravity, magnetic, geochemical surveys, and shallow well observations to determine that the temperature-at-depth profile exhibited lower gradients in the upper sections of several wells and gradients of greater magnitudes in the deeper parts of those wells. The authors stated that these various subsurface measurements indicated a geothermal gradient of 60 °C/kilometer (km). Reference [21] investigated the viability of using geothermal heat at a barley processing plant in northwest Pocatello. Data collected for the study indicated the existence of a geothermal reservoir in the 66 °C to 121 °C range in the Tyhee area. From these measurements and other supporting data, the authors estimated both a 146 °C/km gradient as well as a 169 °C/km gradient at depths below 34 and 122 m, respectively. They concluded that a 110 °C resource could be expected at a depth between 609 to 762 m. Reference [22] provided a summary and analysis of shallow (≤3 km), low-temperature (30–150 °C) geothermal resources in the United States. By using data from known hydrothermal resources and bottom hole temperature measurements from oil, gas, and water wells, the authors performed a geostatistical analysis to estimate the temperature at depth for the entirety of the United States. This data set indicated a mean resource temperature of approximately 120 °C at a 3000 m depth, with a standard deviation of approximately 9 °C, at various locations in Pocatello.

While the reports by Refs. [20] and [21] indicate extremely favorable geothermal gradient estimates, projecting these same gradients to depths greater than 1000 m is prohibitively uncertain. Nonetheless, by using these optimistic reports in tandem with the geostatistical estimates of temperature at depth by Ref. [22], the analysis presented here considers a “base case” geothermal resource of 130 °C at a 3000-m depth, reflecting one standard deviation above the mean resource temperature. In addition, given the favorable nature of the aforementioned resource assessments, this analysis will also consider an “optimistic” case in which the geothermal resource is assumed to be 150 °C at a 2500-m depth, reflecting the 60 °C/km geothermal gradient estimated by Ref. [20], capped at the upper bound of the temperature range for binary power plants analyzed in Ref. [23].

The cost basis for much of this study follows the analysis performed by Ref. [23], which explored economic feasibility of electricity production from binary power plants in low-temperature geothermal areas with fluid temperature below 150 °C. Importantly, the report included supply curves (i.e., system size (kW) vs total normalized installed cost ($/kW)) that are used in REopt.

The Verkis report split the total installed cost into two subcategories: (1) power plant costs, including mechanical equipment, electrical and controls equipment, and civil work; and (2) geothermal field costs, including wells, pumps, and gathering and reinjection systems. The geothermal field costs included the assumptions that the field would consist of two production wells and one injection well all drilled to a depth of 1400 m and capable of producing 40 l/s, resulting in a normalized cost of 100,500 $/(l/s).

We scale this geothermal steam field cost to reflect the depths identified for both the “base case” and “optimistic” case. Lacking further detail on the calculation of the well cost in Ref. [23], the well cost was instead determined using cost correlations developed for the US Department of Energy’s Geothermal Vision study [24]. In particular, the correlation for a small diameter, vertical open-hole well was used to calculate well costs for the “base case” depth of 3000 m and the “optimistic” case depth of 2500 m. The normalized field cost was then recalculated using these new cost estimates for the wells, but using the same assumed costs for the gathering system and reinjection system.

REopt uses the PVWatts application [25] to calculate the electricity production of installed PV systems. Refer to the PVWatts technical reference manual for further modeling assumptions and descriptions [25]. Table 2 summarizes the main technology assumptions (based on PVWatts defaults) and data sources for PV modeling in REopt.

This research considers the community to be electrically connected to the power grid. Energy can be imported and exported. Zero energy buildings and communities will have equal import and export, on an annual basis, of electric kWh if they are all electric. There is no commonly accepted approach for determining the local and/or time-dependent source-to-site conversion factor, so in this work, we use the national annual average source-to-site ratio [26] for utility-supplied electricity.

