On the value of liquid-air and pumped-thermal electricity storage systems in low-carbon electricity systems

Journal Pre-proof On the value of liquid-air and Pumped-Thermal Electricity Storage systems in lowcarbon electricity systems S. Georgiou, M. Aunedi, G. Strbac, C.N. Markides PII:

S0360-5442(19)32375-8

DOI:

https://doi.org/10.1016/j.energy.2019.116680

Reference:

EGY 116680

To appear in:

Energy

Received Date: 14 April 2019 Revised Date:

28 November 2019

Accepted Date: 1 December 2019

Please cite this article as: Georgiou S, Aunedi M, Strbac G, Markides CN, On the value of liquid-air and Pumped-Thermal Electricity Storage systems in low-carbon electricity systems, Energy (2020), doi: https://doi.org/10.1016/j.energy.2019.116680. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

On the Value of Liquid-Air and Pumped-Thermal Electricity Storage Systems in Low-Carbon Electricity Systems S. Georgiou1, M. Aunedi2, G. Strbac2 and C. N. Markides1,* 1

Clean Energy Processes (CEP) Laboratory, Department of Chemical Engineering, Imperial College London, London, UK

2

Department of Electrical and Electronic Engineering, Imperial College London, UK *

Corresponding author: [email protected]

Abstract We consider two medium-to-large scale thermomechanical electricity storage technologies currently under development, namely ‘Liquid-Air Energy Storage’ (LAES) and ‘PumpedThermal Electricity Storage’ (PTES). Consistent thermodynamic models and costing methods based on a unified methodology for the two systems from previous work are presented and used with the objective of integrating the characteristics of the technologies into a wholeelectricity system assessment model and assessing their system-level value in various scenarios for system decarbonization. It is found that the value of storage depends on the cumulative installed capacity of storage in the system, with storage technologies providing greater marginal benefits at low penetrations. The system value of PTES was found to be slightly higher than that of LAES, driven by a higher storage duration and efficiency, although these results must be seen in light of the uncertainty in the (as yet, not demonstrated) performance of key PTES components, namely the reciprocating-piston compressors and expanders. At the same time, PTES was also found to have a higher power capital cost. The results indicate that the complexity of the decarbonization challenge makes it difficult to identify clearly a ‘best’ technology and suggest that the uptake of either technology can provide significant systemlevel benefits.

Keywords: energy storage, liquid-air energy storage, pumped-thermal electricity storage, power system economics, whole-system assessment.

1

1 Introduction The competitiveness of any energy storage technology is strongly affected by its technical and economic characteristics. A storage system that is both efficient and economically competitive is widely acknowledged as an important component of a flexible and efficient low-carbon electricity system. It has the potential to support the cost-efficient integration of intermittent renewable generation, can take advantage of any differences between peak and off-peak electricity prices, as well as provide local and national services to network and system operators. The application potential of any technology in a power system will depend on its characteristics in combination with the requirements of the whole system. Different scales and types of electricity storage technologies are currently being explored. Examples of these include pumped hydroelectricity storage, over-ground compressed air electricity storage, large-scale underground compressed air electricity storage, lithium ion batteries, sodium sulphur batteries, and lead acid batteries. In general, these technologies have been extensively studied and reviewed in literature. For example, Ref. [1] presented an overview of a wide variety of electricity storage solutions based on the technologies available at the time while also exploring their different characteristics, both technical and economic. Similarly, the guide presented by Ref. [2] focused on some of the most common energy storage technologies including pumped hydro, compressed air energy storage, along with a variety of battery based technologies (e.g., lead-acid, lithium-ion, sodium-nickel-chloride, sodiumsulphur, iron-chromium, vanadium redox). The study also presented the main characteristics of each technology as well as their costs at different scales and settings. Furthermore, Ref. [3] considered several energy storage technologies covering their applications, characteristics and their adoption levels. From these studies, it can be seen that different storage solutions can be best positioned for different operations based on their characteristics and abilities. Ref. [4] also studied a wide range of electrical energy storage technologies while identifying the main obstacles for the deployment of energy storage systems. These included regulations and existing processes, costs, and low awareness of the benefits associated with storage [5]. In this paper, we focus on two relatively recently proposed medium-to-large scale thermomechanical electricity storage technologies: Liquid-Air Energy Storage’ (LAES) and ‘Pumped-Thermal Electricity Storage’ (PTES). Some of advantages of these technologies include their application at larger scales (both in terms of energy and power) and longer storage durations, their long expected lifetimes and the absence of geographical constrains

2

usually associated with other more established large-scale electricity storage technologies, such as pumped hydro storage and underground compressed air storage. Even though these technologies are relatively new, they have been studied in literature under several configurations. For example, Refs. [6,7] studied PTES by focusing on the thermodynamic and technical optimization of the system. The PTES configuration investigated in those studies is based on forward and reverser Joule-Brayton cycles, the same configuration considered here. Other studies focusing on PTES but using different configurations are also available in literature (e.g., see Refs. [8,9]). Similarly, studies focusing on LAES are also present in literature. For example, the use of packed bed cold storage in LAES was studied in Ref. [10], while Ref. [11] considered the impact of waste heat and cold in the enhancement of the system’s roundtrip efficiency. It is interesting to note that LAES has had an operational pilot plant for several years. The operation and performance of the pilot plant was studied in Refs. [12,13]. Therefore, LAES can be considered to be at a higher Technology Readiness Level (TRL) than PTES. A comparison between the two systems at the system level from the thermo-economic perspective was presented in Ref. [11]. In that study, the two systems were analysed based on their thermodynamic performance as well as their economic competitiveness. In addition, their expected operation capacities and cost comparison against other technologies was presented. However, Ref. [11] did not consider the application of such technologies as part of a wider low-carbon-electricity system and their potential value when operating within such a system. Therefore, it is considered of interest here to study this aspect further, in order to understand better the potential value of energy storage and the factors affecting this potential. Electricity storage systems can provide flexibility required for cost-effective integration of variable renewables into future power systems [14]. Energy storage technologies have different characteristics, such as power capacity, energy capacity, charge and discharge durations, and can therefore have different purposes in the electricity grid. Although it is important to determine and analyse both their technical and economic properties, it is also vital to assess their realistic value in an electricity system. This assessment can be challenging for newly proposed technologies with limited data, but on the other hand it can provide a first indication of the attractiveness of such technologies at a system level. As assessment framework for quantifying the whole-system value of energy storage in lowcarbon power systems in provided in Ref. [15], including the benefits that arise for

3

generation, transmission and distribution investment as well as system operation. Operational flexibility provided by energy storage and other technologies has been shown to help deliver carbon and renewable targets at a lower cost [16,17], as well as support the integration of variable renewable sources [18,19]. Complex benefits of energy storage result from its ability to not only participate in energy arbitrage, but also at the same time provide reserve services [20], frequency response [21], as well as other balancing and network services [22]. This paper focuses on the potential deployment of LAES and PTES in low-carbon electricity systems. This is performed by incorporating technical characteristics of these technologies, obtained from existing thermo-economic models, into an electricity system model. Furthermore, costing estimates, also obtained from existing thermo-economic models, are compared against each technology’s system value of storage determined from the wholesystem model. Therefore, the main aim of this paper is to quantify the variations in the whole-system value of the two storage technologies based on their characteristics, across several scenarios of transitioning to a low-carbon electricity system.

