Hydropower and Pumped Storage Hydropower Resource Review and Assessment for Alaska’s Railbelt Transmission System

Abstract

The Alaska Railbelt transmission system runs from Fairbanks to Anchorage to Homer and supplies 75% of the state’s population with power. In the near future, this system will experience significant increases in load due to electrification of the transportation and heating sectors. To account for this, several state organizations are working towards the creation of an integrated resource plan and reliability standards.This work encompasses the efforts of researching the operations, cost, and locations of desirable hydropower and pumped storage hydropower (PSH) resources in the areas surrounding the Railbelt transmission system. The aspects of conventional hydropower and PSH as well as adjustable-speed and ternary PSH were analyzed. With Alaska’s diverse and rugged landscape, QGIS was utilized to delineate the positions of energy resources within a reasonable distance from the Railbelt. By incorporating Digital Terrain Models and the QGIS processing toolbox, a least cost path analysis was completed to filter out resources within the designated distance of the Railbelt. Applying existing cost models to the data in this work helped to decide the energy resources that would be studied further. The future of this project includes modeling the selected energy sources in the PSS/e Railbelt model to examine their effects on the reliability and stability.

1. Introduction

In the United States, the federal government has set a net zero carbon emissions goal for 2050 [1]. Hydropower currently provides 38% of the renewable generation in the United States, and pumped storage hydro (PSH) currently provides 93% of grid storage [2]. Hydropower and PSH are currently key components of the net zero carbon emission goal for the United States. The state of Alaska does not currently have any state carbon emission goals; however, individual electric utilities do have carbon reduction goals, and hydropower plays in an important role in energy generation in the state.

Conventional hydropower first began its United States utility-scale journey in the early 1880s [3]. In 2019, there was a hydropower installed capacity of 80.25 gigawatts in the United States, which accounted for 6.7% of the installed generation capacity and 6.6% of all electricity generated in the United States [4]. In Alaska, hydroelectricity has been a cornerstone of powering communities and advancements since the early 1900s. Alaska’s energy portfolio is composed of about 25% hydropower with a capacity of about 483 MW across 50 power plants [5], and it accounts for 90% of the renewable energy generated in Alaska [6].

There is no history of pumped storage hydro methods, such as conventional hydropower, in Alaska; however, it has played a larger role in the contiguous United States since 1930. PSH accounts for 93% of utility-scale storage across 43 PSH plants. This amounts to a generation capacity of about 23 GW [7]. PSH was originally developed in the 1960s to assist with new nuclear generation with limited ramping capabilities being built at the time [8]. The US has not brought any new plants online since 2000. Most plants were brought online in the 1970s [9]. PSH is traditionally used to shift energy between day and night electricity rates; however, in scenarios with high renewables, particularly those with high wind generation, PSH could provide cost savings and reduce wind curtailment [10,11,12].

This paper explores the feasibility of an increased renewable energy capacity in Alaska, specifically in the regional transmission system called the Railbelt outlined in Section 2. Hydropower and PSH were researched in particular because of the large resources available in Alaska. Hydropower has also been a staple for power on the Railbelt and in more remote areas. The hydropower and pumped storage hydro (PSH) resource availability and resource costs were assessed for this region in this paper. Impoundment hydro facilities and three types of PSH are included: conventional, adjustable-speed, and ternary. Multiple characteristics of the resources were collected and calculated, including the operation behavior, location, capacity, and cost.

The purpose of this work is to gather information on these resources that can be used to help make informed decisions on these technologies’ usefulness in Alaska. The contribution of this paper is (1) to provide a review of hydropower and PSH technologies, including how they operate and resource locations in Alaska’s Railbelt, and (2) to provide updated cost estimates for identified PSH sites in Alaska’s Railbelt and outline these technologies’ advantages and disadvantages for the purpose of power system planning and decision making on Alaska’s Railbelt transmission system.

This paper is organized into the following sections: Section 2 details the study area used in this paper, Alaska’s Railbelt transmission system. Section 3 provides a review of the technical operation of hydropower and PSH. Section 4 details the methodologies used to perform the resource assessment and availability and cost assessments in the region. Section 5 provides the resulting hydropower and PSH site locations and costs. Section 6 provides a discussion of other considerations for the development of hydropower and PSH. Section 7 concludes the paper and details future work.