We then performed combined system optimization to size, site, interconnect, and operate the geothermal-enabled zero energy community for maximum cost-effectiveness under appropriate performance constraints.

The geothermal and PV generation (supply) system sizes and dispatch are modeled in REopt. REopt is a techno-economic optimization model formulated as a mixed-integer linear program that selects the size and determines the optimal dispatch of distributed energy resources for a specific building or campus located behind the meter [13]. The most common objective function of the model is to minimize the life cycle cost (LCC) of energy. For additional details on the REopt tool, please refer to Ref. [13]. For this analysis, the constraint of zero energy electric consumption is activated; electric exports offset electric imports from the grid on a one-to-one basis.

Several new components were added to REopt to evaluate geothermal-based electric and thermal supply. For the all-electric load scenarios studied previously, only the additions of geothermal well-field and electric generating plant were needed. For these direct-use thermal scenarios, a geothermal-heated hot water production plant, TES, and district heat distribution network components were added to the optimization. The TES was modeled as a large hot water tank. The LCC of the energy for the community included the cost differences for building energy efficiency measures and end-use thermal equipment to compare the all-electric scenarios to the geothermal district heating and PV supply scenarios.

A simplified community supply and load diagram is shown in Fig. 5. The direct-use thermal supply includes a hot water production plant and thermal storage tank co-located with the electric plant and a district heating distribution network to supply the buildings with hot water for space and hot water heating needs.

The life cycle cost analysis parameters are shown in Table 3. All cost values are reported in 2019 dollars. The analysis period is 30 years with a discount rate of 5%. No tax credits or other subsidies were included. Further details explaining the calculation of the utility rate used in the model can be found in Sec. 2.5.1. The utility electricity rate was calculated by using a load-weighted average retail and commercial electricity from the local utility. The values for general inflation and electricity rate escalation are average values from 2019 to 2049 from the Energy Information Administration 2019 Annual Energy Outlook [27]. The electricity sell-back rate—the credit for electricity exports to the grid under the no-net-metered scenario—is estimated to be 30% of the base electricity rate (e.g., $0.03/kWh). If the system requires an import or export of greater than 10 megawatts electric (MWe), the community is required to pay an interconnection fee of $33/kW based on the maximum imported or exported electricity in addition to electric grid infrastructure upgrade charge [28]. This is most relevant to PV, which requires a much larger system-rated power than the community loads to achieve zero energy.

An overview of the ReOpt modeling and input electrical price used was described previously. The model only had the capability for a single structure; however, Pocatello, Idaho, has a structured rate price depending on time of year and electricity usage. Rates used are available at Ref. [29] for residential and [30] for commercial.

To determine the price to input for the model, the energy demand model from BEopt was obtained. Using the energy demand for the nine building scenarios, the total energy cost for the summer and nonsummer period for each building was obtained. The average cost, including the necessary fixed fees, was then divided by the total energy consumption in those periods to determine a price per kWh of electricity. Then using the total electric load fraction from Table 1, the appropriate contribution to the overall rate was calculated. Table 4 shows the results.

There are costs associated with installing the infrastructure necessary to support using thermal energy and costs associated with deferred infrastructure upgrades otherwise necessary (e.g., if natural gas was installed).

Because the scenario under consideration was installation of a new community and not a retrofit, the costs associated with trenching were negated, as it was already a necessary part of construction regardless of whether using thermal energy from the geothermal resource or installing natural gas lines. This meant the infrastructure costs were limited to distribution piping and any pumps necessary. An extensive literature review and interviews with experts who have implemented district heat solutions determined that pumping outside the energy center would not be necessary [31,32]. The costs then would be associated with designing, commissioning, site-specific work, distribution piping, and the necessary specific electrical and mechanical work. The thermal exchange loop is a closed system. Corrosion and scaling inhibitors are expected to be included in the circulating fluid, thus scaling prevention is a capital cost and not an O&M cost [33]. As outside pumping is not needed, the O&M costs are accounted for in the electric generating plant O&M costs. Data from Tomberlin and Salter indicated that actual project costs were within 5% of expected. The capital costs for the hot water heat exchanger plant and district heating distribution network shown in Fig. 8 were used for the evaluation of the economic feasibility in the modeling. The annual fixed O&M for the district heating distribution network is estimated to be 3.5% of the installed capital cost of the district heating distribution network.