2 Methodology A whole-system assessment approach is adopted here to determine the whole-system value of energy storage in low-carbon electricity systems. The whole-system model, WeSIM, determines optimal decisions for investing into generation, network and/or storage capacity, in order to satisfy the real-time supply-demand balance in a least-cost sense, while at the same time ensuring security of supply. A full description of the modelling approach is given in Ref. [15]. An application of WeSIM was presented in Ref. [22] that quantified the value of energy storage in supporting cost-efficient decarbonization of the electricity system, i.e., delivering carbon reductions at lower total costs. A similar approach was used in Ref. [23] to assess the role and value of pumped hydro-electricity storage in the European power system.

2.1 Electricity storage technologies The LAES system involves four main processes: i) air liquefaction; ii) liquid air storage; iii) waste cold storage; and iv) power generation. A version of the LAES technology is being developed by Highview Power Storage [24]. The configuration considered here is presented in the simplified schematic in Figure 1a. During charging, the system uses inexpensive electricity to liquefy air and gets charged by storing energy in the form of liquid air, and during peak demand periods, when it becomes economically attractive to discharge, the

4

system uses liquid air in the power generation unit where this is pressurized, evaporated, superheated to the temperature of the utilized waste-heat stream (if available) and expanded through a turbine to generate electricity [25]. Although optional, an additional process which was found to be a contributing factor for the enhancement of the system’s performance is the storage of waste cold from the power generation unit during the discharging periods and the utilization of that cold in the air liquefaction process during the charging periods. More details on the workings of a LAES system are given in Refs. [11-13,24,25].

(a)

(b)

Figure 1. Schematics of: (a) a LAES system (adapted from Ref. [12]); and (b) a PTES system (adapted from Ref. [6]). PTES is also a newly proposed electricity storage system, but at a lower TRL than LAES due to the lack of an operational pilot plant (although one is currently under commissioning and early testing in the UK). PTES stores energy in the form of sensible heat in insulated storage tanks containing a storage medium [6]. The configuration of the PTES system considered here operates based on a reverse/forward Joule-Brayton cycle for charging/discharging, respectively [6,7]. The system consists of two thermal reservoirs at different temperatures and pressures

when

charged

(HS:

hot

storage,

CS:

cold

storage),

two

reversible

expansion/compression devices (C: compressor, E: expander) [26,27] and two heat exchangers (HE: heat exchanger) (see Figure 1b). In the charging mode heat is extracted from the cold reservoir and pumped into the hot reservoir, thus resulting in increased temperature difference. During discharging, the flow of the working fluid is reversed to take advantage of the temperature difference, and a heat engine is used to generate electricity. The thermodynamic (first law) performance of energy storage systems is typically expressed in terms of a roundtrip efficiency (
), defined as the net work output ( ) during discharge 5

divided by the net work input ( ) during charge (Eq. 1):
=

 

(1)

where  and  for the two systems can be estimated using the charge and discharge thermodynamic cycles associated with each system. For LAES,  is the power generated during the discharge cycle, whereas  is the work input into the liquefaction unit during the charge cycle (see Figure 1a). Several operational and loss parameters can have an impact on the estimation of
of LAES, for example the amount of waste cold and heat utilization as well as the components’ efficiencies. For PTES,  and  are estimated by considering forward and reverse Joule-Brayton cycles for discharging and charging, respectively. Similarly to LAES, the performance is also significantly affected by a number of operational and component performance variables. The main losses in such a system arise from pressure losses, compression and expansion losses, and thermal losses in reservoirs [6,7,26-29]. Both LAES and PTES are relatively new technologies and consequently information on their costs is limited in the available literature. Therefore, and in order to obtain consistent estimates of the capital costs of both systems, a costing methodology was developed based on simple thermo-economic models and a costing exercise was performed [11,30]. The overriding aim of this exercise was to perform a preliminary economic feasibility assessment of the two technologies that would allow their assessment from a whole-system perspective. In the costing model used specifically for the estimation of capital expenditure, the systems were broken down into their fundamental components for costing and then summed to obtain an estimation of the overall system costs. Where possible, installation costs were considered. The model uses various methods for costing the different components. In addition, various methods from different sources were used to cost some of the components used in the two systems. This was performed in order to provide more overall confident estimate as well as to obtain a range of uncertainty in the costing of these components. In summary, for expanders/turbines, compressors, pumps and storage vessels the costing correlations based on Refs. [31-33] are used along with their associated factors and parameters to estimate capital costs. Each of these methods has cost correlations for these components as well as, in some cases, multiplication factors to account for aspects such as operating conditions. As a result, a range of capital costs is obtained, and the extent of this range in each cases can also be considered an indicator of the uncertainty in the costing results.

6

For the heat exchangers, the model obtains approximate capital costs based on the C-value method [34], which allows for simple costing without the necessity of calculating the heat exchanger area requirement. For the storage material in the storage vessels a specific cost of £100/t is used in the model, assuming magnetite as storage material [28,29]. Finally, for the cost estimation of generators, the model uses a capacity exponent factor approach given that alternative correlations such as the ones presented in Ref. [31] were not found in literature. This approach is based on the following relation in which the exponent factor () used is 0.94 [35]:  = 1.85 · 10 



. · 



(2)

where  is the capital cost in € and  is the power output capacity of the generator in kW. Due to the use of multiple methods to cost the various system components, a range in the total in total capital cost is obtained. From the different cost estimates, upper and lower bounds in addition to average values are found and reported. The average value is used as the primary costs whereas the maximum and minimum are used as estimates of cost uncertainty, as mentioned above. Finally, as the cost correlations used have a cost basis of a past calendar year, they are adjusted to a common money base year using CEPCI index [36]. The reader can refer to more details on the costing approach found in Ref. [11].

2.2 Whole-systems assessment model Analysing future electricity systems at sufficient temporal and spatial granularity is essential for adequately assessing the cost-effectiveness of decarbonization pathways and enabling technologies. In order to accurately quantify system operation and investment cost as well as its carbon performance, quantitative models need to simultaneously consider second-bysecond supply-demand balancing issues as well as multi-year investment decisions. Furthermore, it is also critical to adequately consider the synergies and conflicts between local and national (or trans-national) level infrastructure requirements. To this end, the Whole-electricity System Investment Model (WeSIM) described in Ref. [15] is used here to determine the value of energy storage technologies in supporting efficient investment and operation of future electricity systems. The model minimizes total system cost, comprising: •

Investment cost, which includes the (annualized) capital cost of new generation and storage capacity, the capital cost of new interconnection capacity, and the reinforcement cost of transmission and distribution networks. The investment costs

7

are annualized using the appropriate Weighted-Average Cost of Capital (WACC) and the estimated economic life of the asset. •

System operating cost, which consists of the annual generation operating cost including: (i) variable cost, which is a function of electricity output; (ii) no-load cost; and (iii) start-up cost. Generation operating cost is driven by the assumptions on fuel prices and carbon prices (for those technologies that are carbon emitters).