2. Study Area

Alaska’s Railbelt Transmission system, referred to as the Railbelt, is the largest electrical grid in Alaska, serving about 900 MW of the peak winter load from Homer, AK to Fairbanks, AK. The Alaska Railbelt transmission system supplies 75% of the state’s population with power. Figure 1 illustrates the approximate location and layout of the Railbelt in Alaska.

The Railbelt has four electric cooperatives: the Golden Valley Electric Association, the Matanuska Electric Association, the Chugach Electric Association, and the Homer Electric Association. These utilities are grouped into three main load regions: northern, central, and southern.

Three hydropower plants on the Railbelt account for 18% of the total generation [13]. These three power plants are Bradley Lake, Cooper Lake, and Eklutna Lake. Bradley Lake began its operation in 1991 and currently supplies the Railbelt with a generation capacity of 120 MW. Cooper Lake went online in 1960 and generates 16.7 MW. Eklutna Lake came online in 1955 and currently generates 30 MW [14].

3. Technology Review

Conventional hydropower generates power by converting the gravitational potential energy of raised water to kinetic energy, which flows through a turbine connected to a generator. Hydropower uses one “reservoir” that is usually situated above the powerhouse. The reservoir can change based on the type of hydropower plant. Three types of hydropower exist: impoundment, diversion, and run-of-river [15]. Impoundment is the most common of the three. A dam is created to impound the water and create a reservoir with a much higher head pressure. Diversion does not create a reservoir; it diverts a portion of the river or stream into a penstock or another channel that flows into the powerhouse. Run-of-river hydro creates a low-profile solution that does not require a large impoundment of water, as opposed to full impoundment dams.

Conventional hydropower uses several types of turbines as well as accessory systems to harness the kinetic energy of water and efficiently control it. Hydroelectric turbines are separated into two categories, impulse and reaction. These categories separate the turbines by characterizing the method of water delivery. Each one also has a unique combination of flow and head. Impulse turbines utilize the velocity of the water to discharge into a bucket to transfer the energy to the turbine. Pelton and cross-flow turbines are the most common types of impulse turbines.

Reaction turbines use the pressure generated by the water to turn the turbine. This water flows through a curved penstock called a scroll case. This case directs the water into the turbine. This water is discharged perpendicular to its angle of entry. Two common reaction turbines are Francis and Kaplan, Francis being one of the most common models [15].

Each type of turbine has its own method for controlling the flow of water. For the Pelton turbine, an adjustable “spear” controls the amount of water moving through the nozzle; there is also a deflector that will move in front of the nozzle to control the flow. For the Francis turbine, wicket gates are aligned around the edge of the turbine. The gates are rotated to control the flow, ranging from fully closed to fully open, and are located parallel to the water flow. These controls allow the hydropower plant to regulate its power output and respond to the load demand. Hydropower has the ability to ramp to full generation on a scale of 12–50 MW/s [16].

PSH can be described as a gravity-driven water battery. PSH installations involve two or more reservoirs of water with a substantial amount of elevation difference. Often, this can be conducted by utilizing lakes, existing reservoirs, or even oceans and seas. Another method consists of creating reservoirs using geographic features, such as large “bowl-shaped” features separated by elevation. Depending on the reservoir type, a PSH plant can be characterized as closed- or open-loop [7]. Closed-loop systems are systems that are not connected to any natural water source i.e., lakes, rivers, and oceans. Open-loop systems are systems that utilize a natural reservoir, for example, pumping between two lakes [17].

PSH moves volumes of water between the two reservoirs to follow the demand. When the power demand is high, water flows from the upper reservoir to the lower reservoir to generate power. When the demand is low, water is pumped from the lower reservoir to the upper reservoir using cheap or excess power. The water is transported, via penstocks, to and from the powerhouse at the lower reservoir using pump–turbines.