Capital and O&M cost estimates for geothermal electric generating plants are taken from [23], while the well-field cost to drill to a depth of 3000 m is taken from the concurrent NREL GeoVision [24] study to extract the 130 °C resource temperature from Pocatello, Idaho. The capital cost for the base and optimistic geothermal resource estimates are shown in Fig. 6. The costs are reported on a net-power-production basis, including plant parasitic loads and well pumping power requirements. The main driver for the increased cost at smaller plant sizes is the lower efficiency of these plants, which requires more parasitic and pumping power relative to the electric power production.

The fixed O&M cost is modeled as a function of system size, as shown in Fig. 7. O&M costs are inclusive of personnel (salary for operators, laborers, and security), spare parts and plant consumables, and well replacement. We share the assumption with Ref. [23] that scheduled maintenance occurs once per year and is represented as 1.6% of the capital cost. Furthermore, we assume that additional costs for scaling control within both the well and plant are not required, as it is assumed that the chemical composition of the geothermal fluid was taken into account during the design phase. Scale prevention methods would be site specific as different types of scales are found in various geothermal areas, and it is possible that an additional O&M cost would be require at a site given its composition [23].

The high fixed costs for plants less than 1000 kW are a result of paying the salary for the minimum number of operators at the plant, whereas no additional operators are needed until the plant size exceeds the 1000 kW electric (kWe²) range. The variable O&M is assumed to be constant with respect to system size with a value of $0.027/kWh.

The major subsystems of the direct-use thermal system include the hot water production plant (heat exchanger and balance of plant), thermal storage (if considered and cost-effective), and district thermal distribution network (piping, valves, pumps, and so on).

The hot water production plant produces hot water by exchanging heat with the geothermal brine, and it includes a heat exchanger and balance of plant equipment such as valves, pumps, and plumbing. The district heating distribution network delivers the hot water to the community buildings, and this includes main feeder pipes, branch pipes that go to individual buildings, valves, pumps, and so on. The end-use heat exchange equipment building for space heating or hot water heating is accounted for in building costs.

A simplified model is used for the thermal energy storage tank, with an assumed thermocline separating the hot and cold sides of the tank. The cold return water from the hot water loop is assumed to be delivered back to the hot water production plant and/or the cold side of the hot water tank at the same temperature, independent of load. In this case, the flowrate of water to the community would be modulated in proportion to the load, and transient losses in the distribution network are ignored.

Although transient heat losses and temperature variation are ignored, it is assumed that 15% of the heat transferred from the geothermal brine to the hot water is lost to the environment from a combination of production plant and distribution network heat loss. There is an additional heat loss from the hot water tank, which is proportional to the amount of heat stored in the tank.

The PV capital and O&M cost estimates are based on 2018 community-scale PV array projections from the Annual Technology Baseline [34]. The capital cost sensitivity to system size is shown in Fig. 9. Commercial typically ranges from 100 kW to 1 megawatt (MW), with price increments at 100 kW, 200 kW, 500 kW, and 1 MW. Utility scale typically begins at 5 MW and shows incremental price differences at 5 MW, 10 MW, 50 MW, and 100 MW. The cost for a residential PV system is estimated at $2,897/kW, followed by the 100-kW commercial system at $2003/kW. The fixed O&M is assumed to be constant with respect to the system size with a value of $18.1/kW/yr although O&M for a residential rooftop system is $21/kW/yr [34].