A key feature of WeSIM is that it determines the optimal technology mix for an electricity system while endogenously incorporating a variety of constraints, such as carbon emission constraints, security of supply criteria or scheduling adequate levels of fast and slow reserve services. A detailed model formulation is included in Ref. [15]. Only the key features and constraints are therefore presented here: •

Power balance constraints ensure that supply and demand are balanced at all times.

•

Operating reserve constraints, which include various forms of fast and slow reserve constraints. The amount of operating reserve requirement is calculated as a function of uncertainty in generation and demand across various time horizons. The model distinguishes between two key types of ancillary services: (i) frequency regulation (response), which is delivered in the timeframe of few seconds, considering the impact of system inertia; and (ii) reserves, split between spinning and standing reserve, with delivery occurring within the timeframe of tens of minutes to several hours. Additional constraints ensure that frequency response requirements meet the rate of change of frequency (RoCoF) specification, the minimum nadir frequency, and the steady-state frequency deviation from the nominal frequency.

•

Generator operating constraints, including: (i) Minimum Stable Generation (MSG) and maximum output constraints; (ii) ramp-up and ramp-down constraints; (iii) minimum up and down time constraints; and (iv) available frequency response and reserve constraints.

•

Generation investment: WeSIM optimizes the investment in new generation capacity while considering the generators’ operation costs and CO2 emission constraints, and maintaining the required levels of security of supply.

•

Annual load factor constraints are used to limit the utilization level of thermal generating units, e.g., to account for the effect of planned annual maintenance on plant utilization.

•

For wind, solar, marine, and hydro run-of-river generators, the maximum electricity production is limited by the available energy profile. The model will normally maximize the utilization of these units given their very low marginal cost; nevertheless, 8

in certain conditions when there is oversupply of electricity in the system it may become necessary to curtail renewable output in order to balance the system. •

Demand-side response constraints include constraints for various types of flexible loads, including: (i) weather-independent demand, such as lighting and industrial demand; (ii) heat-driven electricity demand; (iii) demand for charging electric vehicles; and (iv) smart appliance demand. Different demand categories are associated with different levels of flexibility, obtained from the authors’ previous detailed bottom-up modelling of different types of flexible demand.

•

Power flow constraints limit the energy flowing through the lines between the areas in the system, respecting the installed capacity of the network. The model can also invest in increasing network capacity if this is cost efficient.

•

Distribution network constraints are devised to determine the level of distribution network reinforcement cost driven by net peak demand, as informed by detailed modelling of statistically representative UK networks. WeSIM can model different types of distribution networks, e.g., urban, rural, etc. with their respective reinforcement cost.

•

Emission constraints limit the amount of annual carbon emissions from the electricity system, and may affect the choice of generation technologies to invest in.

•

Security constraints ensure that there is sufficient generating and storage capacity in the system to supply the demand with a desired level of security.

2.3 Assessing the value of energy storage in future electricity systems A gross value approach is adopted in this paper to assess the system benefits of energy storage. In the first step, this approach consists of minimizing the total system cost for an appropriately constructed counterfactual scenario, in which there is no energy storage. In the second step, a series of model runs is performed with gradually increasing energy storage capacity, and the resulting reduction in total system cost is interpreted as whole-system benefit of energy storage. Scenarios with energy storage do not assume any cost of storage, hence providing gross (rather than net) system benefits. Gross system benefits can be a useful benchmark to compare against the projected cost of a given energy storage technology. The gross system value approach is adopted in this paper due to the significant uncertainty associated with predicting the investment costs of the storage technologies considered in this work (see Section 3.1). The gross system value focuses on the cost savings that a given storage technology can generate in the power system, and one of the objectives of the paper is

9

to contrast this system value with the estimated cost ranges for LAES and PTES technologies, in order to indicate cost-efficient deployment levels for a given cost point. In this paper the gross whole-system value of storage is quantified in two ways: •

Average whole-system value, obtained by establishing the cost reduction between a given energy storage scenario and the corresponding counterfactual scenario, and then dividing cost savings with the total assumed capacity of energy storage (in kW or kWh). For instance, if the scenario with 10 GW of energy storage results in total system cost savings of £1bn per year, the average gross system value or energy storage is £100/kW per year.

•

Marginal whole-system value, obtained by establishing the cost reduction between a given energy storage scenario and the previous scenario with lower storage capacity, and then dividing it with the incremental capacity of energy storage. For example, if in the scenario with 10 GW of energy storage the total system cost savings are £1bn per year, and the one with 5 GW of storage resulted in £0.6bn of annual savings, the marginal gross system value or energy storage is £0.4bn divided by 5 GW, or £80/kW per year.

More formally, the average and marginal whole-system values of the two energy storage & technologies across different scenarios were found as follows. Let Ξ",$,% denote total system

cost quantified using the whole-system model for region ' (North or South), carbon target ( (100 or 50 g/kWh) and storage scenario ) ∈ +0,1, … ,6. corresponding to deployment levels of /% = 0, 5, …, 25 GW for technology 0. For a given region ' and carbon target ( the average system value of storage technology 0 in storage scenario ) (where ) ≥ 1) is then quantified as: & 3",$,% =

485,6,7 9485,6,: ;:8 9;78

(3)

Similarly, the marginal value of storage technology 0 in each scenario ) ≥ 1 is found as: & <",$,% =

485,6,:=> 9485,6,: 8 ;:8 9;:=>

(4)

The marginal value of storage is a particularly suitable indicator for comparison with estimated costs of storage technologies. It decreases with the installed capacity of storage, as the benefits of first MWs added will be higher than those of subsequently added storage capacity due to diminishing returns and reduced cost savings opportunities. The marginal value provides an indication of the cost-efficient level of deployment, given the basic economic principle that energy storage (or any other technology for that matter) should be deployed up to the level where its gross marginal value equals its cost.