The configuration of the pump/turbine depends on the type of turbine used. The setups researched for this study included conventional (C-PSH), adjustable-speed (AS-PSH), ternary (T-PSH) [18], and quaternary (Q-PSH) [19]. Conventional and adjustable-speed units utilize a reversible pump–turbine, sometimes referred to as pump-as-turbines. Due to the similarity in the design of pumps and turbines, slight modifications can be made to enable the turbine to reverse its direction and pump the water, instead of allowing it to flow through. This also means that the generator, when reversed, acts as a motor. C-PSH most often uses a synchronous machine for its operation. Configurations using synchronous machines are restricted to operating at one rated speed for maximum efficiency, limiting their capability to “Full Off” or “Full On”. AS-PSH can be configured with a Doubly Fed Induction Machine, an asynchronous machine, or a synchronous machine, with the only item changing being the location and size of its power converter. The asynchronous machine is similar to a Type 3 Wind Turbine, while the synchronous machine is similar to a Type 4.

The asynchronous machine has a wider operating range with a high efficiency. Being able to change the speed at which the reversible pump-turbine generates or pumps enables a stronger ability to regulate the frequency or respond to the load. T-PSH can also respond effectively due to its unique setup. T-PSH utilizes a pump and turbine separately on a single shaft connected to a synchronous machine. The separate machines allow the system to pump and generate without reversing the rotation of the shaft. The pump and turbine are fed by the same penstock split in two. C-PSH and AS-PSH are limited to two modes, pump and turbine, allowing only one direction for water flow in the system. Ternary has three modes, pump, turbine, and hydraulic short circuit. This third mode is enabled due to the split penstock and single direction rotation of the shaft. A hydraulic short circuit operates the pump and the turbine at the same time while also supplying the pump with energy from the grid. Volumes of water being moved are segmented into blocks of power, as shown in the Figure 2. This short-circuit mode enables T-PSH to have a very competitive advantage over C-PSH and AS-PSH due to decreased mode switching times and increased frequency regulation capabilities.

Q-PSH has two shafts compared to the one shaft that is available in the other three types of PSH. The two shafts allow for a pump–motor combination and a turbine–generation combination; with a full sized converter between the pump and motor and a direct grid connection with the turbine–generator combination. This two-shaft structure, with the combinations of the pump–motor and turbine–generator, combines the benefits of the AS- and T-PSH and allows faster mode change times. Therefore, it provides the best frequency regulation of the types of PSH available, particularly with high penetration of renewable energy resources [19].

When selecting a PSH configuration, system demand and cost are the primary factors. Depending on the demand, the ability to transition between modes can also quickly become an important factor in the decision-making process. C-PSH is the most basic of the technologies. It also has the slowest switching time with averages in the range of minutes. AS-PSH is slightly faster, ranging from 1 to 4 min for a transition between pumping and generating. T-PSH has a transition time of 30 s to 1.5 min [20]. Q-PSH has the fastest switching times and provides the greatest frequency support [19].

4. Data and Research Methodology

4.1. Resource Assessment and Availability

For conventional hydropower, a database from the Alaska Energy Authority distributed through the State of Alaska Open Data Geoportal [21] was used to map proposed hydroelectric sites in Alaska. For PSH, datasets from the National Renewable Energy Laboratory (NREL) were used [22]. This dataset was created using a GIS-based analysis of potential closed-loop systems in the United States. The PSH sites were filtered in an assessment performed by NREL [22,23], and the number of sites was narrowed down to 1819 sites. The goal of the generated resource maps was to estimate the possible size and number of PSH and hydro resources potentially usable for the Railbelt. The following criteria were used to select sites:

• Cost ($/kW);
• Distance from Railbelt (miles);
• Capacity (MW);
• Land ownership;
• Previously researched potential sites.

To apply these criteria, publicly available infrastructure data for Alaska’s high-voltage transmission lines were imported from the Alaska Energy Authority’s database hosted on the State of Alaska Geoportal. This public transmission data were selected to exclusively show the Railbelt transmission lines. The potential resources were filtered based on their straight-line proximity to the transmission line; for this study, 20 miles was selected as the maximum distance from a Railbelt transmission line. However, due to Alaska’s topography, if a spur line cost was to be estimated, the straight-line distance would not be correct.