Given that the scenario under question is a new community, using the commercial or utility scale prices makes sense as the builder could leverage the economies of scale for purchasing, designing, and installing the system. Equally, the builder could install residential fixed-tilt (25 deg) roof-mounted systems on each of the residences. In the event, the installation is disaggregated and installed on individual residences, clearly each system would operate behind the meter, thus justifying the sell-back pricing used throughout this report. The difference between a residential system and even a 100-kW commercial system is that residential is 44% more expensive than a commercial system installed and has approximately 15%/year higher maintenance costs. While the residential rooftop installation is representative of today’s “business as-usual” zero energy construction practice, community-scale PV is more comparable to the community-scale geothermal electricity generation. We provide both costs for the purposes of evaluating geothermal cost-effectiveness because both rooftop and community-scale PV have healthy sales volumes today.

The fixed and variable electricity costs for both the comparison and baseline case were from Idaho Power [35] and were further confirmed using the U.S. Utility Rate Database [36]. The savings realized from reducing the infrastructure upgrades necessary to connect the community were calculated using the model developed by Ref. [28].

This analysis for community-scale generation assumes that the supply resource (geothermal and/or PV) can effectively displace the aggregate community electric loads as if it were “behind the meter,” even though the plant would really be located in front of any single building meter. Specifically, the electric supply from the geothermal or PV plant can offset a portion or the entire retail electric rate that the community buildings would incur. The validity of this simplification will be investigated further in the future work. For the scenarios presented, electric generation sold back to the grid is credited at 30% of the retail electricity rate (no net-metering).

The scenarios analyzed (shown in Table 6) consider geothermal electricity supply and thermal distribution (with base and optimistic resource estimates) and PV (at residential and commercial scale) electric supply without net-metering contracts. The baseline community size of 1200 residential dwellings (with applicable commercial loads added) is presented in detail for the different scenarios because the average load of about 3 MWe is where the economy of scale of geothermal supply has flattened out [23]. The sensitivity of energy cost to community size is also presented.

The focus of the direct-use analysis is to investigate the opportunity for a geothermal resource to directly provide the significant heat loads of a cold climate community without converting all space and hot water heating needs to electric loads. Business as usual for this electric-plus-thermal zero energy community would be PV, sized so its exported electricity offsets all electric imports from the grid. The geothermal well and generator can deliver a mix of electricity and thermal energy to meet the needs of the community. Only no-net-metering scenarios are explicitly discussed to maintain the focus on geothermal solutions.

All results are presented for analysis of the systems using weather data and geothermal resource estimates for Pocatello, Idaho.

The case for geothermal electric and direct-use thermal supply was analyzed to contrast with a baseline community, which represent a business-as-usual scenario. The direct-use thermal supply leverages the economy of scale of the geothermal well-field that is needed for the electric generating plant, so it is a relatively small cost addition to increase the size of the well-field and add a heat exchanger to generate hot water. There is additional cost to the community for direct use, including thermal energy storage (hot water tank) and the district heating distribution network, but the economics of such additions are explored.

Similar to the all-electric case in Ref. [37], a baseline community was established to represent a business-as-usual scenario including buildings with PV (at either residential or community scales) and grid electricity. This baseline community requires different space and water heating equipment compared to the geothermal direct-use community. Building options for the code-minimum residential buildings were optimized in BEopt, resulting in an average source energy savings of approximately 19%. This is accomplished by improving wood stud wall insulation, basement insulation, air sealing, lighting, furnace efficiency, and water heater efficiency. The starting point for this optimization does not include a high-efficiency HVAC system like the all-electric baseline, which is the primary reason for the difference in source energy reductions. Even with these efficiency improvements, the total energy consumption (electric and thermal) is 30% higher than the all-electric community design.

Table 7 presents the results of the life cycle cost analysis for the direct-use scenarios described in Sec. 2.6. The first column has the reference case for the community supplied by electricity from the local utility; the net present value (NPV) of all the other scenarios is relative to this reference. The PV residential and PV commercial scenarios only differ in the capital and O&M costs, and the lower costs for community-scale PV achieve a positive NPV. The geothermal base scenario forces all of the thermal load to be supplied by the geothermal direct-use hot water in real time, and the geothermal electric plant just has to produce as much as the annual community electric load. The thermal energy is provided at a much lower rate than electric energy, as expected—it is about 12% of the LCOE of the electric generation. The large negative NPV for the geothermal base scenario is predominantly due to the high capital cost. The geothermal optimistic scenario supplies the same energy with roughly half of the capital cost, and this results in better economics than the PV residential scenario even with much higher (30%) energy needs of the community.

Figure 10 shows the life cycle cost breakdown graphically for all scenarios. This shows how the geothermal supply is insulated from increases in electricity prices in the future because of its small contribution to the life cycle costs. In contrast to geothermal, where the supply matches well with the load, the PV community scenarios has utility purchases accounting for nearly half of the total life cycle cost. It should be noted that PV is much more dependent on the local utility rates, where the life cycle cost would increase significantly if electricity prices increased.

A sensitivity of the combined electric and thermal LCOE to community size is shown in Fig. 11; the levelized cost of energy is a weighted average based on the electric and thermal supply. The base geothermal scenario achieves lower LCOE than the residential-scale PV scenario at a community size of about 1600 residential buildings. The optimistic geothermal scenario achieves a lower LCOE than community-scale PV at a community size of about 1200 residential buildings. The community-scale PV and optimistic geothermal scenarios achieve lower LCOE than the utility-supplied community at about 400 and 1100 residential buildings, respectively. The base geothermal scenario only approaches the LCOE of the utility-supplied community at sizes greater than 3600 residential buildings.

This study attempted to quantify the grid impacts of importing and exporting additional power by estimating the distribution service upgrade cost required if importing or exporting more than 10 MW. However, there may be larger consequences of the variable nature of PV and the large mismatch of electric supply and load that are not captured by this cost. This is especially concerning as more PV penetration exists on the grid and there becomes an excess of supply relative to demand during times of high solar resource and a grid stabilization challenge as the PV resource declines. Table 8 shows the metrics related to the amount of load that is supplied by the geothermal and PV resource in real time, and the peak import and export power that the supply exchanges with the local grid.

This illustrates the superior match between local supply and load for geothermal compared to PV, with PV exporting more than half of its production to the grid (and subsequently having to purchase the same amount at other times to achieve zero energy). PV has about three times the peak import/export power compared to geothermal. The bidirectional flow of energy on the grid serving these communities are depicted, on an hourly basis, for the electric and direct-use community in Fig. 12.

Resilience is not explicitly accounted for in the techno-economic analysis, but the variability of geothermal is much lower than PV, while the supply matches the load to a much greater extent, as illustrated in Sec. 3.4. The geothermal plant(s) have expected and unexpected maintenance requirements that could reduce their capacity, but with multiple modular plants running in parallel and proper maintenance, the geothermal supply should have minimal downtime. PV hardware has high reliability, but the variable resource on which it relies can result in significant production variation on the scale of seconds (e.g., a thick cloud blocks the sunlight), and this poses significant challenges to the local grid and other generation assets on which the grid relies.

This report shows that geothermal electricity generation can be cost competitive with PV in support of zero energy construction practice under a range of conditions. As the community size increases, economies of scale resulted in improved cost-effectiveness of the geothermal plant. Community-scale geothermal electricity has a lower LCOE compared to residential rooftop PV—the dominant zero energy construction practice today—for communities of roughly 1200 homes and larger; this is true even for very conservative estimates of geothermal resource potential. If high-temperature geothermal resources can be found, as in the optimistic scenario, geothermal is found to reach grid electricity cost parity at a community of about 1100 homes and comes very close to achieving cost parity with community-scale PV solar gardens as well. It would take only modest cost or performance improvements for geothermal to reach cost parity with community-scale PV if excellent geothermal resources are available.

Geothermal supply achieves competitive LCOE with PV using the optimistic geothermal resource estimates at the Pocatello, Idaho, site of 150 °C at a depth of 2500 m from Ref. [23]; this represents a cost target for geothermal of about $10,000/kWe_net, including the well-field and electric generating plant. The main driver of cost is the well-field (about 70% from the base resource estimates), but there are two aspects to lowering the well-field portion of the total installed cost:

• Lower the actual well-field drilling costs through improved techniques, reusing of existing wells from oil and gas drilling, or other means.

• Improve thermal-to-electric conversion efficiency of the generating plant, which would use less well-field relative to power output.

Thermal energy storage was seen to directly affect the well costs. By including thermal energy storage, the well was not required to be sized to meet the coincident peak electric and thermal demand, which permitted the well size to be reduced. Improvements in thermal energy storage performance and cost can therefore enhance the value proposition for geothermal technology’s electric and direct-use applications.

The installation of the geothermal plant in close proximity to the community has the potential to reduce the need to install grid infrastructure to support the new community. It was clear that overall power flow across the utility infrastructure—both kWh on an annual basis and maximum peak demand—were lower by approximately 66% for all geothermal zero energy community scenarios, compared to the business-as-usual communities. As a continuous generating resource, it can help overcome the “duck curve” generated by the solar resource [38]. Geothermal generation was better aligned with the energy demand of buildings.

In no case was significant battery capacity determined to be optimal. REopt selected a modest battery in some cases, mainly to balance demand costs and/or to participate in some price arbitrage in time of use rate scenarios. Thermal energy storage was found to be cost-effective in direct-use scenarios. It was exercised to balance out variable thermal demands from buildings with the geothermal well’s resource availability, reducing well requirements.

The ability for less geographically centralized communities to achieve zero energy should be greatly enhanced with the introduction of the geothermal plant because it is able to offset significant infrastructure investment that would otherwise be required.

Separate analysis, not presented in this study, evaluated the same community with a combined electric-plus-natural-gas utility supply. Because of the low cost of natural gas in the United States, direct-use thermal resources did not achieve cost-effectiveness with conservative assumptions about thermal production, storage, and distribution costs. In this case, geothermal energy supply could achieve market parity and cost-effectiveness if one or more technology performance improvements occur, costs are reduced, or policy incentives are enacted.

We did uncover an interesting challenge for geothermal applications at community scale. Because thermal energy is available at very low marginal cost, energy efficiency was less cost-effective leading to increased energy consumption. This largely offset the benefits of low LCOE due to higher energy use in the buildings. This merits further investigation.

We reemphasize that this work did not analyze GHPs and makes no claims about their potential cost-effectiveness in this application. We also did not review the application of central geothermal plants for an off-grid community; that also is left for the future work, but it is expected that the economics would be even more favorable for geothermal as a result of higher costs of power in small, isolated grids.

This report documents an initial study, of which the bulk of our effort was applied to establishing a multi-software-tool workflow that enables the desired analysis. Multiple enhancements could make the analysis methodology more robust and permit its application to real-world projects.

The authors would like to thank our colleagues and funders for enabling, advising, and reviewing this work. That list includes Arlene Anderson, Sheri Anstedt, Deanna Cook, Molly Hames, Scott Horowitz, Amanda Kolker, Shanti Pless, Ben Polly, Billy Roberts, David Roberts, Caity Smith, Linh Truong, Eric Wilson, Jeff Winick, and Kate Young. We appreciate permission from Colin Williams at USGS to reuse Fig. 1 in this article.

This work was authored by the National Renewable Energy Laboratory (NREL), operated by Alliance for Sustainable Energy, LLC, for the U.S. Department of Energy DOE under Contract No. DE-AC36-08GO28308. Funding provided by DOE’s Office of Energy Efficiency and Renewable Energy, Geothermal Technologies Office. The views expressed herein do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes.

The authors are employed by the same institution as the Editor of this Special Section Issue and Associate Editor, of ASME Journal of Energy Resources Technology.