10

2.4 Description of the scenarios used in the analysis Scenarios used to assess the system value of energy storage technologies in this paper are constructed to capture the key drivers for the value of flexibility provided by energy storage. In all scenarios the power system is designed and operated to meet one of the two levels of carbon emission intensity: 100 g/kWh or 50 g/kWh. These carbon targets broadly correspond to the targets for the UK power system in the 2030-2040 horizon. All scenarios are constructed by optimizing the portfolio of generation technologies to meet the carbon target, while meeting electricity demand with adequate level of security of supply. Technologies available for adding to the system included: wind, solar PV, nuclear and Carbon Capture and Storage (CCS), as well as conventional generation technologies such as Combined-Cycle Gas Turbines (CCGT) and peaking gas generation such as Open-Cycle Gas Turbines (OCGT). In order to represent typical variations in renewable output and demand across different geographies, the scenarios were developed to broadly represent the North and South of Europe. The annual utilization factor for wind in the North was 39%, which was higher than in the South (33%), while the utilization of PV in the North was lower than in the South – 11% vs. 18%. To represent geographical variations in climate, the peak demand was assumed to occur during winter in the North and during summer in the South. It is assumed that the electricity system can be represented by a single node without power exchanges with neighbouring systems. The system electricity demand was sized to broadly correspond to the GB system demand at 347 TWh annually, of which 8.4% and 7.8% was associated with electrified transport and heat demand, respectively. Peak electricity demand in the North was assumed to occur in winter, at the level of 71.8 GW, while in the South the peak demand of 67.0 GW occurred in summer (note that these represents the peaks before accounting for any demand-side response actions). The central assumption in all scenarios was that the uptake of demand-side response (DSR) was 25% of its theoretical potential, allowing a proportion of demand to be shifted in time. DSR in the model is provided by flexible electric vehicles, heat pumps, residential appliances and industrial and commercial demand. Sensitivity studies were also carried out for DSR uptake levels of 0% and 50% to study the competition between DSR and energy storage. The assumed DSR penetration refers to the uptake level of DSR relative to its maximum theoretical potential. The whole-system model employed in this paper, which is presented in detail in Ref. [15], distinguishes between flexible electric vehicles, heat pumps, residential

11

appliances, and industrial and commercial demand. For each of these categories, a different proportion of demand was assumed to allow temporal shifting, based on previous bottom-up modelling of demand flexibility [37,38]. For instance, it was assumed that at full (i.e., 100%) DSR penetration, 80% of the EV demand can be shifted in time, while only 10% of the I&C demand can be shifted within each 24-hour period. For a given DSR uptake level (e.g., 25%), this assumption was scaled down accordingly, so that, for example, only 20% of the EV demand and 2.5% of the I&C demand was considered shiftable within each 24-hour period. The counterfactual scenarios were assumed not to contain any (pre-existing) energy storage. The deployed capacity of each of the two energy storage technologies studied in this paper, LAES and PTES, was varied between 0 and 25 GW in 5 GW increments. In line with Ref. [11] and the economic analysis of LAES and PTES technologies presented in Section 3.1, the durations for LAES and PTES (i.e., the ratios between their energy capacity and installed power) were taken to be 4.0 hours and 5.8 hours, respectively, while their assumed cycle efficiencies were 55% and 70%, respectively (see Table 1 for more details on technical and economic parameters of LAES and PTES). Given the expected scale of deployment of LAES and PTES in the MW range, it was assumed that both technologies were connected to the high-voltage electricity distribution grid, so that their impact on distribution network cost is adequately evaluated in whole-system modelling. The costs of the generation technologies were assumed largely based on Ref. [39] as well as the authors’ own projections. The costs of renewables from Ref. [39] were converted into GBP using the exchange rate of £1 = $1.375, resulting in a levelized cost of electricity (LCOE) for wind of £41/MWh in the North of Europe and £48/MWh in the South, and for solar PV of £69/MWh in the North and £42/MWh in the South. The LCOEs of nuclear and CCS (assuming 90% annual load factor) were £134/MWh and £93/MWh, respectively, based on the estimates reported in Ref. [39] and correcting for the duration of construction of these technologies. Investment costs for CCGT and OCGT generators were assumed at £687/kW and £568/kW, respectively, as in Ref. [39]. Based on the authors’ own estimates, the assumed conversion efficiencies at full output for CCGT, CCS and OCGT were 54.7%, 49.3% and 39.0%, respectively, but the model also took into account the reduction in efficiency when these thermal generators operate part-loaded. The assumed efficiencies resulted in carbon emission factors for these technologies at full output of 367, 41 and 515 gCO2/kWh, respectively (CCS was assumed to have the carbon capture rate of 90%). Fuel and carbon costs were also taken from [39], with the assumed cost of gas of £22.6/kWh, and the carbon price of £29.1/t. 12

3 Results and discussion 3.1 Electricity storage technology costs In the case studies considered in this paper, the unit sizes of the LAES and PTES systems were taken from values proposed by companies developing these technologies: for LAES, a power output capacity of 12 MW and energy capacity of 50 MWh were used (taken from Ref. [11]), and for PTES a power output capacity of 2 MW and energy capacity of 11.5 MWh were used (taken from Ref. [11]). These capacities and power-to-energy ratios are similar to the ones found in literature (e.g., Refs. [6,7,25]) and can be potentially considered viable for early stage commercial-scale plants. A summary of technology characteristics assumed in the present analysis is provided in Table 1. Thermodynamic and costing models of LAES and PTES were used to estimate their power capital cost (total capital expenditure divided by the power capacity), and energy capital cost (total capital expenditure divided by the energy capacity). The estimated power capital cost of PTES and LAES was found to be around £3,185/kW and £1,425/kW based on mean values, respectively. The equivalent values in terms of the energy capital costs for PTES and LAES were estimated at about £540/kWh and £340/kWh, respectively. The difference between the two technologies was found to be significantly less in terms of energy capital cost than power capital cost. The ranges of power capital cost were estimated to be 930-1800 £/kW and 2200-4525 £/kW for LAES and PTES respectively. Similarly, in terms of energy capital cost, the two systems were estimated to be in the range of 225-435 £/kWh for LAES and 375-775 £/kWh for PTES. The source of these cost variations is the different costing methods used. Some variation in cost estimates is usually expected due to the large number of factors affecting the costing of components and therefore the overall system cost. These values were obtained in the present work and correspond to the capacities and efficiencies of the systems under consideration. Table 1. Characteristics of unit LAES and PTES systems considered in this work. Characteristic

LAES

PTES

Power capacity

12 MW

2 MW

Energy capacity

50 MWh

11.5 MWh

Duration

4.0 hours

5.8 hours

55%

70%

Power capital cost

930-1800 £/kW

2200-4525 £/kW

Energy capital cost

225-435 £/kWh

375-775 £/kWh

Cycle efficiency

13

Notes As in Ref. [11]

It is recognized that at different capacities and power to energy ratios the power and/or capital costs might change and this represents an area for future work. For the cases considered here, these cost estimations indicate a slight competitive advantage of LAES in terms of both power and energy capital cost (Figure 2 and 3). However, the capital cost estimates do not reflect the system value of each technology in a given electricity system. Therefore, a combination of both cost and value estimates is adopted to assess the attractiveness of these systems in low-carbon electricity systems.

PTES

LAES

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Power capital cost (£/kW) Lower bound

Mean

Upper bound

Figure 2. Estimated power capital cost of PTES and LAES. N.B.: For systems of a different size, in each case as proposed in literature and by the technology developers (see Table 1).

PTES

LAES

0

100

200

300

400

500

600

700

800

900

Energy capital cost (£/kWh) Lower bound

Mean

Upper bound

Figure 3. Estimated energy capital cost of PTES and LAES. N.B.: For systems of a different size, in each case as proposed in literature (see Table 1).