To estimate the cost of a spur line, a possible path for this line must be created. To avoid mountains and rapid elevation changes, it was assumed that the spur line would follow a path with the least amount of slope. To simulate this path in QGIS, a least cost path analysis was completed with the cost being the slope. A Digital Terrain Model of the area of interest surrounding the Railbelt was imported from the State of Alaska’s Geoportal. The Digital Terrain Model held elevation data about the area of interest’s terrain. This excluded trees and buildings that would be included in a Digital Surface Model. Using the slope tool in QGIS, a slope map was created. To keep the paths reasonable, all bodies of water were excluded from the slope map using the intrinsic slope of water, 0, and converting all 0 values to NULL values, leading them to be ignored by the least cost path tool. The least cost path tool assessed the values held in each pixel of the slope map. This assessment was used to draw a path from the resource site to the nearest point on the Railbelt whilst collecting the least amount of cost (slope).

The Railbelt-wide hydro and PSH sites with the least cost paths are shown in Figure 3. A close-up view of Figure 3 is shown in Figure 4 to illustrate this method.

Alaska’s land ownership was the next factor considered for resource location. Alaska has native, federal, state, and borough lands, protected wildlife, national parks, military bases, etc. To visualize the land ownership, a map of Administered Lands, courtesy of the Alaska division of the US Bureau of Land Management, was used [24]. This map specifies the owner of each parcel of land in Alaska, allowing potential resource sites to be removed if they are on non-buildable land, such as National Parks, military installations, and other protected areas. Locations that exist on public lands were selected as these are more likely to be used for potential development. The map of Administered Lands was also used for potential resource sites from the Alaska Energy Authority.

4.2. Resource Cost

Water-based energy resources are commonly expensive to construct due to their low energy density (dams, reservoirs, etc.) and environmental impact. The conventional hydropower initial capital cost (ICC) is dependent on several factors. Hydropower projects are often filtered based on their generation capacity, which contributes to other cost factors, such as the head and reservoir size. Capital costs of hydropower include electromechanical systems, C&I, environmental impact studies, permitting, and environmental impact mitigation procedures. A cost analysis of hydropower by the International Renewable Energy Agency identified the largest contributors to capital costs for large- and small-scale hydropower projects [25]. Large-scale hydropower projects have capital costs consisting mainly of C&I, such as the impoundment, tunneling, and powerhouse costs associated with moving large volumes of water. Capital costs of small-scale hydropower projects are dominated by electromechanical systems, such as generators, turbines, control systems, etc. Multiple models and studies have shown a general relationship between the generation capacity and cost. Head also has an effect on this relationship [25]. For most projects, the ICC ($/kW) will increase as the capacity decreases. For small hydro projects, 5–100 MW, the cost can range from $1300 to $8000/kW ($2010 USD). Large hydro projects, greater than 100 MW, have a slightly lower range of $1050–7650/kW [25]. The Oak Ridge National Laboratory (ORNL) completed their own analysis of hydropower costs and created a model to estimate new project costs [26]. The ORNL analyzed 84 New-Stream Development projects, meaning situations where no dam existed prior to the project. The ORNL assessed this database using project characteristics and a confidence score system to develop an ICC model equation, Equation (1), where P is the capacity in MW, and H is the head in feet. The ICC is measured in USD for 2012.

$$\mathrm{ICC}=8717830{\mathrm{P}}^{0.975}{\mathrm{H}}^{-0.12}$$

(1)

Using this model, the ORNL estimated that, on average, new-stream development projects cost 3882 $/kW. This cost was verified against 17 constructed projects with an average cost of 3885 $/kW [26].