3.2 System value of energy storage technologies Generation portfolios in counterfactual scenarios, where the generation portfolios were costoptimized to meet carbon intensity targets without any energy storage, are shown in Figure 4 for North and South of Europe scenarios and 100 and 50 g/kWh carbon targets. In the North the carbon target is achieved mostly by installing around 80 GW of wind generation, and CCS capacity (more so at 50 g/kWh). The remainder of the portfolio consists of CCGT and

14

OCGT generation to ensure sufficient capacity margin. South scenarios contain a mix of wind and PV capacity, as well as CCS capacity that is higher than in comparable North scenarios.

Figure 4. Generation technology mix in counterfactual scenarios (without energy storage). System values of LAES and PTES were then quantified as described earlier. Given that the system optimization provided annual cost savings, these annualized values were converted to capitalized values assuming the same system value would be generated over the lifetime of the storage asset. The assumed lifetime for both LAES and PTES was 20 years and the cost of capital was 7%. Figure 5 shows the capitalized average and marginal values of the two energy storage technologies expressed in £/kW across a range of scenarios and uptake levels.

Figure 5. Average and marginal system value of proposed LAES and PTES per unit power. Several key observations can be made from Figure 5. First, the system value of storage decreases significantly with higher uptake levels, as expected, and the marginal value decreases faster than the average value. Due to the saturation effect, an incremental unit of energy storage capacity brings less value if it is added into the system with an already high

15

energy storage capacity. The rate of decrease in marginal value is similar for both technologies and both carbon target scenarios. On the other hand, the value reduces more slowly in a more solar-dominated South system. As the deployed storage capacity increases from 5 to 25 GW, the marginal value in the North reduces by 55% for LAES and 58-60% for PTES, while in the South the reductions are 42-43% for LAES and 43% for PTES. This result can be explained by a higher variability of predominantly solar PV generation in the South and hence more opportunity for using energy storage. Overall, the marginal system value of LAES was found to vary across the two scenarios and two geographies between £1,8002,100/kW for 5 GW uptake and £800-1,200/kW for 25 GW uptake. Higher values are observed for PTES, with £2,000-2,500/kW at 5 GW penetration and £800-1,400/kW at 25 GW penetration. Second, the value of PTES appears to be higher than for LAES for comparable scenarios and deployment levels. This difference can be attributed to the positive effect of higher duration and higher efficiency of PTES when compared to LAES. In the North, the value advantage of PTES is 13-17% at low penetrations (5 GW), but only 4-5% at higher penetrations (25 GW). In the South the value differentials are even more prominent: 17-20% at 5 GW and 14-21% at 25 GW. Third, geographical drivers for electricity generation and demand also affect the value of energy storage: results suggest that the system value is considerably higher in the South than in the North, primarily driven by higher variability of PV generation compared to wind. The values of LAES and PTES are 11-14%, i.e., 14-18% higher, respectively, in the South than in the North at 5 GW uptake; however at the uptake of 25 GW the value difference increases so that the value in the South exceeds that in the North by 40-46% for LAES and by 61-62% for PTES. Finally, the carbon ambition level is also a key driver for the whole-system value of storage, with values observed in the 50 g/kWh scenarios consistently higher than those in the 100 g/kWh scenarios, although the difference is not very significant. The differences may become more pronounced with even lower carbon targets, but further research is needed to quantify the exact magnitude of the impact of very low carbon intensity on the system value of storage. For some purposes it may be more convenient to express the system value of energy storage technologies per unit energy capacity, i.e., in monetary units per kWh of capacity. This can be evaluated by dividing the system benefits in terms of specific cost (£/kW) with the duration of a given storage technology, as illustrated in Figure 6.

16

Figure 6. Average and marginal system value of proposed LAES and PTES per unit energy. As expected, the relative trends of the system value of storage decreasing with higher penetrations and being the highest in the South and for 50 g/kWh carbon target remain unchanged. However, given the different assumed durations of the two technologies (4.0 hours for LAES and 5.8 hours for PTES) as suggested in Ref. [11], the value of PTES per unit energy now becomes smaller than the value of LAES, in contrast to the values expressed in £/kW (Figure 5). This implies that for PTES to become a more competitive option, its perunit energy cost will need to be lower than for LAES. The value of LAES in the North system was observed to vary between around £440-470/kWh at 5 GW penetration and £200-210/kWh at 25 GW penetration, while its value in the South for the same penetrations varied between £500-520/kWh and £290-300/kWh. The corresponding values of PTES are lower due to the lower duration parameter, reducing from £350-380/kWh to £140-150/kWh in the North, and from £410-440/kWh to £230-250/kWh in the South. The whole-system modelling approach adopted here also allows us to specify the breakdown of the marginal values of LAES and PTES into their constituent components, namely: the investment cost (Gen. CAPEX (low-C)) and operating cost (OPEX (low-C)) of low-carbon generation, Gen. CAPEX (other) and OPEX (other) of conventional generation and Distr. CAPEX of distribution networks. Such breakdowns are illustrated in Figure 7 for the 50 g/kWh scenarios, demonstrating that the value of storage can materialize in different segments of the electricity system, and that while the overall cost is reduced some cost components can increase while the others decrease. The net system value is the sum of savings across the different OPEX and CAPEX components.

17

Figure 7. Breakdown of marginal system value of proposed LAES and PTES (50 g/kWh). The following key system value components arising from energy storage can be identified: i) avoided Gen. CAPEX (low-C) and OPEX (low-C) of low-carbon generation (largely CCS), resulting from higher operational efficiency and lower renewable curtailment; ii) avoided Gen. CAPEX (other) of conventional power-generation, by storage displacing conventional generation and contributing to the capacity margin; and iii) for lower levels of storage penetration in the North there is avoided Distr. CAPEX of distribution driven by energy storage reducing peaks in the distribution grid. In most cases, it is found that deploying energy storage leads to a slight increase in the OPEX (low-C) of conventional generation given that a part of CCS generation is replaced by less expensive but more carbon-intensive CCGT, but with a positive overall cost savings. It emerges that the value of storage can materialize in different segments of the electricity system. In reality this would mean that to maximize its economic value, an energy storage operator would need to simultaneously deliver multiple services to the system [22]. The results confirm that one of the key drivers of the value of energy storage is its ability to deliver a lower-cost portfolio of low-carbon generation, in this case by replacing costly CCS capacity with less expensive (although also less flexible) wind and PV capacity. As an example, adding 25 GW of LAES or PTES in the North system allows the carbon target of 50 g/kWh to be reached with 6 GW less CCS capacity, and about 8 GW more wind capacity; in the South the same amount of storage displaces 9-11 GW of CCS and 2-5 GW of wind capacity but increases PV capacity by 31-38 GW (note that the disproportionately higher increase in PV capacity is a consequence of a lower capacity factor than CCS and wind).

18

To further illustrate the effect of adding energy storage on the composition of the cost-optimal generation mix, Figure 8 quantifies the changes in the generation mix compared to the counterfactual scenario for the 50 g/kWh scenarios. Of interest is that an increasing energy storage capacity generally results in progressively more CCS capacity being replaced with wind (in the North) or PV (in the South). At the same time, given that energy storage can contribute to the requirements for firm capacity, its deployment also results in lower requirements for peaking generation capacity (OCGT) in the system in order to maintain the security margin.

Figure 8. Changes in installed capacity driven by deployment of energy storage (50 g/kWh). Similarly to 50 g/kWh scenarios, the breakdown of the marginal system value for the two storage technologies is provided in Figure 9 for the 100 g/kWh scenarios. The results are comparable to the 50 g/kWh scenarios, with key system value components being the avoided OPEX (low-C) of low-carbon generation (i.e., CCS), avoided Gen. CAPEX (other) of conventional generation and avoided Gen. CAPEX (low-C) of low-carbon generation (as the net result of having less CCS but more wind and PV in the system with energy storage).

19

Figure 9. Breakdown of marginal system value of proposed LAES and PTES (100 g/kWh).

3.3 Impact of demand-side response Finally, energy storage is understood to be in competition with demand-side response when providing flexibility services and, therefore, an attempt is made here to quantify the impact of competing flexible providers on the whole-system value of energy storage. Two sets of sensitivity studies were performed where the uptake level of DSR was set either at a low (0%) or high (50%) level. The effect on the marginal system value of LAES and PTES is shown in Figure 10. As expected, a higher DSR uptake would result in a lower value of storage and vice versa. This occurs because the cost saving opportunities that are accessible to storage are also accessible to DSR, hence there is direct competition between the two flexible options. The effect of higher DSR uptake is moderate, with the average reduction in value across all scenarios around 10%. On the other hand, a low DSR uptake would increase the system value of energy storage by 25% on average, also implying that a cost-efficient deployment of energy storage should increase in the absence of DSR.

Figure 10. Impact of DSR uptake on marginal system value of proposed LAES and PTES. 20

3.4 Sensitivity analysis on PTES performance At the time of writing, PTES is still at a relatively early stage of development and therefore, as with most emerging technologies, there are noteworthy uncertainties around the eventual performance of its components and of the overall system. Some of the losses in PTES emerge from the potential inefficiencies of the (customized) reciprocating compression and expansion devices, which feature a number of innovative elements and are the least mature components of this technology [40,41]. In the present work, baseline values for the performance of the compressor and expander were based on values suggested and used in literature and also by the developers of this technology, who are aiming to achieve a very high performance [7,11,40]. In order to examine the impact of the performance of the compression and expansion devices on the competitiveness of the PTES technology, and given that such devices do not yet exist as envisaged by the developers of the PTES system being examined in the present work, we performed simulations for values of polytropic efficiency (which captures some of the inefficiencies of compression and expansion devices) that were lower than the baseline assumptions. Lowering the assumed polytropic efficiency from 97% to 95% reduces the roundtrip efficiency of the PTES system from 70% to 65%. Reducing it even further to 90% results in an estimated roundtrip efficiency drop to 52%. These more conservative efficiency assumptions for the PTES systems were then used in additional WeSIM simulations to estimate how the whole-system value of PTES, and consequently its competitiveness in a low carbon grid, are affected by the varying PTES system roundtrip efficiency. Similar to the analysis presented earlier in this section, PTES was evaluated in North and South of Europe scenarios for two levels of carbon reduction target (as described in Section 2.4). The whole-system value of storage for PTES under different performance parameters was estimated both per unit of power (Figure 11) and per unit of energy (Figure 12). It is evident from these figures that a reduced PTES roundtrip efficiency also results in a lower system value both in terms of power and energy, with the magnitude of this reduction being relatively small and depending on the scenario considered, and the overall trends and major qualitative conclusions remaining unchanged. Specifically, a greater reduction in system value was observed in the PV-dominated South of Europe scenario than the winddominated North. The estimated whole-system values of PTES deployment in the South decreased by 3-4% and 12-17% when using roundtrip efficiency values of 65% and 52%, respectively, compared to the baseline assumption of 70%. The equivalent value reductions in the North amounted to 1-2% and 4-8% (at 65% and 52% efficiency), respectively. Similarly, 21

a greater reduction in system value driven by a lower PTES roundtrip efficiency was observed with the carbon target of 50 gCO2/kWh than with 100 gCO2/kWh. When using a 52% PTES roundtrip efficiency value, the whole-system value of PTES (per kW) became comparable to the system value estimated for LAES. Nevertheless, given the higher power capital cost estimated for PTES it can be argued that a higher system value of storage would be desirable to justify investment.

Figure 11. Average and marginal system value of PTES per unit power under different roundtrip efficiencies.

Figure 12. Average and marginal system value of PTES per unit energy under different roundtrip efficiencies. This sensitivity analysis demonstrates the impact of the performance of compression and expansion devices on the overall performance, i.e., the roundtrip efficiency, and the wholesystem value of PTES. However, these deviations also need to be considered in light of the

22

overall uncertainty in these thermo-economic analyses and predictions, which are typically of the order of 10-20%, or higher in some cases, especially in terms of economic indicators. Differences in the system value of PTES driven by variations in roundtrip efficiency can also be used to indicate the benefits of employing higher performance components when designing energy storage systems. Depending on the system context, it may be justified to invest in more efficient but potentially also more expensive storage system components if the system benefits resulting from improved performance outweigh the additional expenditure. Exploring this relationship further represents an interesting area for future research.

4 Conclusions Consistent thermodynamic and economic models were developed and applied to determine the characteristics of LAES and PTES systems. These two thermomechanical technologies where selected as being suitable for electricity storage at medium-to-large scales (both in terms of energy and power), long storage durations, long lifetimes, multiple cycle operation, and lack of geographical constrains. Differences in key system characteristics in earlier work [11] indicated these should be tested in a network-scale model to identify the conditions in which each technology is more valuable. Therefore, their application in a whole-system model was investigated in order to determine the system value of storage under different scenarios. The aim of this paper was to estimate and proceed to explore how the system value of storage of these two technologies varies depending on the characteristics of the technologies but also of the electricity system within which they are installed. Key system assumptions were therefore varied with respect to: 1) the energy storage penetration levels; 2) the composition of low-carbon generation mix (North vs. South); and 3) the carbon emission targets (50 and 100 gCO2/kWh). The assumed DSR uptake was also varied as part of the sensitivity analysis in the paper in order to further evaluate the impact of different assumptions on the value of energy storage and capture the impact of key drivers on the value of energy storage. For both LAES and PTES, proposed unit systems sizes were taken from companies developing these technologies. The LAES system was considered at a power capacity of 12 MW and an energy capacity of 50 MWh, whereas PTES was considered at a power and energy capacities of 2 MW and 11.5 MWh [11], respectively. Through the consideration of multiple costing methods, a range capital cost was obtained for the two systems. As a result, the systems’ cost competitiveness as well as the potential for having a capital cost lower than the system value