The pumped storage hydropower capital cost is dependent on multiple factors. Those with the greatest effect include the head height, flow rate, and storage/generation capacity. The NREL modified the simplified PSH cost model from the Australian National University [27], which was adapted from a detailed model created in a report from Entura [28]. Equation (2) is the total cost model, and seven equations from [22] go into this equation.

$$C_T=(C_P+C_t+C_u+C_l)\frac{1.33}{1.2}+C_s$$

(2)

where $C_T$ is the total cost in 2018$, $C_P$ is the cost of the powerhouse in 2018$, $C_t$ is the cost of the tunnel in 2018$, $C_u$ is the cost of the upper reservoir in 2018$, $C_l$ is the cost of the lower reservoir in 2018$, the ratio 1.33/1.2 is the contingency cost adjustment factor used to adjust the contingency cost of 20% used in the ANU model to the contingency cost of 33% used for other NREL technologies, and $C_s$ is the cost of the spur line in 2018$. The NREL made two adjustments in Equation (2): $C_s$ was added, as the ANU’s model did not include spur line costs, and a cost correction factor of 1.51 was added to this simplified model for plants with a capacity of 900–1100 MW. This cost correction factor was arrived at by comparing the median cost of 900–1100 MW plants, created by the ANU’s cost model, to the average cost of a 1000 MW plant modeled as a result of the Energy Storage Grand Challenge. The final range of costs for PSH plants from the NREL [28] was 1407–2137 $/kW. The Energy Storage Grand Challenge created a comprehensive table outlining the costs and lifetime performance for a 100 MW, 1000 MWh system [16]. This table includes the powerhouse C&I, reservoir, contingency, O&M, electromechanical powertrain, and performance metrics based on the lifespan, efficiency, and ramp rates. It does not include O&M, which was modeled using a function of the plant capacity (P) in MW and the annual energy throughput ($AE$) in MWh, as shown in Equation (3).

$$C_{\mathrm{O}\&\mathrm{M}}=34730P^{0.32}A{E}^{0.33}$$

(3)

Alaska’s construction costs were calculated based on the cost assessments from the NREL [28]. Alaska’s construction costs included an adjustment for inflation from 2018 to 2023. An estimated increased cost of construction in Alaska was determined by comparing average new-stream development costs from the United States, 4800 $/kW (2015$) [26] adjusted to 6144 $/kW (2023$), to the most recently proposed hydro development in Alaska. The most recently proposed Alaskan hydro project is Grant Lake, which has a capacity of 5 MW. Grant Lake’s estimated construction costs total $53,878,050 (2018$), adjusted to $65,192,441 (2023$) [29], which equates to 13,038 $/kW. Therefore, the an Alaskan construction cost factor of 2.12 was used to calculate the increased cost of construction in Alaska.

5. Results

5.1. Pumped Storage Hydro

Resource locations for each type of technology were narrowed down from the locations identified by the Alaska Energy Authority for hydro and the NREL locations for PSH as well as the Eklutna and Right Mountain locations identified by the Alaska Institute for Climate and Energy, which are located on the Railbelt [30]. Cost was the ultimate deciding factor for the sites selected. The PSH site locations were narrowed down to the top ten lowest-cost sites by $/kW using the Alaska construction costs. The capital costs of the Eklutna and Right Mountain sites reported in the Alaska Institute for Climate and Energy report were increased to adjust to 2023$, and no additional Alaska construction cost multiplier was added; however, it should be noted that these costs were calculated differently and may not be suited for direct comparison to the other site costs. The details of each of the top ten lowest-cost sites within 3 miles of the Railbelt are listed in

Table 1. Top ten lowest-cost ($/kW) PSH sites within 3 miles of the Railbelt.

Site	Nearest Water Body	Power Capacity (MW)	Energy Storage (GWh)	Capital Cost Estimate (2023 $/kW)
1	Eklutna Lake	426	507	2328
2	Sustina River	744	7.44	2979
3	Lewis River	707	7.07	3400
4	Nellie Juan River	461	4.61	3540
5	Ship Creek	542	5.42	3645
6	Jakolof Creek	432	4.32	3753
7	Yanert Fork	502	5.02	3963
8	Pierce Creek	407	4.08	3997
9	Willow Creek	296	2.96	4146
10	Little Willow Creek	445	4.45	4176

, where the first site is the Eklutna PSH site.