23

threshold can depend also on the range of costs and the specific value used within the range. Therefore, it was interesting to consider this variation, which can also provide a confidence level in the costing of the system. To provide a common comparison basis, the costs of the two systems were normalized based on their corresponding power and energy capacities. The power capital costs of PTES and LAES were found to be in ranges of 930-1800 £/kW and 2200-4525 £/kW, respectively. The variation determined for PTES was observed to be greater for PTES than LAES for both power and energy capital costs. It must be noted though that these cost estimates are based on a costing framework that aims to provide a preliminary and approximate cost of what could be potentially expected for the technologies, and that they do not capture cost reductions from mass production (i.e., learning curves). It is also assumed that all system components are readily available and their costs can be captured by existing standard cost correlations used in the framework. Therefore, inevitably there is a level of uncertainty regarding the actual costs of these technologies. Future studies into refining the cost estimates as these technologies progress through the technology readiness levels (TRLs) are expected to provide improved cost estimates. The whole-system value of electricity storage was found to greatly vary depending on the cumulative installed capacity of storage in the system. Considering that the marginal system value of storage can be considered equivalent to the maximum acceptable cost of the storage system at a given penetration, we can use the cost estimates of LAES and PTES to say if the systems are attractive for implementation under different system scenarios, and at what level of installed capacity. Storage technologies provide greater marginal benefits at low penetrations and can therefore be viable in these conditions at a higher capital cost. The two carbon target scenarios showed comparable results with a positive effect of more ambitious carbon targets on the system value of storage. On the other hand, the location and installed capacity were found to have a greater impact on the system value and, hence, the acceptable cost of the technologies. One of the key drivers for the system value of flexibility provided by energy storage is its ability to reduce the requirements to invest in low-carbon generation capacity while meeting the same system-level carbon target. The results indicate that in the presence of energy storage it becomes cost-efficient to replace more flexible but also more costly low-carbon capacity (e.g., CCS) with lower-cost but less flexible variable renewables. The whole-system value of PTES was observed to be slightly higher than for LAES, driven by higher duration and efficiency, however, the higher power capital cost of PTES makes this becomes less attractive for implementation at lower volumes. Based on the lower end of cost 24

estimates, PTES was found to be close to its whole-system value at the minimum capacity considered, except in cases with low DSR uptake in which it becomes more attractive. LAES, on the other hand, based on the average estimated power capital cost, is found to be attractive for implementation at installed capacities between 5 and 10 GW in the North of Europe and between 10 and 15 GW in the South. The cost-efficient volume of LAES increases even further in scenarios with low DSR uptake. Overall, these results indicate that the complexity of the decarbonization challenge makes it difficult to identify a clear ‘best’ technology, especially considering the accuracy of earlystage modelling approaches and in particular given the difficulty of predicting the eventual costs of technologies that have yet to be built or tested at full scale, and whose components are in some cases highly customized and still in early stages of development. Therefore, the estimated performance of these systems is contingent on the development of their components at a performance level (particularly concerning the efficiencies of the large reciprocating compressors and expanders in PTES), which is based on target performance values specified by the developers and also elsewhere in the literature [7,11,40]. It is likely that the successful uptake of either technology will depend to some extent also on early adoption. In either case, the present work also suggests that the uptake of either technology can provide significant system-level benefits and that R&D into such solutions should be encouraged. Future research in this area will include exploring other costing methods/correlations for these technologies that might result in lower costs. Also, if learning curves are considered as a result of incremental installed capacity which can contribute to the reduction in cost estimates, it is possible that the systems will be economically attractive at even higher installed capacities. Also, investigating LAES and PTES at different capacities and power to energy ratios can be an interesting avenue for future work. This was illustrated in previous works [11] in which LAES was downsized to the same capacity as PTES. The impact of scale and power to energy ratio can therefore influence the economic metrics of each system. Also, the roundtrip efficiency of each technology can also be another factor that might influence the both technical and economic performance of the systems. Therefore, operating under different conditions can also impact the costing of the system, especially when economic metrics are normalized based on power and energy outputs. Furthermore, it is noteworthy that performance estimates of the technologies in this work depend on the successful development of critical components to enable their operation as assumed. Further sensitivity analyses around the impact of component

25

performance on their value as well as the competition between LAES and PTES under different assumptions on their performance could also be a potential avenue for future work.

Acknowledgements This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) [grant numbers EP/J006041/1 and EP/R045518/1] and UK Natural Environment Research Council (NERC) [grant number NE/L002515/1]. Data supporting this publication can be obtained on request from [email protected].

References [1]

Luo X., Wang J., Dooner M., Clarke J., “Overview of current development in electrical energy storage technologies and the application potential in power system operation”, Applied Energy, 2015, 137:511–36, doi:10.1016/j.apenergy.2014.09.081

[2]

Sandia National Laboratories, “DOE/EPRI 2013 Electricity Storage Handbook in Collaboration with NRECA”, New Mexico and California (USA), Sandia National Laboratories, 2013, Report No. SAN2013-5131.

[3]

Chen H., Cong T. N., Yang W., Tan C., Li Y., Ding Y., “Progress in electrical energy storage system: A critical review”, Progress in Natural Science, 2009, 19(3):291–312, doi:10.1016/j.pnsc.2008.07.014.

[4]

Mahlia T. M. I., Saktisahdan T. J., Jannifar A., Hasan M. H., Matseelar H. S. C., “A review of available methods and development on energy storage; technology update”, Renewable

and

Sustainable

Energy

Reviews,

2014,

33:532–545,

doi:10.1016/j.rser.2014.01.068. [5]

Elkind E. N., “The power of energy storage. How to increase deployment in California to reduce greenhouse gas emissions”, 2010.

[6]

White, A., Parks, G., Markides, C.N., “Thermodynamic analysis of pumped thermal electricity

storage”,

Applied

Thermal

Engineering,

2013,

53(2):291–298,

doi:10.1016/j.applthermaleng.2012.03.030. [7]

McTigue, J.D., White, A.J., Markides, C.N., “Parametric studies and optimisation of pumped

thermal

electricity

storage”,

doi:10.1016/j.apenergy.2014.08.039.

26

Applied

Energy,

2015,

137:800–811,

[8]

Frate G.F., Antonelli M., Desideri U., “A novel pumped thermal electricity storage (PTES) system with thermal integration”, Applied Thermal Engineering, 2017, 121:1051–8, doi:10.1016/j.applthermaleng.2017.04.127

[9]

Roskosch D., Atakan B., “Pumped heat electricity storage: Potential analysis and ORC requirements”, Energy Procedia, 2017, 129:1026–33, doi:10.1016/j.egypro.2017.09.235