5.2. Hydropower

Hydropower locations were identified based solely on previously studied sites as well as some upgrades to current hydropower sites. The largest site identified was the Susitna–Watana Hydro Project. This project has an annual energy budget of 2.8 million MWh with an installed generation capacity of 459 MW [14]. The Dixon Diversion was also included. This is a project to divert glacial meltwater from the Dixon Glacier, increasing the annual energy budget of Bradley hydro by 150 GWh. This additional capacity would require the Bradley Lake dam to be raised; however, the total nameplate capacity of the Bradley Lake power plant would stay the same. The Grant Lake hydro project on the Kenai Peninsula was also identified. This project is scheduled to begin construction in 2023 and has a 5 MW generation capacity and an annual energy production of 18,600 MWh. The selected hydro sites along with the top ten lowest-cost PSH sites are shown in Figure 5.

6. Other Considerations

Hydropower can also provide additional services to electric grids beyond energy. There are also negative consequences associated with both hydropower and PSH. These other considerations are further outlined here and classified as either a strength or weakness of hydropower or PSH.

6.1. Strengths

Hydropower generators are capable of providing frequency regulation and reserves, and their hourly ramping flexibility is the most extensively used out of any other type of resource in the United States [2]. This flexibility and frequency regulation and reserve capability are key grid services, especially when integrating variable renewable energy resources. Hydropower generators can also be black-start capable, meaning that they can be used to restart the electrical grid after a system-wide blackout [2]. Black-start capability is not found with all types of resources and is a critical component of grid resilience.

Hydropower also has non-energy-related benefits including food security through providing a source of water for irrigation, water quality and quantity through the control of water release, high-paying jobs, flood protection through capturing floodwaters, and recreation and tourism [31].

6.2. Weaknesses

Energy resources that use water as the generation medium can have ecological and environmental (flora, fauna, landscape, and historical remains [32]) impacts. Water must be diverted or stored to create power. This often creates reservoirs that can span hundreds of miles. In Alaska, there are several factors that come into consideration when deciding where to build a PSH or hydropower plant. For hydropower, much of Alaska’s most energy dense river systems are considered anadromous streams [33]. These streams are used by salmon for spawning, meaning that interference would impact salmon runs, which is a subsistence resource and a cornerstone of Alaska’s economy. Other issues include the obstruction of natural views, which is also important for tourism.

PSH can have a similar effect to conventional hydropower, depending on how it is configured. Closed-loop systems have the potential to impact habitats and species, depending on where the reservoirs are placed. This, for example, could be a concern for caribou habitats in much of Alaska. Open-loop systems pose the same threat as hydropower due to constantly changing water levels disabling the ability for the reservoirs to be used by wildlife.

Studies have identified that the water bodies of dams of hydropower facilities release CO₂ and CH₄ emissions due to both human-made and natural processes [34,35]. Therefore, hydropower generation is not a completely carbon-free resource. However, it has been found that hydropower generation facilities do have lower emissions in comparison to thermo-based generation methods, such as coal, oil, diesel, and natural gas generation facilities [35].

7. Conclusions

This paper covered how PSH and hydropower operate on a technical basis as well as the cost, and potential locations of these resources. PSH and hydropower generate and store power using the potential energy of elevated water sources. Hydropower uses conventional turbine generators for operation and can ramp up to full generation on a scale of 12–50 MW/s [16]. PSH is similar to hydropower; however, it utilizes two reservoirs of water separated by an elevation difference to create a large water battery. The water is pumped to the upper reservoir when the demand is low and generated down to the lower reservoir when the demand is high. PSH powerhouses use reversible pump–turbines to pump and generate between the two reservoirs. The transition time between pumping and generating changes depending on the PSH configuration: conventional, adjustable-speed, ternary, and quaternary. Quaternary has the fastest switching times and greatest frequency regulation capabilities.

PSH and hydropower have comparable ICC categories: C&I, electromechanical systems, environmental protection procedures, and permitting. Cost models for hydropower plants created by ORNL showed that new-stream development hydropower has an average ICC of around 3882 $/kW. PSH cost models from the NREL HydroWIRES report resulted in a ICC range of 1407–2137 $/kW. The construction costs accounting for inflation and Alaskan construction costs resulted in a range of 2090–2931 $/KW.

The potential locations of hydropower and PSH were analyzed from several sets of data. Ultimately, two hydropower sites were identified: the Susitna–Watana Hydro and Grant Lake. Ten potential PSH sites were also identified based on cost.