[10] Sciacovelli A., Vecchi A., Ding Y., “Liquid air energy storage (LAES) with packed bed cold thermal storage – from component to system level performance through dynamic modelling”, Applied Energy, 2017, 190:84–98, doi:10.1016/j.apenergy.2016.12.118 [11] Georgiou, S., Shah, N., Markides, C.N., “A thermo-economic analysis and comparison of pumped-thermal and liquid-air electricity storage systems”, Applied Energy, 2018, 226:1119–33, doi:10.1016/j.apenergy.2018.04.128 [12] Morgan R., Nelmes S., Gibson E., Brett G., “Liquid air energy storage – analysis and first results from a pilot scale demonstration plant”, Applied Energy, 2015, 137:845–53, doi:10.1016/j.apenergy.2014.07.109 [13] Sciacovelli A., Smith D., Navarro H., Li Y., Ding Y., “Liquid air energy storage – operation and performance of the first pilot plant in the world”, Proceedings of the 29th international

conference

on

efficiency,

cost,

optimization,

simulation

and

environmental impact of energy systems, 2016, Portorož, Slovenia. [14] Markides, C.N., “The role of pumped and waste heat technologies in a high-efficiency sustainable energy future for the UK”, Applied Thermal Engineering, 2013, 53(2):197– 209, doi:10.1016/j.applthermaleng.2012.02.037. [15] Pudjianto, D., Aunedi, M., Djapic, P., Strbac, G., “Whole-system assessment of value of energy storage in low-carbon electricity systems”, IEEE Transactions on Smart Grid, 2014, 5(2):1098–1109, doi:10.1109/tsg.2013.2282039. [16] Palmintier, B.S., Webster, M.D., “Impact of operational flexibility on electricity generation planning with renewable and carbon targets”, IEEE Transactions on Sustainable Energy, 2016, 7(2):672–684, doi:10.1109/tste.2015.2498640. [17] de Sisternes F.J., Jenkins J.D., Botterud A., “The value of energy storage in decarbonizing

the

electricity

sector”,

Applied

Energy,

2016,

175:368–379,

doi:10.1016/j.apenergy.2016.05.014. [18] Bitaraf H., Rahman S., “Reducing curtailed wind energy through energy storage and demand response”, IEEE Transactions on Sustainable Energy, 2018, 9:228–236, doi:10.1109/tste.2017.2724546.

27

[19] Aunedi M, Pudjianto D., Strbac G., “Calculating system integration costs of low-carbon generation technologies in future GB electricity system”, 5th IET International Conference on Renewable Power Generation (RPG) 2016, doi:10.1049/cp.2016.0529. [20] Xu B., Wang Y., Dvorkin Y., Fernández-Blanco R., Silva-Monroy C.A., Watson J.P., Kirschen D.S., “Scalable planning for energy storage in energy and reserve markets”, IEEE

Transactions

on

Power

Systems,

2017,

32:4515–4527,

doi:10.1109/tpwrs.2017.2682790. [21] Carrión M., Dvorkin Y., Pandžić H., “Primary frequency response in capacity expansion with energy storage”, IEEE Transactions on Power Systems, 33:1824–1835, 2018, doi:10.1109/tpwrs.2017.2735807. [22] Strbac, G., Aunedi, M., Konstantelos, I., Moreira, R., Teng, F., Moreno, R., Pudjianto, D., Laguna, A., Papadopoulos, P., “Opportunities for energy storage: Assessing wholesystem economic benefits of energy storage in future electricity systems”, IEEE Power and Energy Magazine, 2017, 15:32–41, doi:10.1109/mpe.2017.2708858. [23] Teng, F., Pudjianto, D., Aunedi, M., Strbac, G., “Assessment of future whole-system value of large-scale pumped storage plants in Europe”, Energies, 2018, 11(1):246, doi:10.3390/en11010246. [24] Highview Power Storage, “Liquid air energy storage 2016”, available at: www.highview-power.com [accessed 15.2.2016]. [25] Centre for Low Carbon Futures, “Liquid air in the energy and transport systems: Opportunities for industry and innovation in the UK”, York, UK: Centre for Low Carbon Futures; 2013, Report No. 021. [26] Mathie, R., Markides, C.N., White, A.J., “A framework for the analysis of thermal losses in reciprocating compressors and expanders”, Heat Transfer Engineering, 2014, doi:10.1080/01457632.2014.889460. [27] Willich, C., Markides, C.N., White, A.J., “An investigation of heat transfer losses in reciprocating devices”, Applied Thermal Engineering, 2017, 35(16-17):1435–1449, doi:10.1016/j.applthermaleng.2016.09.136. [28] White, A.J, McTigue, J.D., Markides, C.N., “Analysis and optimisation of packed-bed thermal reservoirs for electricity storage applications”, Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 2016, 230(7):739–754, doi:10.1177/0957650916668447.

28

[29] White, A., McTigue, J., Markides, C.,” Wave propagation and thermodynamic losses in packed-bed thermal reservoirs for energy storage”, Applied Energy, 2014, 130:648– 657, doi:10.1016/j.apenergy.2014.02.071. [30] Georgiou, S., Mac Dowell, N., Shah, N., Markides, C.N., “Thermo-economic comparison of liquid-air and pumped-thermal electricity storage”, Proceedings of ECOS 2017, San Diego, California, USA. [31] Turton, R., Bailie, R.C., Whiting, W.B., Shaeiwitz, J.A., “Analysis, synthesis and design of chemical processes”, London, UK, Pearson Education International, 4th Ed., 2013. [32] Seider W.D., Seader J.D., Lewin D.R., Widagdo S., “Product and design principles: synthesis analysis, and evaluation”, 3rd ed, Asia: John Wiley & Sons (Asia) Pte Ltd, 2010. [33] Couper J.R., Penney W.R., Fair J.R., Walas S.M., “Chemical process equipment: Selection and design”, revised second ed. MA (USA), Butterworth-Heinemann, Elsevier Inc., 2010. [34] Hewitt, G.F., Pugh, S.J., “Approximate design and costing methods for heat exchangers”,

Heat

Transfer

Engineering,

2007,

28(2):76–86,

doi:10.1080/01457630601023229. [35] Lemmens, S., “Cost engineering techniques and their applicability for cost estimation of organic Rankine cycle systems”, Energies, 2016, 9(7):485, doi:10.3390/en9070485. [36] Jenkins S., “Economic indicators: CEPCI”, Chemical engineering essentials for the CPI professional,

available

at:

www.chemengonline.com/economic-indicatorscepci

[accessed 5.11.2016]. [37] Pudjianto D., Djapic P., Aunedi M., Gan C. K., Strbac G., Huang S., Infield D., “Smart control for minimizing distribution network reinforcement cost due to electrification”, Energy Policy, 2013, 52:76–84. [38] Papadaskalopoulos D., Strbac G., Mancarella P., Aunedi M., Stanojevic V., “Decentralized participation of flexible demand in electricity markets—Part II: Application with electric vehicles and heat pump systems,” IEEE Transactions on Power Systems, 2013, 28(4):3667–3674, doi:10.1109/tpwrs.2013.2245687. [39] “BP Technology Outlook 2018: How technology could change the way energy is produced and consumed”, available at: www.bp.com/technologyoutlook [accessed 15.3.2018].

29

[40] Howes J., “Concept and development of a pumped heat electricity storage device”, Proceedings of the IEEE, 2012, 100(2):493–503, doi:10.1109/jproc.2011.2174529. [41] Willich C., White A.J., “Heat transfer losses in reciprocating compressors with valve actuation for energy storage applications”, Journal of Energy Storage, 2017, 14(2):322– 328, doi:10.1016/j.est.2017.07.024.

30

Highlights • • • • •

PTES has higher power and energy capital costs than LAES The variation in the capital costs of PTES is higher than that of LAES The whole-system value of PTES is slightly higher than LAES The whole-system value of the technologies varies depending on deployment Lower carbon targets increase the system value of energy storage

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: