Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Techno-economic analysis of the Li-ion batteries and reversible fuel cells as energy-storage systems used in green and energy-efficient buildings

Techno-economic analysis of the Li-ion batteries and reversible fuel cells as energy-storage... PV Modules Combustible Gas Detector DC input One-way Valve DC/DC Convertor Dryer Hydrogen Storage DC/AC Inverter Back Pressure Stack Regulator Oxygen Pump Vent 80-90°C Controllable City Water Water Valve Water Cleaner 20°C Expander Air Compressor FC Mode EC Mode Keywords: green-building energy-storage systems; fuel cell; hydrogen; Li-ion batteries; reversible fuel cells energy-efficient through design modifications if connected Introduction to the power grid. The most popular renewable-energy sys- The industrial, residential and commercial sectors were tems for commercial buildings are either in stand-alone responsible for 35%, 16% and 12% of the total energy con- configuration or connected to the grid that is fed by a sumption in the USA in 2019 [1]. Globally, energy con- large portion of renewable energy, photovoltaic (PV) solar sumption of the building sector accounts for ~30% of the and wind systems. For the stand-alone configuration, it world’s total energy demand, which includes electrical seems that PV panels are more practical to use and easier power, heating and cooling loads [2]. Green buildings and to install. PV panels can be installed on the surface of the retrofit energy-efficient buildings have become a trend building, which allows combining electrical-energy pro- in building-energy science research in the last few years. duction with other functions of the building structures [7]. The performance of net-zero-energy buildings has gained Integrating PV panels with the building façade has also be- more attention since the publication in 2010 of the EU come a popular choice for many modern building designs Energy Performance of Buildings Directive recast [3]. In [4]. Recent studies also recommend using concentrated the USA, the federal government and many state govern- solar panels (CSP), which can decrease the area needed to ments have promoted ‘marketable zero energy homes in install PV cells by using low-cost transparent material in 2020 and commercial zero energy buildings in 2025’ or smaller areas, thus decreasing the overall system cost and similar programmes [4]. Japan also attempts to promote increasing the temperature of the heat source. This fact ‘carbon-neutralized buildings’, including existing build- also implies that CSP can perform better in several func- ings, by 2050 [5]. Thus, green-building design is becoming tions such as in heating, refrigeration, dehumidification broadly adopted in commercial and residential sectors. By and lighting for such buildings [2]. definition, green buildings are designed to minimize the Alongside the extensive research on the new de- impacts on the environment by applying techniques to sign and required policies for green and energy-efficient reduce energy usage and water usage, and by minimizing buildings, many researchers have also analysed the the environmental disturbances from the building site [6]. cost of energy production, consumption and storage in Green buildings also aim to improve human health and these energy-efficient buildings. In their comprehen- workplace environments through the design of healthier sive review of the energy-storage systems (ESSs) used in indoor environments [6]. energy-efficient buildings, Chatzivasileiadi et al. [8] found Green buildings can be 100% self-sufficient if connected that battery storage systems (including Li-ion, Zn-air and to a suitably sized renewable-energy system or can be more Compressor Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 275 NaNiCl batteries) are among the most promising and eco- last type of ESS technology is the thermal energy-storage nomic ESSs given the short-term storage needs for ESSs (TES) system, which is capable of storing heat or cold in connected to commercial buildings. In another study a storage medium at certain temperatures for further made by Tribioli and Gonzolino [9 ], the authors modelled usage, under different conditions such as temperature, and discussed the economics of energy production and place or power [16]. TES systems are generally classified storage using batteries and reversible fuel cells (RFCs). into three different categories, depending on the thermo- In this paper, the authors modelled a stand-alone poly- dynamic properties of the storage medium, which include generation power plant for a strip mall located in eight sensible heat, latent heat, absorption and adsorption different climate zones in the USA (Minneapolis, MN; system. TES systems are commonly used as ESSs for in- Houston, TX; Las Vegas, NV; Los Angeles, CA; Miami, FL; dustrial and residential purposes, such as space heating New York, NY; Denver, CO; and Seattle, WA). In their model, or cooling, process heating and cooling, hot-water produc- they coupled a photovoltaic panel array to a battery and a tion and circulation in buildings, and applications that re- unitized regenerative polymer electrolyte membrane fuel quire phase-change material [13–15]. A  great comparison cell as primary storage and a diesel generator as a sec- of different ESSs was made by Mostafa et  al. [17]. In that ondary backup system. Their cost model indicates that paper, the authors studied several ESSs for short-term and this poly-generation system can produce energy at a long-term storage needs and found that battery storage levelized cost as low as 8.6 cents/kWh in Seattle, WA and systems like Li-ion, acid, nickel-cadmium (NiCd), sodium- ≤42.6 cents/kWh in Denver, CO. Interestingly, this study sulphur (NaS), vanadium redox and zinc-bromine (ZnBr) found that the adoption of this poly-generation system are more suitable for short- and medium-term storage ap- can mitigate between 61.23% and 99.60% of the carbon- plications, while compressed air, hydrogen-based energy dioxide emissions. storage and pumped hydro are more suitable for medium- Among the different challenges that are still facing to long-term storage applications. They also found that the green buildings are the high initial cost, energy efficiency annualized life-cycle cost of storage (LCCOS) in $/kWh- of the installed systems and intermittent nature of the year for long-term storage systems such as underground electricity generated by the renewable-energy systems. compressed air, aboveground compressed air and pumped Having an efficient ESS is still considered a big challenge hydro tend to decrease with a system lifetime of ≤40 years. that engineers need to carefully plan for. The authors found that, unlike long-term storage systems, ESSs can be classified into five main classes: chem- short- and medium-term storage systems such as battery ical, electrochemical, electrical, mechanical and thermal.stor age technologies and hydrogen-based energy storage Chemical ESSs store energy in the chemical bonds of can see a decrease in the LCCOS in the first 20 years, then atoms and molecules, which then can be released in an increase in the LCCOS because the variable operation chemical reactions to recover the stored energy. After the and maintenance costs for medium-term ESSs are less release of chemical energy, the substance is often changed than the corresponding costs calculated for long-term into an entirely different substance [10]. The stored chem- ESSs, which can be directly linked to the fact that the dis- ical energy can be released via electron-transfer reac- charge periods of short- and medium-term ESSs are lower tions that produce electricity. Chemical energy storage than the discharge periods of long-term ESSs. focuses on the production of hydrogen, ammonia and Among different ESSs, it seems that rechargeable bat- synthetic natural gas as secondary energy carriers [11, tery technologies such as lead-acid, NaS and Li-ion bat- 12]. Electrochemical ESSs convert chemical energy into teries (LIBs) are common ESS technologies, especially for electrical energy. Electrochemical cells can be cate -gor small-scale stationary energy-storage applications [18], ized into four types, depending on their function: primary but still possess some limitations that limit their wide- cells or batteries, secondary cells or secondary batteries, spread deployment such as having shorter life compared reserve cells and fuel cells [13]. The third type of ESS tech- to other battery energy storage and other ESS technologies, nology is electrical ESSs, which also can be grouped into low energy density and, to a lesser extent, high cost in dol- two subgroups: electrostatic systems, including capaci- lars/kWh of storage capacity, especially for Li-ion and NaS tors and supercapacitors; and magnetic/current ESSs [14]. technologies [19, 20]. New battery technologies (e.g. redox The fourth ESS technology is mechanical energy-storage flow and zinc-iron batteries) are promising ESS technolo- systems (MESSs), which are usually classified based on gies, but do not come without limitations. Such limitations their working principles. Examples of MESSs include sys- include their high initial cost and low energy density [21]. tems that store energy based on the forces associated with In their paper, Mostafa et  al. [22] developed a techno- pressurized gas, forced springs, kinetic energy and poten- economic model and optimal energy management for tial energy [15]. The most useful advantage of mechanical a grid-connected micro grid that was dependent on the ESSs is that they can readily deliver the energy whenever renewable-energy resources and different battery storage required in a short period of time, besides being adapt- technologies including LA, NiCd, Li-ion and NaS. They able [13]. Flywheels, pumped hydro storage systems and studied the effect of certain design and performance compressed-air ESSs are also classified as MESSs [13]. The parameters such as initial charge, depth of discharge and Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 276 | Clean Energy, 2021, Vol. 5, No. 2 the number of charging/discharging cycles with the goal The selection of the PEM-RFC and LIB systems is based of minimizing the total operating cost of the system and on their flexibility to operate as load-following systems maximizing the benefits of battery storage systems. Other and because of their short response times to the changing goals also included minimizing the investment and re- loads and ability to work on partial loads [24–26]. Many placement costs of a battery storage system, and minim- published techno-economic analyses have been focusing izing the operation and maintenance costs of a distributed on the relative comparison between ESSs from economic generation system. Mostafa and co-authors used a gen- and performance perspectives and their useful applica- eral algebraic modelling system to solve the deterministic tion in the short or long term. In this paper, we attempt to optimization problem; stochastic optimization was also compare two competing technologies that are suitable for used to model the market-price uncertainty and gap deci- smaller-scale applications (<1 MW), which allow them to sion theory to model the electric-load uncertainty. Among be connected in a stand-alone configuration to commer - many interesting findings, this group of researchers found cial or industrial buildings. This paper discusses the re- that the total cost per day of the battery storage system sults of the techno-economic model that we developed for through the lifetime of the project decreased when the a PEM-RFC system that is designed to have dual functions: value of the depth of discharge (DoD) increased and that (i) an ESS to produce hydrogen in the electrolysis mode NaS batteries had better potential in reducing the total and power in the fuel-cell mode and (ii) as a hydrogen- operating cost of microgrids. production system to use excess produced hydrogen as a Hydrogen has been investigated as a medium for fuel for electric fuel-cell vehicles or other end uses. Then, long-term energy-storage applications and as a tech- the same model was used to assess and compare LIBs nology to improve the resiliency of power grids [ ].8 The with different sizes and to evaluate their relative perform- advancements in fuel-cell technologies have encouraged ance with RFCs. Parametric sensitivity analysis aims at as- researchers to study RFCs as a viable ESS to store energy sessing certain designs and financial parameters and to for medium to long periods of time (months to years). study their effects on the LCOS. Users can use this ESS as a flexible system that can pro- The rest of the paper is organized as follows: Section 1 duce and store hydrogen fuel for several months or years discusses the methodology used in sizing the PV system without losing the nominal storage capacity. Rechargeable and ESS using the building-energy profile, the key param- batteries, on the other hand, are known for their gradual eters used in developing the techno-economic models for losses of stored energy through self-discharging if left LIB and RFC, and the development of the mathematical without use for longer periods of time [23]. Unitized RFCs formula to calculate the LCOS. Section 2 includes the re- use bifunctional anode/cathode electrodes, allowing the sults and discussion of the LCOS models, and includes a RFC to operate in both electrolysis and fuel-cell modes discussion on the sizing of the hydrogen storage for RFCs using the same stack [11]. This configuration is similar to based on the building demand and supply curves; com- the rechargeable batteries that operate under both c-har parison between LIBs and RFCs, and sensitivity analysis ging and discharging conditions. In electrolysis mode, the are presented. Section 3 concludes with a summary of the RFC takes in electricity and water and produces hydrogen high-level findings and insights from this work. as a fuel and oxygen as a by-product. In fuel-cell mode, the RFC utilizes hydrogen and oxygen (from air), and pro- 1 Methodology duces electricity and water as by-products. RFCs have a wide range of applications that range from energy- In this section, we discuss the electricity profile for com- storage devices coupled to renewable-energy sources and mercial buildings in Section 1.1, followed by discussion of intermittent power grids, backup power plants and in the design of the PV solar system and its technical param- spacecrafts [11]. eters in Section 1.2. Working principles and logic of the two It seems that the attractiveness of promising storage selected ESSs (RFC and LIB) and the technical and financial technologies, such as reversible proton-exchange mem- factors used in developing the techno-economic model are brane (PEM) fuel cells, which can compete with recharge- all discussed in Section 1.3. Finally, the mathematical for - able stationary LIBs, has not been thoroughly explored mulas and the underlying assumptions used in developing in the literature. Thus, this paper aims to fill this gap by the LCOS are presented in Section 1.4. The flow diagram evaluating the flexibility and economics of energy storage in Fig. 1 shows the sequence of steps and analyses con- using LIBs as a mature ESS technology and RFCs as an ducted in this study. The first step was the analysis of the emergent ESS technology. A  techno-economic analysis building-energy profiles using an average year from the TM is conducted for this purpose to assess the performance EnergyPlus program [27], then we used building-energy and economics of energy storage using LIBs and unitized- profiles to estimate the PV solar-system size in kW. The PV stack reversible PEM fuel cells (PEM-RFC). The effects of system was designed to ensure a self-sufficient building system design and some selected financial parameters on in a stand-alone configuration—that is, this building is not the levelized cost of energy storage (LCOS) are particularly connected to any power grids other than its own PV solar analysed and discussed in this paper. power system. The economics of energy production from Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 277 Analysis of Building Energy Demand and Supply PV Solar System (Design and Economics (LCOE)) Energy Storage System – RFC Energy Storage System – LIB (System Design and Economics) (System Design and Economics) Hydrogen Energy Storage Dynamics – LIB Production/Consumption (Charging/Discharging Cycles) Sensitivity Analysis - RFC Sensitivity Analysis - LIB (Design Parameters; Financial Factors) (Design Parameters; Financial Factors) Fig. 1: Flow diagram showing the approach and analysis used in this study El Paso, TX [kWh] the PV system were estimated using the levelized cost of energy (LCOE) to estimate the cost component of electricity in the energy-storage equation (Equation (1)). The LCOS was used as the main metric to assess the studied ESSs (LIB 140 and RFC) from an economic perspective, while resiliency was estimated using the ability of the system to provide all required building demands without stopping. Finally, sen- sitivity analysis was conducted to study the effect of the most important factors that play key roles in determining the final value of the energy storage. These factors include two sets: the first set includes system-design parameters (ESS capital cost, operating and maintenance (O&M) cost, 0 16-Dec 4-Feb 26-Mar 15-May4-July 23-Aug 12-Oct 1-Dec electricity cost and roundtrip efficiency), while the second set includes discount rate and system lifetime. Fig. 2: Energy profiles for a medium office building in El Paso, TX simulation program that is used to model both energy 1.1 Commercial-building electricity profiles consumption in heating, cooling, ventilation, lighting and plug and process loads, and water use in buildings. In The building-performance simulations are used to identify this analysis, we use the energy profile for medium office each energy end use, which includes space heating, space buildings in El Paso, TX. According to the EnergyPlus def- cooling, domestic hot water, and appliances and lighting. initions, the medium office building has an area of 4982 In many cases, electricity may be used to meet  all of the 2 2 1 m (53 628 ft ) and consists of three storeys . Fig. 2 shows end uses, especially for those buildings that are not con- sample energy profiles for a medium office building in El nected to the natural-gas distribution network. Nearly Paso, TX. A  quick look at this energy profile shows that all space-cooling, appliance and lighting needs are pro- energy consumption increases in the summer months in vided by electricity; however, space heating and domestic El Paso, TX due to space-cooling requirements. The min- hot water can be met with other energy sources, such as imum, mean and maximum electricity consumptions in natural gas, heating oil and wood, or via central district the medium office building in El Paso, TX are 11.4, 53.0 and heating systems that supply hot water to several buildings 164.2 kWh, respectively. using a central heating source [28]. The electricity profiles used in this work were pulled from the EnergyPlus platform provided from the US https://www.energy.gov/eere/buildings/commercial-reference- buildings. Department of Energy [27]. EnergyPlus is a building-energy Electricty (kWh) Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 278 | Clean Energy, 2021, Vol. 5, No. 2 positioned itself as one of the most competitive hydrogen- 1.2 PV-system design production and energy-storage technologies because of The design parameters for the PV system were deter - the high energy-conversion efficiency and low operating mined based on the building load profile with the goal temperatures (<120°C) [3233 , ]. The high capital cost, which of having a stand-alone self-sustained system that can is partially attributed to the use of precious platinum- supply enough electricity to the building during the day- group metals (PGMs) and expensive selective PEMs in the time and can divert excess energy during the daytime electrode assemblies, is one of the biggest limitations to the ESS to be stored and used later when there is no that limit the adoption of PEM technology relative to the sunshine. Based on the load profiles, the estimated nom- other electrolysis/fuel-cell technologies such as alkaline inal PV-system size was set to 400 kW . The selected PV DC electrolysers and solid-oxide fuel cells [33]. In the USA, module was SunPower SPR-E19-310-COM, which has a Proton (NEL), Giner and other fuel-cell manufacturers nominal efficiency of 19.02% and can provide a maximum have started to produce megawatt-scale PEM electrolysers power of 310.15 W . The DC/AC inverter is SMA America: DC as a means to reduce the manufacturing cost. In addition, SC750CP-US, which has an average DC-to-AC conversion many R&D institutes are also working on improving the efficiency of 97.59%. Other cost and design parameters conversion efficiency of the PEM electrolysers—another are summarized in Table 1. The System Advisor Model factor that has historically limited the adoption of this (SAM 2020.2.29) from the Department of Energy’s National technology. Today, PEM electrolysers have achieved an im- Renewable Energy Laboratory (NREL) was used to model portant milestone in terms of the higher current densities the PV system. The SAM is a free techno-economic soft- (~5× that of the alkaline electrolysers) and higher effi- ware model that is used to model the central large-scale ciencies (nearly 6% higher than that value for alkaline and distributed small-scale renewable-energy systems electrolysers) [3435 , ]. and facilitates the decision-making process for people The unitized-stack reversible fuel cell (URFC) is a new working in the renewable-energy industry [29]. energy-storage technology that utilizes the same stack to perform dual functions: hydrogen production (electrolysis 1.3 ESS design and functional parameters mode) and electricity generation (fuel-cell mode). The overall reaction mechanism inside the URFC is as follows: Proton-exchange membrane fuel cells (PEMFCs) and + − electrolysers (PEMECs) are the most favourable fuel-cell (1) In the fuel cell mode : 4H + 4e + O → 2H O 2 2 and electrolysis technologies when the applications re- + − In the electrolysis mode : 2H O → 4H + O + 4e(2) 2 2 quire a fast response due to the changing nature of the load and when the system is required to work in partial-load A schematic of the PEM-RFC that is connected to the PV mode (also called turndown ratio) [ ]. PEM fuel-cell tec 9 h- system and the office building is shown in Fig. 3. The PV nology is considered the best type of fuel-cell technology system is the main source of power, while the PEM-RFC for transport applications in fuel-cell electric vehicles andher e represents the primary ESS. Inside the RFC system, material-handling equipment, and in backup power sys- we can see that the stack represents the core subsystem tems thanks to its low operating temperatures (<120°C) in which electrochemical reactions take place. The stack and good efficiency relative to other fuel-cell technologies is made of tens to hundreds of repeated cells called mem- [30, 31]. PEM-electrolysis technology, a sister technology of brane electrode assemblies that are sandwiched between the PEM fuel cells, is used to split water molecules into metallic or composite bipolar plates that should have hydrogen and oxygen. PEM-electrolysis technology has good corrosion-resistance properties [36]. The name of Table 1: Design parameters for the PV solar system Parameter Value Unit Nominal system size 400 kW DC PV-module type SunPower SPR-E19-310-COM PV-module capacity Up to 310.15 W DC DC/AC inverter SMA America: SC750CP-US DC/AC inverter efficiency 97.587% Maximum DC voltage 820 V DC System losses 5.44% Annual DC-degradation rate 0.50% per year Total module area 2094 m PV-system capital cost $578 351 (including contingency = 3% of subtotal cost) Total indirect cost $72 723 (engineering, permitting and environmental, land and sales tax) Total installed system cost $651 074 Total installed cost per capacity $1.63/W DC Incentives and taxes Variable (location-specific) including 26% in federal investment tax credit Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 279 PV Modules Combustible Gas Detector DC input One-way Valve DC/DC Convertor Dryer Hydrogen Storage DC/AC Inverter Back Pressure Stack Regulator Oxygen Pump Vent 80-90°C Controllable City Water Water Valve Water Cleaner 20°C Expander Air Compressor FC Mode EC Mode Control DC/AC Unit Inverter PV Modules DC input DC/DC Convertor Fig. 3: Schematics of (a) a unitized-stack PEM-RFC ESS and (b) an LIB ESS this technology comes from the structure of the mem- of different parts that are assembled and attached to the brane, which is usually made of a solid polymer electrolyte stack to supply the power and transport fluids used in the membrane that is deposited with a catalyst layer made electrochemical reactions inside the stack and to collect from PGMs, which include platinum, iridium and ruthe- and dry the gases (mainly hydrogen) produced from the nium. The catalyst coated membrane, or CCM (the mem- electrolysis process. Several subsystems are included in brane with electrode layers), is then attached to carbon- or the BOP, such as power electronics (AC/DC rectifier, DC/ titanium-based gas-diffusion layers from both sides [11]. AC inverter or a bidirectional inverter that can replace The main function of the gas-diffusion layer is to facili- the individual AC/DC rectifier and DC/AC inverter in one tate the diffusion of water and gas molecules to and from dual-function system, transducers, electric wiring, re- the catalytic layers. The second important system in the lays and switches), water-management system (water PEM-RFC is the balance of plant (BOP), which is composed pumps, heat exchangers, pipes and hoses, manifolds, Compressor Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 280 | Clean Energy, 2021, Vol. 5, No. 2 etc.), thermal-management system (e.g. heat exchan- add the membrane electrode assembly parts and for the gers, thermal host to recover waste heat, air blowers and gasket/seal on the edges. The estimated cell power density manifolds) and hydrogen-processing and storage unit for the electrolyser stack is 0.85 W/cm . The estimated (hydrogen dryer, single- or multiple-stage compressors power density of the fuel-cell stack is 0.409 W/cm . and stationary storage tanks) [11 33 , , 36]. The sizes of the RFCs and LIBs are based on the elec- Fig. 4 shows the polarization curve for a PEM-RFC. In tricity profile for the building. It was also assumed that the base case, we set the current densities for the fuel cell the building will have a stand-alone PV-ESS so that there and electrolyser at 0.56 and 0.50 A/cm , respectively. The is no need to connect the building to the power grid. The reference potential, j, for the fuel cell is set at 0.73 V, while estimated gross power for the fuel cell and electrolyser are the reference potential is set at 1.7 V for the electrolysis 162 and 396 kW, respectively. Parasitic losses, power losses operating mode. These values give an estimated stack effi- in the BOP, were estimated at 12% and 15% of the gross ciency of ~42.9% in the ideal case. The stack roundtrip effi- power for the electrolysis and fuel-cell operating modes, ciency can be calculated using the following equation: respectively. The functional specifications for the LIB ESS are sum- Q V FC FC η = η .η = (3) RT,stack FC EC marized in Table 3. The base system is an LIB that con- Q V EC EC tains nickel–manganese–cobalt in the cathode layer and where Q is the charge transferred during the fuel-cell FC graphite in the anode layer. The electrolyte is usually made operating mode, Q is the charge transferred in the elec- EC from lithium salts (LiPF ). The LIB ESS was designed to have trolysis mode, V is the supplied voltage to the cell in the FC charging and discharging rates (c-rates) of 0.50, which is fuel-cell operating mode and V is the supplied voltage EC equivalent to 250 kWh/h for a 500-kWh LIB system, with to the stack in the electrolysis mode. Under the ideal con- the lowest DoD of 125 kWh. The roundtrip efficiency was version cycle of the closed-loop H /H O storage system 2 2 assumed to be 90% [36, 37]. Unlike LIBs, RFCs can have zero with no current leakage, Equation (3) can be reduced to DoD, which happens when the fuel-cell mode (discharging Equation (4) as follows: mode) in the RFC can no longer operate due to the unavail- FC ability of hydrogen in the storage tanks. For consistency η = (4) RT,stack EC in the economic analysis, we assumed the same financial It is important to mention that the system roundtrip effi- ciency is usually less than the stack roundtrip efficiency Table 2: Functional specification of the RFC ESS because of the parasitic losses in the electrolysis process Parameter Electrolyser Fuel cell Unit/notes and fuel-cell BOP components. The system roundtrip effi- ciency can be expressed as: System power 396 161.88 kW Gross system power 450 190.45 kW (W − W ) stack BOP FC η = (5) RT,system Stack operating tem- 60–90 60–90 °C (W + W ) stack BOP EC perature range where W is the energy consumed by the stack and Stack operating 1–20 1–20 bar stack W is the parasitic loss that represents the energy con- pressure range BOP Total platinum- 1.5 1.5 g/m sumed by the BOP parts. Subscripts FC and EC in the above group-metal equations refer to the fuel-cell and electrolysis operating loading modes, respectively. Total cell area 800 800 cm The functional specifications of the cell, stack and key Cell active area 680 680 cm parameters for the BOP system are summarized in Table 2. Current density 0.5 0.56 A/cm The assumed cell total area is 800 cm with an active area Reference potential 1.7 0.73 V 2 2 of 680  cm , leaving 120  cm for the binding perimeter to Power density 0.850 0.409 W/cm Single-cell power 578.00 277.984 W Polarisation Curve for PEM-RFC Cells per system 685 2.5 Cells per stack 98 Q V FC FC RT, stack = Q V Stack per system 7 EC EC 2.0 Stack electrical 75% 48% efficiency 1.5 BOP power 12% 15% Of the 1.0 consumptions rated system 0.5 power Electrolysis Mode Fuel Cell Mode System efficiency 36.0% 0.0 –1,500 –1,000 –500 0500 1,000 1,500 2 Based on the target values by the US Department of Energy’s Hydrogen Current Density (mA/cm ) and Fuel Cell Technologies Office [44, 45]. Fig. 4: Polarization curve for a PEM-RFC stack Percentage of the nominal system power. Potential (V) Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 281 Table 3. ESS design and cost model parameters Parameter RFC LIB Units Notes System-design parameters Nominal system 450 (EC) 500 (nom- kW for RFC Unitized stack that operates in FC/EC modes size 162 (FC) inal size) kWh for LIB Battery/fuel-cell Proton-exchange NMC111 NMC111 (cathode) and graphite (anode) chemistry membrane Discharge rate 162 250 kW For RFC: fuel-cell efficiency = 48% and can supply (kWh/h) ≤162 kWh/h LIB: can discharge 250 kWh/h Lowest depth of 0 125 kWh Fuel cell can operate until the hydrogen is no discharge (DoD) longer available in the hydrogen-storage tanks Storage-system cost Variable ($20–750 N/a Using cost values for stationary steel–molyb- per kg-H ) denum tanks from [46] and for underground storage in salt cavern from [47] Roundtrip efficiency 36% 90% Average roundtrip efficiencies for PEM-RFCs [24] and LIBs [26, 40] Economic-analysis parameters Discount rate 10% 10% System lifetime 20 20 Year Estimated stack 4000 5000 Cycle Assuming 10 000 cycles for RFCs, where voltage lifetime fade does not to exceed 20% of the original ref- erence voltage [38, 48] Assuming 5000 cycles for nickel–manganese–co- balt LIBs, with storage-capacity fade of 20% [29] Price of electricity 0.022 0.022 $/kWh Using the real levelized cost of electricity (LCOS) generated by the PV system O&M cost 0.02 0.02 $/kWh Estimated value using values from Reznicek and Braun [21] for similar ESS technology System capital cost 600 350 $ Adjusted value based on the system capital cost for PEM electrolysers discussed in Mayyas et al. [33] and average values for LIBs from [26, 49, 50] Installation factor 25% 25% $/kW References [36, 46] Installed system 337 500 218 175 capital cost (TIC) for ESSs Cycles per year Variable Variable Transition from charging to discharging modes is considered one cycle NMC111 formula is LiNi Mn Co O . 0.33 0.33 0.33 2 We assumed a stationary steel–molybdenum storage tank is used for a low storage amount that is ≤1000 kg) (based on the cost estimates obtained from [46]) and a bulk storage system using underground salt caverns when the storage size is >1000 kg (cost estimates obtained from Ahluwalia et al. [47]). O&M cost per kWh discharged electricity from the ESS. Equivalent to 3.8–6.6% of the annual amortized installed system cost for LIBs and 8.6% of the annual amortized installed system cost for RFCs. Excluding the hydrogen-storage system for RFCs. parameters (discount rate, system lifetime and installa- temperature, charging and discharging rates, lowest DoD tion factor) for RFCs and LIBs. We also assumed the O&M and cathode chemistry [8 ]. The estimated lifetime for an cost to be 2 cents/kWh discharged for both ESSs [11]. The NMC111 nickel–manganese–cobalt LIB is 5000 cycles as- O&M cost is estimated based on similar values from the suming the lowest DoD of 25% [29] with a remaining literature, which accounts for the cost of operating and storage capacity of 75% of the original nominal storage maintaining the system during its expected useful life. capacity in kWh. It is important to highlight that the ESS Most of the O&M cost is associated with the cost of refur - lifetime is 20  years, which indicates that RFC and LIB bishing or replacing the old cells in the RFC stack and LIB. stacks should be replaced several times during this time For PEM-RFC, the estimated lifetime of the stack is 10 000 period. The replacement cost is included in the O&M cost. cycles, assuming that acceptable voltage degradation does We also assume flexible storage capacity for the not exceed 20% of the original reference voltage (based hydrogen ESS, which means that the RFC can produce on the values given in 38, 39). The useful lifetime of the hydrogen at all times when the PV supplies power to LIB is dependent on several factors, such as the operating the RFC system; thus, idle time for RFCs is theoretically Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 282 | Clean Energy, 2021, Vol. 5, No. 2 equal to zero (excluding scheduled maintenance and where TIC is the total installed system cost, r is the dis- unexpected interruptions). The roundtrip efficiency count rate (base value = 10%), N is the system lifetime in (η ) is another key parameter that plays a key role in years (base value = 20 years), kWh is the total discharge RTE dis determining the LCOS values. The PEM-RFC system is electricity from the system in one cycle, is the a P verage known for its low η , which is in the range of 28–36% cost of electricity used in charging the system, t is the RTE [23], whereas LIBs are known for their high η , which number of hours (max t  =  8760  h/year), η is the system RTE RTE can reach 85–95% in most cases [26, 40]. In this study, roundtrip efficiency and O&M cost is the estimated oper - we calculated the roundtrip efficiency of the RFC as 36% ation and maintenance cost ($0.02/kWh discharged from using the average efficiency values for PEM fuel cells and the ESS). PEM electrolysers from literature, and the corresponding roundtrip efficiency for the LIB system was assumed to be 90% based on the typical values in the literature [26, 2 Results and discussion 40]. It is important to highlight that the number of cycles 2.1 LCOE production from a PV solar system (charge/discharge) is determined based on the PV power As mentioned above, the building-energy system is de- profile, which determines the charging vs. discharging signed to have a self-sustained energy system that can modes for RFCs and charging/discharging/idle modes produce and store energy when the electricity supplied for LIBs. This factor plays a key role in shaping the final by the PV system exceeds the building demand and can LCOS values. convert it back to the building when needed in the night- Among the flexibilities that the RFC system provides is time or when the PV production cannot cover the building the ability to design the system to act as: (i) an ESS and demand. Fig. 5 summarizes the results obtained from the (ii) a hydrogen-production system to produce hydrogen SAM model for El Paso, TX. The nominal PV-system size for other end uses (e.g. as a fuel for fuel-cell electric ve- is 400  kW , which can produce ≤757  931 MWh in the hicles) [11, 41]. The latter design parameter is discussed DC first year. The peak electricity generation happens in the here as an indicative parameter to guide researchers who month of May in El Paso, TX and the lowest generation are interested in such dual-function systems. Equation (6) occurs in December. The size of the PV system was de- describes the status of the charging state (expressed in kg termined based on the building-energy profile and the of hydrogen stored in the hydrogen tanks): ability of the ESS to store excess energy so that curtailed S = S + H + S (6) t t−1 RFC 0 electricity from the PV system was at a minimum. The where S is the amount of stored hydrogen at time t (kg), profile of the electricity produced by the PV system is S is the amount of stored hydrogen at time t–1 (kg), H shown in Fig. 6 with a maximum electricity production t–1 RFC is the amount of hydrogen produced or consumed by the of 389 kWh. RFC system at time t (kg) and S is the strategic hydrogen- As shown in Fig. 5, the capacity factor of the PV system storage position (kg), which can be set to zero if the RFC is only 21.7%, which seems to be lower than the average system begins with empty tanks. capacity factor for wind turbines in Texas [42]. Even with this low capacity factor, solar energy still supplies low-cost electricity at a cost of 2.21 cents/kWh, which is comparable 1.4  LCOS to that of wind energy [42]. The LCOS is a quantitative metric that is used to evaluate the cost of storing and converting energy into a usable form using an ESS. As shown in Equation (7) below, the 2.2 LCOS in RFCs and LIBs LCOS is a function of several parameters, including capital As discussed previously, Equation (6) was used to calculate cost of the ESS, electricity price, number of cycles between the cumulative amount of hydrogen produced and con- charging/discharging modes, efficiency of the ESS and cap- sumed by the RFC system (Fig. 7). With a strategic starting acity of the ESS. The following formula was adapted from hydrogen-storage position of 250 kg-H at the beginning of [21] and then modified to calculate the LCOS for both LIBs the year, the RFC will not fall to a point at which there is and PEM-RFCs: no hydrogen in the hydrogen tanks—that is, this starting point will ensure that the cumulative hydrogen curve will r P TIC −N 1−(1+r) t=0 not fall below zero. The maximum amount of hydrogen LCOS = + + C(7) O & M t. η RTE stored in the hydrogen tanks reaches 1336  kg-H on 25 kWh 2 dis t=0 June. After that, the hydrogen level fluctuates around this value for several weeks and then starts to decline with time until it reaches 674  kg-H at the end of the year. Careful 2 2 For ESSs, roundtrip efficiency is defined as the ratio of the energy planning is needed here as to whether to sell part of this put in (charging mode) to the energy retrieved from storage in stored hydrogen for other end uses (e.g. for fuel-cell elec- the discharging mode. Inefficiencies include losses in the ESS tric vehicles) or to design the hydrogen-storage system to itself and the losses in transmitting and converting this energy from electricity to other electrochemical energy forms. accommodate all of the hydrogen produced in the whole Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 283 Metric Value Monthly Energy Production 80 000 Annual energy (year 1) 757,931 kwh Capacity factor (year 1) 21.7% Energy yield (year 1) 1,903 kWh/kW 70 000 Performance ratio (year 1) 0.79 Levelized COE (nominal) 2.77 ¢/kWh 60 000 Levelized COE (real) 2.21 ¢/kWh Electricity bill without system (year 1) $72,758 50 000 Electricity bill with system (year 1) $32,633 Net savings with system (year 1) $40,126 Net present value $155,734 40 000 Simple payback period 12.8 years Discounted payback period NaN 30 000 Net capital cost $651,074 Equity $0 20 000 Debt $651,074 10 000 Jan Feb Mar Apr MayJun JulAug Sep Oct NovDec Fig. 5: Summary report of the PV system from the SAM model PV [kWh] generated by the RFC system is 148.3 kWh on 3 August, which seems to be one of the hottest days in the year in El 400 Paso, TX, so that the energy demand in that day exceeds 350 what the PV can produce. Unlike the RFC system, the duty cycle in the LIB is a bit different. Charging and discharging do not always sum up to 100% because of the limited storage capacity of the LIB. Another issue with the LIB is the ability to supply electricity during peak times for several hours at a discharge rate that is suitable to meet the building-energy profile. For example, a hot summer in El Paso, TX may result in a building that 16-Dec 4-Feb 26-Mar 15-May 4-Jul 23-Aug 12-Oct 1-Dec –50 demands large amounts of electricity for several hours. To Fig. 6: Profile of the electricity generated by the PV system avoid such a case, the storage capacity of the LIB in kWh and the discharge rate (kWh/h) can be increased to accom- modate such a scenario, but this does not automatically Cumulative Hydrogen Production/Consumption eliminate the higher percentage in the idle state. The duty 1,600 1,400 cycles for a 400-kWh LIB for the month of April are shown 1,200 in Fig. 9a. The battery can store electricity until it reaches its 1,000 maximum storage capacity (nominal = 400 kWh). Similarly, the LIB is also limited in terms of the duration over which it can supply the power; e.g. a 400-kWh LIB with 100-kWh lowest DoD can supply 300 kWh during the discharge time. In any case, increasing the size of the LIB system can reduce 16-Dec 4-Feb 26-Mar 15-May 4-Jul 23-Aug 12-Oct 1-Dec the percentage of idle time but, on the other hand, this will Time/Date compromise the total system cost, and thus may increase the capital-cost component in Equation (7). For a larger LIB Fig. 7: Cumulative hydrogen production/consumption in the RFC system (starting H level = 250 kg) size (1000 kWh) in Fig. 9b, one can see that the idle time in April (also applies to other months in the year) can be re- year. The electricity generated by the RFC is shown in Fig. 8 , duced substantially and, on the other hand, the discharge with the total of 170  964 kWh, which contributes ~36.8% time will increase substantially as well. Fig. 10 shows the of the total annual building demand. The peak electricity percentages of time for each operating mode as a function Electricity Generated by PV (kWh) kg-H2 kWh Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 284 | Clean Energy, 2021, Vol. 5, No. 2 kWh from RFC of the LIB size in kWh. The percentages of the charging and discharging modes can be increased from 8.4% and 42.5%, respectively, for a 400-kWh battery to 15.1% and 61.8% for charging and discharging times, respectively, for a 2000- kWh LIB system. The biggest advantage here is the ability to lower the time that the LIB spends in idle mode, which can be reduced from 49.1% for a 400-kWh LIB to 23.1% for a 2000-kWh LIB system. The effect of changing the LIB size on the LCOS and the expected amount of kWh discharged from the LIB during 16-Dec 4-Feb 26-Mar 15-May4-Jul 23-Aug 12-Oct 1-Dec the whole year are shown in Fig. 11. The LCOS increases from 25.5 cents/kWh for a 400-kWh LIB to 65.3cents/kWh Fig. 8: Electricity generated by the RFC system for a 2000-kWh LIB. The main reason behind such an in- crease is that the change in the numerator in the capital- cost component (the first term after the equals sign in Equation (7)) is increasing much more than the expected increase in the denominator, which implies that such a so- lution may not be so competitive from a cost perspective. In terms of the amount of electricity that the LIB system can 200 supply to the building when the PV system is off (night- 150 time) or when it cannot supply the whole amount of the building need at certain times, we can see that a larger LIB system can supply more electricity to the building. A 400- kWh LIB system can supply 97  905 kWh/year (equivalent 31-Mar 10-Apr 20-Apr 30-Apr to 57.3% of the required electricity to be supplied by the Date ESS) and a 2000-kWh LIB system can supply ≤168  832 1,200 kWh/year (98.8% of the required electricity from the ESS). 1,000 However, energy engineers should also weigh up the cost of connecting the building to the grid and whether they want to design the building to be a stand-alone 100% green building or design the building to rely on both a renewable- 400 energy source and the power grid when the demand ex- ceeds the total supply from the PV and LIB systems. Such a decision can be extended to include other selection cri- teria and incentives offered to green buildings, which are 31-Mar 10-Apr 20-Apr 30-Apr beyond the scope of this paper. Date Thus far, this model was developed to evaluate the Fig. 9: State of charge of the LIB in April for: (a) a 400-kWh LIB (min- economics and the resiliency of the ESS. It should be imum value = lowest depth of discharge = 100 kWh) (note: zero values highlighted that other metrics can be used in similar com- represent idle states); and (b) a 1000-kWh LIB (minimum value = lowest parisons for energy-efficient buildings. In their compara- depth of discharge = 250 kWh) tive study, Harkouss et al. [ 43] used five metrics to assess the 0.70 180,000 70% 160,000 0.60 60% 140,000 0.50 50% 120,000 0.40 40% 100,000 80,000 30% 0.30 60,000 20% 0.20 40,000 LCOS 10% 0.10 20,000 KWh Discharged 0% 400 600 800 1000 1200 1400 1600 1800 2000 - - 400 600 800 1000 1200 1400 1600 1800 2000 LIB Size (kWh) LIB Size (kWh) % Charge% Discharge % Idle Fig. 11: Effect of changing the LIB size on the LCOS in the LIB storage Fig. 10: Effect of changing the LIB size (in kWh) on the percentage of system (note: the total amount of the electricity that is not directly the time in the operating modes (charge, discharge and  idle) Lowest depth of discharge = 0.25 of the nominal LIB size in kWh. covered by the PV system = 170 964 kWh/year) Percent of time kWh in LIB kWh from RFC kWh in LIB LCOS ($/kWh) kWh Discharged Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 285 performance of the net-zero-energy buildings. These metrics sensitivity-analysis purposes: capital cost and roundtrip include (i) energy-efficiency indicator (total energy consump- efficiency as the key system-design parameters, electricity tion); (ii) energy balance and self-sufficiency indicator, which cost and O&M as the key operating-cost parameters, and is measured using a load matching index; (iii) grid-stress in- lifetime and discount rate as the most important financial dicator using a grid-interaction index; (iv) economic indica- parameters. Tornado charts in Fig. 12a–c show the effect of tors using life-cycle cost, LCOE and simple payback period; changing such parameters on the LCOS in 1000-kWh LIB, and (v) environment indicator using CO-equivalent emis- 2000-kWh LIB and 450 kW RFC systems (equivalent dis- sions. We believe that these performance indicators can be charge rate = 162 kWh/h), respectively. used in a more holistic approach in the future to evaluate the It seems that the LCOS models for both LIB and RFC tech- performance of energy-efficient buildings and different ESSs. nologies are more sensitive to capital cost followed by dis- count rate and expected system lifetime. Operating costs (electricity cost and O&M cost) seem to have a lower effect 2.3 Sensitivity analysis on the final LCOS in both LIBs and RFCs. Another important Sensitivity analysis was performed to assess the effect remark for the LIB system is the impact of changing the of changing some financial and system-design param- nominal storage size in kWh, which will automatically in- eters on the LCOS. Based on the LCOS formula in Equation crease the value of the supplied energy by the system (de- (7), the following parameters were selected for the nominator in the capital-cost component in Equation (7)), which in turn does not seem to outweigh the substantial in- crease in the capital cost in the numerator; thus, increasing LCOS-2,000 KWH LIB (NOMINAL=65.3 CENT/KWH) the LIB size will probably increase the LCOS. Capital Cost Discount Rate 3 Conclusions Lifetime Green-energy buildings that totally or partially rely on renewable-energy sources such as solar or wind energies Rondtrip Efficiency need cost-effective ESSs to store energy during energy- Electricity Price production periods and the ability to reverse it during non- +50% -%50 O&M Cost production times. RFCs and LIBs were studied in this paper to evaluate their ability to increase green-building resili- –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 Change in LCOS ($/KWH) ency and reduce the cost of energy storage. Results confirm B the attractiveness of both technologies as electricity- LCOS-1,000 KWH LIB (NOMINAL=32.1 CENT/KWH) storage systems that can be connected to commercial Capital Cost buildings. Cost results show that a 400-kW PV system can generate electricity at a cost of 2.21 cents/kWh in El Paso, Discount Rate TX. In terms of energy storage, we found that the average Lifetime LCOS using a 450-kW RFC is ~31.3 cents/kWh, while this could reach as low as 25.5 cents/kWh using a small LIB Rondtrip Efficiency ESS (400-kWh). However, while the RFC provides the flexi- Electricity Price bility required to meet building-energy profiles, LIBs may +50% not be able to meet building needs unless the storage -%50 O&M Cost size is increased substantially, which in turn incurs more –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 Change in LCOS ($/KWH) energy-storage costs, making LIBs less favourable from an economic perspective. Given the nature of the energy pro- LCOS-RFC (NOMINAL=31.3 CENT/KWH) duction and energy consumption of this selected commer - Capital Cost cial building, it seems that LIBs provide a better ESS at a Discount Rate lower cost in $/kWh and can be designed to meet building needs even at peak times. A hydrogen-based ESS is still a Lifetime viable solution for such buildings, but may need to be con- Rondtrip Efficiency sidered when the goal is to store energy for longer periods or when it is are connected to a nearby hydrogen station, Electricity Price where it can be used primarily as a hydrogen-production +50% O&M Cost -%50 system and as an ESS as a secondary goal. Parametric sen- –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 sitivity analysis of the stand-alone PV-ESS showed that Change in LCOS ($/KWH) the LCOSs of the LIB and RFC systems are sensitive to system capital cost, financial parameters (discount rate Fig. 12: Tornado charts showing the effect of changing certain param- and system lifetime) and, to a lesser extent, the roundtrip eters on the LCOS for: (a) a 1000-kWh LIB, (b) a 2000-kWh LIB and (c) a 450-kW (162 kWh/h) RFC system efficiency and the generated-electricity cost. Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 286 | Clean Energy, 2021, Vol. 5, No. 2 in Thermal Energy Storage Systems. Cambridge: Woodhead Acknowledgement Publishing, 2015, 1–28. The authors gratefully acknowledge the support and funding from [17] Mostafa MH, Abdel Aleem SHE, Ali SGet  , al. Techno-economic Khalifa University - Grant number 8474000238/FSU-2020-14 to con- assessment of energy storage systems using annualized life duct this research. cycle cost of storage (LCCOS) and levelized cost of energy (LCOE) metrics. Journal of Energy Storage, 2020, 29:101345. [18] Nikolaidis P, Poullikkas A. A comparative review of electrical energy storage systems for better sustainability. Journal of Conflict of Interest Power Technologies, 2017, 97:220–245. None declared. [19] Dunn B, Kamath H, Tarascon JM. Electrical energy storage for the grid: a battery of choices. Science, 2011, 334:928–935. [20] Argyrou MC, Christodoulides P, Kalogirou SA. Energy storage References for electricity generation and related processes: technolo- [1] EIA (Energy Information Administration). US Energy Facts gies appraisal and grid scale applications. Renewable and Explained. 2019. https://www.eia.gov/energyexplained/ Sustainable Energy Reviews, 2018, 94:804–21. us-energy-facts/ (20 March 2021, date last accessed). [21] Reznicek  E, Braun  RJ. Techno-economic and off-design ana- [2] Li G, Xuan Q, Akram MW, et al. Building integrated solar con- lysis of stand-alone, distributed-scale reversible solid oxide centrating systems: a review. Applied Energy, 2020, 260:114288. cell energy storage systems. Energy Conversion and Management, [3] The Directive 2010/31/EU of the European Parliament and of the 2018, 175:263–277. Council of 19 May 2010 on the energy performance of buildings, [22] Mostafa MH, Aleem SHA, Ali SG,et  al. Robust energy manage- Official Journal of the European Union. https://eur-lex.europa. ment and economic analysis of microgrids considering dif- eu/LexUriServ/LexUriServ.do?uri=OJ:L:2010:153:0013:0035:en: ferent battery characteristics. IEEE Access, 2020, 8:54751–54775. PDF (20 March 2021, date last accessed). [23] Hammond GP, Hazeldine T. Indicative energy technology as- [4] Aeleneia  L, Gonçalves  H. From solar building design to net sessment of advanced rechargeable batteries. Applied Energy, zero energy buildings: performance insights of an office 2015, 138:559–571. building. Energy Procedia, 2014, 48:1236–1243. [24] Ito H, Miyazaki N, Ishida M, et al. Efficiency of unitized revers- [5] Balaban O, de Oliveira JAP. Sustainable buildings for healthier ible fuel cell systems. International Journal of Hydrogen Energy, cities: assessing the co-benefits of green buildings in Japan. 2016, 41:5803–5815. Journal of Cleaner Production, 2017, 163:S68–S78. [25] Hur  W, Moon  Y, Shin  K, et  al. Economic value of Li-ion en- [6] Allen JG,et  al. Green buildings and health. Current Environmental ergy storage system in frequency regulation application from Health Reports, 2015, 2:250–258. utility firm’s perspective in Korea. Energies, 2015, 8:5000–5017. [7] Bakos GC, Soursos M, Tsagas NF. Technoeconomic assessment [26] IRENA. Battery storage technology improvements and cost of a building-integrated PV system for electrical energy saving reductions to 2030: a deep dive. In: International Renewable in residential sector. Energy and Buildings, 2003, 35:757–762. Energy Agency Workshop Dusseldorf, Kai-Philipp Kairies, [8] Tribioli L, Cozzolino R. Techno-economic analysis of a stand- ISEA/RWTH Aachen, 17 March 2017. https://www.irena. alone microgrid for a commercial building in eight different org/-/media/Files/IRENA/Agency/Events/2017/Mar/15/2017_ climate zones. Energy Conversion and Management, 2019, Kairies_Battery_Cost_and_Performance_01.pdf (20 March 1:58–71. 2021, date last accessed). [9] Chatzivasileiadi  A, Ampatzi  E, Knight  I. Characteristics of [27] EnergyPlus. https://energyplus.net/ (10 May 2021, date last electrical energy storage technologies and their applications accessed). in buildings. Renewable and Sustainable Energy Reviews, 2013, [28] Pezzutto S, Croce S, Zambotti S, et al. Assessment of the space 25:814–830. heating and domestic hot water market in Europe—open data [10] Lucchese FC, Canha LN, Brignol WS,et  al. Energy storage sys- and results. Energies, 2019, 12:1760. tems role in supporting renewable resources: global overview. [29] SAM (System Advisor Model, 2020.2.29). National Renewable In: 2019 54th International Universities Power Engineering Energy Laboratory. https://sam.nrel.gov/ (10 May 2021, date last Conference (UPEC). IEEE. Bucharest, Romania, 3–6 September accessed). 2019, 1–6. doi: 10.1109/UPEC.2019.8893470. [30] Mayyas A, Wei M, Chan SH, et al. Fuel cell forklift deployment [11] Mayyas A, Wei M, Levis G. Hydrogen as a long-term, large-scale in the USA. In: Stolten  D, Samsun  RC, Garland  N (eds). Fuels energy storage solution when coupled with renewable energy Cells: Data, Facts and Figures. Weinheim: Wiley-VCH, 2016, sources or grids with dynamic electricity pricing schemes. 334–342. International Journal of Hydrogen Energy, 2020, 45:16311–16325. [31] Wei  M, Chan  SH, Mayyas  A, et  al. Deployment and cap- [12] Kharel S, Shabani B. Hydrogen as a long-term large-scale energy acity trends for stationary fuel cell systems in the USA. In: storage solution to support renewables. Energies, 2018, 11:2825. Stolten D, Samsun RC, Garland N (eds). Fuels Cells: Data, Facts [13] Guney  MS, Tepe  Y. Classification and assessment of energy and Figures. Weinheim: Wiley-VCH, 2016, 257–269. storage systems. Renewable and Sustainable Energy Reviews, [32] Ayers KE, Capuano CB, Anderson EB. Recent advances in cell 2017, 75:1187–1197. cost and efficiency for PEM-based water electrolysis. ECS [14] Chen H, Cong TN, Yang W, et  al. Progress in electrical energy Transactions, 2012, 41:15–22. storage system: a critical review. Progress in Natural Science, [33] Mayyas  AT, Ruth  MF, Pivovar  BS, et  al. Manufacturing Cost 2009, 19:291–312. Analysis for Proton Exchange Membrane Water Electrolyzers (No. [15] Wagner L. Overview of Energy Storage Methods, Research Report, NREL/TP-6A20-72740). Golden, CO: National Renewable Energy 2007. http://www.moraassociates.com/reports/0712%20 Laboratory (NREL), 2019. Energy%20storage.pdf (20 March 2021, date last accessed). [34] Carmo M, Fritz D, Maier W, et al. Alkaline water electrolysis vs. [16] Cabeza  LF, Martorell  I, Miró  L, et  al. Introduction to thermal PEM water electrolysis—exploring their full performance. In: energy storage (TES) systems. In: Cabeza  LF (ed). Advances 227th ECS Meeting. Chicago, 24–28 May 2015. Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 287 [35] Schmidt  O, Gambhir  A, Staffell  I, et  al. Future cost and [44] DOE (Department of Energy). DOE Technical performance of water electrolysis: an expert elicit- Targets for Hydrogen Production from Electrolysis. ation study. International Journal of Hydrogen Energy, 2017, https://www .ener gy .go v/eer e/fuelcells/ 42:30470–30492. doe-technical-targets-hydrogen-production-electrolysis [36] Wei M, Mayyas A, Lipman TE, et al. Fuel cell systems: total cost [45] DOE (Department of Energy). DOE Technical Targets for Fuel Cell of ownership. In: Lipman  TE, Weber  AZ (eds). Fuel Cells and Systems for Stationary (Combined Heat and Power) Applications. Hydrogen Production, A Volume in the Encyclopedia of Sustainability https://www.energy.gov/eere/fuelcells/doe-technical-targets- Science and Technology, Second Edition. New York City: Springer fuel-cell-systems-stationary-combined-heat-and-power (20 Nature, 2019, 27–81. March 2021, date last accessed). [37] IRENA. Utility-scale Batteries Innovation Landscape Brief, [46] Mayyas  A, Mann  M. Manufacturing competitiveness ana- 2019. https://www.irena.org/-/media/Files/IRENA/Agency/ lysis for hydrogen refueling stations. International Journal of Publication/2019/Sep/IRENA_Utility-scale-batteries_2019.pdf Hydrogen Energy, 2019, 44:9121–9142. (10 May 2021, date last accessed). [47] Ahluwalia RK, Papadias DD, Peng JK, et  al. System level ana- [38] Regmi  YN, Peng  X, Fornaciari  JC, et  al. A low tempera- lysis of hydrogen storage options. In: U.S. DOE Hydrogen and ture unitized regenerative fuel cell realizing 60% round Fuel Cells Program 2019 Annual Merit Review and Peer Evaluation trip efficiency and 10000 cycles of durability for energy Meeting, Washington, DC, 29 April–1 May 2019. https://www. storage applications. Energy & Environmental Science, 2020, hydrogen.energy.gov/pdfs/review19/st001_ahluwalia_2019_o. 13:2096–2105. pdf (10 May 2021, date last accessed). [39] Wu J, Yuan XZ, Martin JJ, et  al. A review of PEM fuel cell dur - [48] Bruijn FD. PEMFC Lifetime and Durability: An Overview. ability: degradation mechanisms and mitigation strategies. Thessaloniki, 21 September 2011. https://www.sintef.no/ Journal of Power Sources 2008, 184:104–119. globalassets/upload/materialer_kjemi/energikonvertering- [40] Dufo-López  R, Bernal-Agustín  JL. Techno-economic ana- og-materialer/bilder/pemfc-overview-thessaloniki-fa- lysis of grid-connected battery storage. Energy Conversion and debruijn---f.pdf (20 March 2021, date last accessed). Management, 2015, 91:394–404. [49] Tsiropoulos I, Tarvydas D, Lebedeva N. Li-Ion Batteries [41] Nguyen VN, Blum L. Reversible fuel cells. In: Barbir F, Basile A, for Mobility and Stationary Storage Applications. European Veziroğlu  TN (eds). Compendium of Hydrogen Energy, Volume 3: Union’s Joint Research Center (JRC). EUR 29440 EN, 2018. Hydrogen Energy Conversion. Cambridge: Woodhead Publishing, https://ec.europa.eu/jrc/en/publication/eur-scientific-and- 2016, 115–145. technical-research-reports/Li-ion-batteries-mobility-and- [42] U.S. Energy Information Administration. Wind Is a Growing Part of stationary-storage-applications (20 March 2021, date last the Electricity Mix in Texas. 15 October 2020. https://www.eia.gov/ accessed). todayinenergy/detail.php?id=45476#:~:text=In%20April%20 [50] Larsson P, Börjesson, P. Cost Models for Battery Energy 2019%2C%20wind%20generators,United%20States%20as%20 Storage Systems. KTH School of Industrial Engineering a%20whole. (22 March 2021, date last accessed). and Management. Energy Technology EGI-2018, TRITA- [43] Harkouss  F, Fardoun  F, Biwole  PH. Optimal design of renew- ITM-EX 2018:428. http://www.diva-portal.org/smash/get/ able energy solution sets for net zero energy buildings. Ener gy, diva2:1254196/FULLTEXT01.pdf (20 March 2021, date last 2019, 179:1155–1175. accessed). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clean Energy Oxford University Press

Techno-economic analysis of the Li-ion batteries and reversible fuel cells as energy-storage systems used in green and energy-efficient buildings

Loading next page...
 
/lp/oxford-university-press/techno-economic-analysis-of-the-li-ion-batteries-and-reversible-fuel-0c790h0CMR
Publisher
Oxford University Press
Copyright
Copyright © 2021 National Institute of Clean-and-Low-Carbon Energy
ISSN
2515-4230
eISSN
2515-396X
DOI
10.1093/ce/zkab009
Publisher site
See Article on Publisher Site

Abstract

PV Modules Combustible Gas Detector DC input One-way Valve DC/DC Convertor Dryer Hydrogen Storage DC/AC Inverter Back Pressure Stack Regulator Oxygen Pump Vent 80-90°C Controllable City Water Water Valve Water Cleaner 20°C Expander Air Compressor FC Mode EC Mode Keywords: green-building energy-storage systems; fuel cell; hydrogen; Li-ion batteries; reversible fuel cells energy-efficient through design modifications if connected Introduction to the power grid. The most popular renewable-energy sys- The industrial, residential and commercial sectors were tems for commercial buildings are either in stand-alone responsible for 35%, 16% and 12% of the total energy con- configuration or connected to the grid that is fed by a sumption in the USA in 2019 [1]. Globally, energy con- large portion of renewable energy, photovoltaic (PV) solar sumption of the building sector accounts for ~30% of the and wind systems. For the stand-alone configuration, it world’s total energy demand, which includes electrical seems that PV panels are more practical to use and easier power, heating and cooling loads [2]. Green buildings and to install. PV panels can be installed on the surface of the retrofit energy-efficient buildings have become a trend building, which allows combining electrical-energy pro- in building-energy science research in the last few years. duction with other functions of the building structures [7]. The performance of net-zero-energy buildings has gained Integrating PV panels with the building façade has also be- more attention since the publication in 2010 of the EU come a popular choice for many modern building designs Energy Performance of Buildings Directive recast [3]. In [4]. Recent studies also recommend using concentrated the USA, the federal government and many state govern- solar panels (CSP), which can decrease the area needed to ments have promoted ‘marketable zero energy homes in install PV cells by using low-cost transparent material in 2020 and commercial zero energy buildings in 2025’ or smaller areas, thus decreasing the overall system cost and similar programmes [4]. Japan also attempts to promote increasing the temperature of the heat source. This fact ‘carbon-neutralized buildings’, including existing build- also implies that CSP can perform better in several func- ings, by 2050 [5]. Thus, green-building design is becoming tions such as in heating, refrigeration, dehumidification broadly adopted in commercial and residential sectors. By and lighting for such buildings [2]. definition, green buildings are designed to minimize the Alongside the extensive research on the new de- impacts on the environment by applying techniques to sign and required policies for green and energy-efficient reduce energy usage and water usage, and by minimizing buildings, many researchers have also analysed the the environmental disturbances from the building site [6]. cost of energy production, consumption and storage in Green buildings also aim to improve human health and these energy-efficient buildings. In their comprehen- workplace environments through the design of healthier sive review of the energy-storage systems (ESSs) used in indoor environments [6]. energy-efficient buildings, Chatzivasileiadi et al. [8] found Green buildings can be 100% self-sufficient if connected that battery storage systems (including Li-ion, Zn-air and to a suitably sized renewable-energy system or can be more Compressor Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 275 NaNiCl batteries) are among the most promising and eco- last type of ESS technology is the thermal energy-storage nomic ESSs given the short-term storage needs for ESSs (TES) system, which is capable of storing heat or cold in connected to commercial buildings. In another study a storage medium at certain temperatures for further made by Tribioli and Gonzolino [9 ], the authors modelled usage, under different conditions such as temperature, and discussed the economics of energy production and place or power [16]. TES systems are generally classified storage using batteries and reversible fuel cells (RFCs). into three different categories, depending on the thermo- In this paper, the authors modelled a stand-alone poly- dynamic properties of the storage medium, which include generation power plant for a strip mall located in eight sensible heat, latent heat, absorption and adsorption different climate zones in the USA (Minneapolis, MN; system. TES systems are commonly used as ESSs for in- Houston, TX; Las Vegas, NV; Los Angeles, CA; Miami, FL; dustrial and residential purposes, such as space heating New York, NY; Denver, CO; and Seattle, WA). In their model, or cooling, process heating and cooling, hot-water produc- they coupled a photovoltaic panel array to a battery and a tion and circulation in buildings, and applications that re- unitized regenerative polymer electrolyte membrane fuel quire phase-change material [13–15]. A  great comparison cell as primary storage and a diesel generator as a sec- of different ESSs was made by Mostafa et  al. [17]. In that ondary backup system. Their cost model indicates that paper, the authors studied several ESSs for short-term and this poly-generation system can produce energy at a long-term storage needs and found that battery storage levelized cost as low as 8.6 cents/kWh in Seattle, WA and systems like Li-ion, acid, nickel-cadmium (NiCd), sodium- ≤42.6 cents/kWh in Denver, CO. Interestingly, this study sulphur (NaS), vanadium redox and zinc-bromine (ZnBr) found that the adoption of this poly-generation system are more suitable for short- and medium-term storage ap- can mitigate between 61.23% and 99.60% of the carbon- plications, while compressed air, hydrogen-based energy dioxide emissions. storage and pumped hydro are more suitable for medium- Among the different challenges that are still facing to long-term storage applications. They also found that the green buildings are the high initial cost, energy efficiency annualized life-cycle cost of storage (LCCOS) in $/kWh- of the installed systems and intermittent nature of the year for long-term storage systems such as underground electricity generated by the renewable-energy systems. compressed air, aboveground compressed air and pumped Having an efficient ESS is still considered a big challenge hydro tend to decrease with a system lifetime of ≤40 years. that engineers need to carefully plan for. The authors found that, unlike long-term storage systems, ESSs can be classified into five main classes: chem- short- and medium-term storage systems such as battery ical, electrochemical, electrical, mechanical and thermal.stor age technologies and hydrogen-based energy storage Chemical ESSs store energy in the chemical bonds of can see a decrease in the LCCOS in the first 20 years, then atoms and molecules, which then can be released in an increase in the LCCOS because the variable operation chemical reactions to recover the stored energy. After the and maintenance costs for medium-term ESSs are less release of chemical energy, the substance is often changed than the corresponding costs calculated for long-term into an entirely different substance [10]. The stored chem- ESSs, which can be directly linked to the fact that the dis- ical energy can be released via electron-transfer reac- charge periods of short- and medium-term ESSs are lower tions that produce electricity. Chemical energy storage than the discharge periods of long-term ESSs. focuses on the production of hydrogen, ammonia and Among different ESSs, it seems that rechargeable bat- synthetic natural gas as secondary energy carriers [11, tery technologies such as lead-acid, NaS and Li-ion bat- 12]. Electrochemical ESSs convert chemical energy into teries (LIBs) are common ESS technologies, especially for electrical energy. Electrochemical cells can be cate -gor small-scale stationary energy-storage applications [18], ized into four types, depending on their function: primary but still possess some limitations that limit their wide- cells or batteries, secondary cells or secondary batteries, spread deployment such as having shorter life compared reserve cells and fuel cells [13]. The third type of ESS tech- to other battery energy storage and other ESS technologies, nology is electrical ESSs, which also can be grouped into low energy density and, to a lesser extent, high cost in dol- two subgroups: electrostatic systems, including capaci- lars/kWh of storage capacity, especially for Li-ion and NaS tors and supercapacitors; and magnetic/current ESSs [14]. technologies [19, 20]. New battery technologies (e.g. redox The fourth ESS technology is mechanical energy-storage flow and zinc-iron batteries) are promising ESS technolo- systems (MESSs), which are usually classified based on gies, but do not come without limitations. Such limitations their working principles. Examples of MESSs include sys- include their high initial cost and low energy density [21]. tems that store energy based on the forces associated with In their paper, Mostafa et  al. [22] developed a techno- pressurized gas, forced springs, kinetic energy and poten- economic model and optimal energy management for tial energy [15]. The most useful advantage of mechanical a grid-connected micro grid that was dependent on the ESSs is that they can readily deliver the energy whenever renewable-energy resources and different battery storage required in a short period of time, besides being adapt- technologies including LA, NiCd, Li-ion and NaS. They able [13]. Flywheels, pumped hydro storage systems and studied the effect of certain design and performance compressed-air ESSs are also classified as MESSs [13]. The parameters such as initial charge, depth of discharge and Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 276 | Clean Energy, 2021, Vol. 5, No. 2 the number of charging/discharging cycles with the goal The selection of the PEM-RFC and LIB systems is based of minimizing the total operating cost of the system and on their flexibility to operate as load-following systems maximizing the benefits of battery storage systems. Other and because of their short response times to the changing goals also included minimizing the investment and re- loads and ability to work on partial loads [24–26]. Many placement costs of a battery storage system, and minim- published techno-economic analyses have been focusing izing the operation and maintenance costs of a distributed on the relative comparison between ESSs from economic generation system. Mostafa and co-authors used a gen- and performance perspectives and their useful applica- eral algebraic modelling system to solve the deterministic tion in the short or long term. In this paper, we attempt to optimization problem; stochastic optimization was also compare two competing technologies that are suitable for used to model the market-price uncertainty and gap deci- smaller-scale applications (<1 MW), which allow them to sion theory to model the electric-load uncertainty. Among be connected in a stand-alone configuration to commer - many interesting findings, this group of researchers found cial or industrial buildings. This paper discusses the re- that the total cost per day of the battery storage system sults of the techno-economic model that we developed for through the lifetime of the project decreased when the a PEM-RFC system that is designed to have dual functions: value of the depth of discharge (DoD) increased and that (i) an ESS to produce hydrogen in the electrolysis mode NaS batteries had better potential in reducing the total and power in the fuel-cell mode and (ii) as a hydrogen- operating cost of microgrids. production system to use excess produced hydrogen as a Hydrogen has been investigated as a medium for fuel for electric fuel-cell vehicles or other end uses. Then, long-term energy-storage applications and as a tech- the same model was used to assess and compare LIBs nology to improve the resiliency of power grids [ ].8 The with different sizes and to evaluate their relative perform- advancements in fuel-cell technologies have encouraged ance with RFCs. Parametric sensitivity analysis aims at as- researchers to study RFCs as a viable ESS to store energy sessing certain designs and financial parameters and to for medium to long periods of time (months to years). study their effects on the LCOS. Users can use this ESS as a flexible system that can pro- The rest of the paper is organized as follows: Section 1 duce and store hydrogen fuel for several months or years discusses the methodology used in sizing the PV system without losing the nominal storage capacity. Rechargeable and ESS using the building-energy profile, the key param- batteries, on the other hand, are known for their gradual eters used in developing the techno-economic models for losses of stored energy through self-discharging if left LIB and RFC, and the development of the mathematical without use for longer periods of time [23]. Unitized RFCs formula to calculate the LCOS. Section 2 includes the re- use bifunctional anode/cathode electrodes, allowing the sults and discussion of the LCOS models, and includes a RFC to operate in both electrolysis and fuel-cell modes discussion on the sizing of the hydrogen storage for RFCs using the same stack [11]. This configuration is similar to based on the building demand and supply curves; com- the rechargeable batteries that operate under both c-har parison between LIBs and RFCs, and sensitivity analysis ging and discharging conditions. In electrolysis mode, the are presented. Section 3 concludes with a summary of the RFC takes in electricity and water and produces hydrogen high-level findings and insights from this work. as a fuel and oxygen as a by-product. In fuel-cell mode, the RFC utilizes hydrogen and oxygen (from air), and pro- 1 Methodology duces electricity and water as by-products. RFCs have a wide range of applications that range from energy- In this section, we discuss the electricity profile for com- storage devices coupled to renewable-energy sources and mercial buildings in Section 1.1, followed by discussion of intermittent power grids, backup power plants and in the design of the PV solar system and its technical param- spacecrafts [11]. eters in Section 1.2. Working principles and logic of the two It seems that the attractiveness of promising storage selected ESSs (RFC and LIB) and the technical and financial technologies, such as reversible proton-exchange mem- factors used in developing the techno-economic model are brane (PEM) fuel cells, which can compete with recharge- all discussed in Section 1.3. Finally, the mathematical for - able stationary LIBs, has not been thoroughly explored mulas and the underlying assumptions used in developing in the literature. Thus, this paper aims to fill this gap by the LCOS are presented in Section 1.4. The flow diagram evaluating the flexibility and economics of energy storage in Fig. 1 shows the sequence of steps and analyses con- using LIBs as a mature ESS technology and RFCs as an ducted in this study. The first step was the analysis of the emergent ESS technology. A  techno-economic analysis building-energy profiles using an average year from the TM is conducted for this purpose to assess the performance EnergyPlus program [27], then we used building-energy and economics of energy storage using LIBs and unitized- profiles to estimate the PV solar-system size in kW. The PV stack reversible PEM fuel cells (PEM-RFC). The effects of system was designed to ensure a self-sufficient building system design and some selected financial parameters on in a stand-alone configuration—that is, this building is not the levelized cost of energy storage (LCOS) are particularly connected to any power grids other than its own PV solar analysed and discussed in this paper. power system. The economics of energy production from Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 277 Analysis of Building Energy Demand and Supply PV Solar System (Design and Economics (LCOE)) Energy Storage System – RFC Energy Storage System – LIB (System Design and Economics) (System Design and Economics) Hydrogen Energy Storage Dynamics – LIB Production/Consumption (Charging/Discharging Cycles) Sensitivity Analysis - RFC Sensitivity Analysis - LIB (Design Parameters; Financial Factors) (Design Parameters; Financial Factors) Fig. 1: Flow diagram showing the approach and analysis used in this study El Paso, TX [kWh] the PV system were estimated using the levelized cost of energy (LCOE) to estimate the cost component of electricity in the energy-storage equation (Equation (1)). The LCOS was used as the main metric to assess the studied ESSs (LIB 140 and RFC) from an economic perspective, while resiliency was estimated using the ability of the system to provide all required building demands without stopping. Finally, sen- sitivity analysis was conducted to study the effect of the most important factors that play key roles in determining the final value of the energy storage. These factors include two sets: the first set includes system-design parameters (ESS capital cost, operating and maintenance (O&M) cost, 0 16-Dec 4-Feb 26-Mar 15-May4-July 23-Aug 12-Oct 1-Dec electricity cost and roundtrip efficiency), while the second set includes discount rate and system lifetime. Fig. 2: Energy profiles for a medium office building in El Paso, TX simulation program that is used to model both energy 1.1 Commercial-building electricity profiles consumption in heating, cooling, ventilation, lighting and plug and process loads, and water use in buildings. In The building-performance simulations are used to identify this analysis, we use the energy profile for medium office each energy end use, which includes space heating, space buildings in El Paso, TX. According to the EnergyPlus def- cooling, domestic hot water, and appliances and lighting. initions, the medium office building has an area of 4982 In many cases, electricity may be used to meet  all of the 2 2 1 m (53 628 ft ) and consists of three storeys . Fig. 2 shows end uses, especially for those buildings that are not con- sample energy profiles for a medium office building in El nected to the natural-gas distribution network. Nearly Paso, TX. A  quick look at this energy profile shows that all space-cooling, appliance and lighting needs are pro- energy consumption increases in the summer months in vided by electricity; however, space heating and domestic El Paso, TX due to space-cooling requirements. The min- hot water can be met with other energy sources, such as imum, mean and maximum electricity consumptions in natural gas, heating oil and wood, or via central district the medium office building in El Paso, TX are 11.4, 53.0 and heating systems that supply hot water to several buildings 164.2 kWh, respectively. using a central heating source [28]. The electricity profiles used in this work were pulled from the EnergyPlus platform provided from the US https://www.energy.gov/eere/buildings/commercial-reference- buildings. Department of Energy [27]. EnergyPlus is a building-energy Electricty (kWh) Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 278 | Clean Energy, 2021, Vol. 5, No. 2 positioned itself as one of the most competitive hydrogen- 1.2 PV-system design production and energy-storage technologies because of The design parameters for the PV system were deter - the high energy-conversion efficiency and low operating mined based on the building load profile with the goal temperatures (<120°C) [3233 , ]. The high capital cost, which of having a stand-alone self-sustained system that can is partially attributed to the use of precious platinum- supply enough electricity to the building during the day- group metals (PGMs) and expensive selective PEMs in the time and can divert excess energy during the daytime electrode assemblies, is one of the biggest limitations to the ESS to be stored and used later when there is no that limit the adoption of PEM technology relative to the sunshine. Based on the load profiles, the estimated nom- other electrolysis/fuel-cell technologies such as alkaline inal PV-system size was set to 400 kW . The selected PV DC electrolysers and solid-oxide fuel cells [33]. In the USA, module was SunPower SPR-E19-310-COM, which has a Proton (NEL), Giner and other fuel-cell manufacturers nominal efficiency of 19.02% and can provide a maximum have started to produce megawatt-scale PEM electrolysers power of 310.15 W . The DC/AC inverter is SMA America: DC as a means to reduce the manufacturing cost. In addition, SC750CP-US, which has an average DC-to-AC conversion many R&D institutes are also working on improving the efficiency of 97.59%. Other cost and design parameters conversion efficiency of the PEM electrolysers—another are summarized in Table 1. The System Advisor Model factor that has historically limited the adoption of this (SAM 2020.2.29) from the Department of Energy’s National technology. Today, PEM electrolysers have achieved an im- Renewable Energy Laboratory (NREL) was used to model portant milestone in terms of the higher current densities the PV system. The SAM is a free techno-economic soft- (~5× that of the alkaline electrolysers) and higher effi- ware model that is used to model the central large-scale ciencies (nearly 6% higher than that value for alkaline and distributed small-scale renewable-energy systems electrolysers) [3435 , ]. and facilitates the decision-making process for people The unitized-stack reversible fuel cell (URFC) is a new working in the renewable-energy industry [29]. energy-storage technology that utilizes the same stack to perform dual functions: hydrogen production (electrolysis 1.3 ESS design and functional parameters mode) and electricity generation (fuel-cell mode). The overall reaction mechanism inside the URFC is as follows: Proton-exchange membrane fuel cells (PEMFCs) and + − electrolysers (PEMECs) are the most favourable fuel-cell (1) In the fuel cell mode : 4H + 4e + O → 2H O 2 2 and electrolysis technologies when the applications re- + − In the electrolysis mode : 2H O → 4H + O + 4e(2) 2 2 quire a fast response due to the changing nature of the load and when the system is required to work in partial-load A schematic of the PEM-RFC that is connected to the PV mode (also called turndown ratio) [ ]. PEM fuel-cell tec 9 h- system and the office building is shown in Fig. 3. The PV nology is considered the best type of fuel-cell technology system is the main source of power, while the PEM-RFC for transport applications in fuel-cell electric vehicles andher e represents the primary ESS. Inside the RFC system, material-handling equipment, and in backup power sys- we can see that the stack represents the core subsystem tems thanks to its low operating temperatures (<120°C) in which electrochemical reactions take place. The stack and good efficiency relative to other fuel-cell technologies is made of tens to hundreds of repeated cells called mem- [30, 31]. PEM-electrolysis technology, a sister technology of brane electrode assemblies that are sandwiched between the PEM fuel cells, is used to split water molecules into metallic or composite bipolar plates that should have hydrogen and oxygen. PEM-electrolysis technology has good corrosion-resistance properties [36]. The name of Table 1: Design parameters for the PV solar system Parameter Value Unit Nominal system size 400 kW DC PV-module type SunPower SPR-E19-310-COM PV-module capacity Up to 310.15 W DC DC/AC inverter SMA America: SC750CP-US DC/AC inverter efficiency 97.587% Maximum DC voltage 820 V DC System losses 5.44% Annual DC-degradation rate 0.50% per year Total module area 2094 m PV-system capital cost $578 351 (including contingency = 3% of subtotal cost) Total indirect cost $72 723 (engineering, permitting and environmental, land and sales tax) Total installed system cost $651 074 Total installed cost per capacity $1.63/W DC Incentives and taxes Variable (location-specific) including 26% in federal investment tax credit Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 279 PV Modules Combustible Gas Detector DC input One-way Valve DC/DC Convertor Dryer Hydrogen Storage DC/AC Inverter Back Pressure Stack Regulator Oxygen Pump Vent 80-90°C Controllable City Water Water Valve Water Cleaner 20°C Expander Air Compressor FC Mode EC Mode Control DC/AC Unit Inverter PV Modules DC input DC/DC Convertor Fig. 3: Schematics of (a) a unitized-stack PEM-RFC ESS and (b) an LIB ESS this technology comes from the structure of the mem- of different parts that are assembled and attached to the brane, which is usually made of a solid polymer electrolyte stack to supply the power and transport fluids used in the membrane that is deposited with a catalyst layer made electrochemical reactions inside the stack and to collect from PGMs, which include platinum, iridium and ruthe- and dry the gases (mainly hydrogen) produced from the nium. The catalyst coated membrane, or CCM (the mem- electrolysis process. Several subsystems are included in brane with electrode layers), is then attached to carbon- or the BOP, such as power electronics (AC/DC rectifier, DC/ titanium-based gas-diffusion layers from both sides [11]. AC inverter or a bidirectional inverter that can replace The main function of the gas-diffusion layer is to facili- the individual AC/DC rectifier and DC/AC inverter in one tate the diffusion of water and gas molecules to and from dual-function system, transducers, electric wiring, re- the catalytic layers. The second important system in the lays and switches), water-management system (water PEM-RFC is the balance of plant (BOP), which is composed pumps, heat exchangers, pipes and hoses, manifolds, Compressor Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 280 | Clean Energy, 2021, Vol. 5, No. 2 etc.), thermal-management system (e.g. heat exchan- add the membrane electrode assembly parts and for the gers, thermal host to recover waste heat, air blowers and gasket/seal on the edges. The estimated cell power density manifolds) and hydrogen-processing and storage unit for the electrolyser stack is 0.85 W/cm . The estimated (hydrogen dryer, single- or multiple-stage compressors power density of the fuel-cell stack is 0.409 W/cm . and stationary storage tanks) [11 33 , , 36]. The sizes of the RFCs and LIBs are based on the elec- Fig. 4 shows the polarization curve for a PEM-RFC. In tricity profile for the building. It was also assumed that the base case, we set the current densities for the fuel cell the building will have a stand-alone PV-ESS so that there and electrolyser at 0.56 and 0.50 A/cm , respectively. The is no need to connect the building to the power grid. The reference potential, j, for the fuel cell is set at 0.73 V, while estimated gross power for the fuel cell and electrolyser are the reference potential is set at 1.7 V for the electrolysis 162 and 396 kW, respectively. Parasitic losses, power losses operating mode. These values give an estimated stack effi- in the BOP, were estimated at 12% and 15% of the gross ciency of ~42.9% in the ideal case. The stack roundtrip effi- power for the electrolysis and fuel-cell operating modes, ciency can be calculated using the following equation: respectively. The functional specifications for the LIB ESS are sum- Q V FC FC η = η .η = (3) RT,stack FC EC marized in Table 3. The base system is an LIB that con- Q V EC EC tains nickel–manganese–cobalt in the cathode layer and where Q is the charge transferred during the fuel-cell FC graphite in the anode layer. The electrolyte is usually made operating mode, Q is the charge transferred in the elec- EC from lithium salts (LiPF ). The LIB ESS was designed to have trolysis mode, V is the supplied voltage to the cell in the FC charging and discharging rates (c-rates) of 0.50, which is fuel-cell operating mode and V is the supplied voltage EC equivalent to 250 kWh/h for a 500-kWh LIB system, with to the stack in the electrolysis mode. Under the ideal con- the lowest DoD of 125 kWh. The roundtrip efficiency was version cycle of the closed-loop H /H O storage system 2 2 assumed to be 90% [36, 37]. Unlike LIBs, RFCs can have zero with no current leakage, Equation (3) can be reduced to DoD, which happens when the fuel-cell mode (discharging Equation (4) as follows: mode) in the RFC can no longer operate due to the unavail- FC ability of hydrogen in the storage tanks. For consistency η = (4) RT,stack EC in the economic analysis, we assumed the same financial It is important to mention that the system roundtrip effi- ciency is usually less than the stack roundtrip efficiency Table 2: Functional specification of the RFC ESS because of the parasitic losses in the electrolysis process Parameter Electrolyser Fuel cell Unit/notes and fuel-cell BOP components. The system roundtrip effi- ciency can be expressed as: System power 396 161.88 kW Gross system power 450 190.45 kW (W − W ) stack BOP FC η = (5) RT,system Stack operating tem- 60–90 60–90 °C (W + W ) stack BOP EC perature range where W is the energy consumed by the stack and Stack operating 1–20 1–20 bar stack W is the parasitic loss that represents the energy con- pressure range BOP Total platinum- 1.5 1.5 g/m sumed by the BOP parts. Subscripts FC and EC in the above group-metal equations refer to the fuel-cell and electrolysis operating loading modes, respectively. Total cell area 800 800 cm The functional specifications of the cell, stack and key Cell active area 680 680 cm parameters for the BOP system are summarized in Table 2. Current density 0.5 0.56 A/cm The assumed cell total area is 800 cm with an active area Reference potential 1.7 0.73 V 2 2 of 680  cm , leaving 120  cm for the binding perimeter to Power density 0.850 0.409 W/cm Single-cell power 578.00 277.984 W Polarisation Curve for PEM-RFC Cells per system 685 2.5 Cells per stack 98 Q V FC FC RT, stack = Q V Stack per system 7 EC EC 2.0 Stack electrical 75% 48% efficiency 1.5 BOP power 12% 15% Of the 1.0 consumptions rated system 0.5 power Electrolysis Mode Fuel Cell Mode System efficiency 36.0% 0.0 –1,500 –1,000 –500 0500 1,000 1,500 2 Based on the target values by the US Department of Energy’s Hydrogen Current Density (mA/cm ) and Fuel Cell Technologies Office [44, 45]. Fig. 4: Polarization curve for a PEM-RFC stack Percentage of the nominal system power. Potential (V) Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 281 Table 3. ESS design and cost model parameters Parameter RFC LIB Units Notes System-design parameters Nominal system 450 (EC) 500 (nom- kW for RFC Unitized stack that operates in FC/EC modes size 162 (FC) inal size) kWh for LIB Battery/fuel-cell Proton-exchange NMC111 NMC111 (cathode) and graphite (anode) chemistry membrane Discharge rate 162 250 kW For RFC: fuel-cell efficiency = 48% and can supply (kWh/h) ≤162 kWh/h LIB: can discharge 250 kWh/h Lowest depth of 0 125 kWh Fuel cell can operate until the hydrogen is no discharge (DoD) longer available in the hydrogen-storage tanks Storage-system cost Variable ($20–750 N/a Using cost values for stationary steel–molyb- per kg-H ) denum tanks from [46] and for underground storage in salt cavern from [47] Roundtrip efficiency 36% 90% Average roundtrip efficiencies for PEM-RFCs [24] and LIBs [26, 40] Economic-analysis parameters Discount rate 10% 10% System lifetime 20 20 Year Estimated stack 4000 5000 Cycle Assuming 10 000 cycles for RFCs, where voltage lifetime fade does not to exceed 20% of the original ref- erence voltage [38, 48] Assuming 5000 cycles for nickel–manganese–co- balt LIBs, with storage-capacity fade of 20% [29] Price of electricity 0.022 0.022 $/kWh Using the real levelized cost of electricity (LCOS) generated by the PV system O&M cost 0.02 0.02 $/kWh Estimated value using values from Reznicek and Braun [21] for similar ESS technology System capital cost 600 350 $ Adjusted value based on the system capital cost for PEM electrolysers discussed in Mayyas et al. [33] and average values for LIBs from [26, 49, 50] Installation factor 25% 25% $/kW References [36, 46] Installed system 337 500 218 175 capital cost (TIC) for ESSs Cycles per year Variable Variable Transition from charging to discharging modes is considered one cycle NMC111 formula is LiNi Mn Co O . 0.33 0.33 0.33 2 We assumed a stationary steel–molybdenum storage tank is used for a low storage amount that is ≤1000 kg) (based on the cost estimates obtained from [46]) and a bulk storage system using underground salt caverns when the storage size is >1000 kg (cost estimates obtained from Ahluwalia et al. [47]). O&M cost per kWh discharged electricity from the ESS. Equivalent to 3.8–6.6% of the annual amortized installed system cost for LIBs and 8.6% of the annual amortized installed system cost for RFCs. Excluding the hydrogen-storage system for RFCs. parameters (discount rate, system lifetime and installa- temperature, charging and discharging rates, lowest DoD tion factor) for RFCs and LIBs. We also assumed the O&M and cathode chemistry [8 ]. The estimated lifetime for an cost to be 2 cents/kWh discharged for both ESSs [11]. The NMC111 nickel–manganese–cobalt LIB is 5000 cycles as- O&M cost is estimated based on similar values from the suming the lowest DoD of 25% [29] with a remaining literature, which accounts for the cost of operating and storage capacity of 75% of the original nominal storage maintaining the system during its expected useful life. capacity in kWh. It is important to highlight that the ESS Most of the O&M cost is associated with the cost of refur - lifetime is 20  years, which indicates that RFC and LIB bishing or replacing the old cells in the RFC stack and LIB. stacks should be replaced several times during this time For PEM-RFC, the estimated lifetime of the stack is 10 000 period. The replacement cost is included in the O&M cost. cycles, assuming that acceptable voltage degradation does We also assume flexible storage capacity for the not exceed 20% of the original reference voltage (based hydrogen ESS, which means that the RFC can produce on the values given in 38, 39). The useful lifetime of the hydrogen at all times when the PV supplies power to LIB is dependent on several factors, such as the operating the RFC system; thus, idle time for RFCs is theoretically Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 282 | Clean Energy, 2021, Vol. 5, No. 2 equal to zero (excluding scheduled maintenance and where TIC is the total installed system cost, r is the dis- unexpected interruptions). The roundtrip efficiency count rate (base value = 10%), N is the system lifetime in (η ) is another key parameter that plays a key role in years (base value = 20 years), kWh is the total discharge RTE dis determining the LCOS values. The PEM-RFC system is electricity from the system in one cycle, is the a P verage known for its low η , which is in the range of 28–36% cost of electricity used in charging the system, t is the RTE [23], whereas LIBs are known for their high η , which number of hours (max t  =  8760  h/year), η is the system RTE RTE can reach 85–95% in most cases [26, 40]. In this study, roundtrip efficiency and O&M cost is the estimated oper - we calculated the roundtrip efficiency of the RFC as 36% ation and maintenance cost ($0.02/kWh discharged from using the average efficiency values for PEM fuel cells and the ESS). PEM electrolysers from literature, and the corresponding roundtrip efficiency for the LIB system was assumed to be 90% based on the typical values in the literature [26, 2 Results and discussion 40]. It is important to highlight that the number of cycles 2.1 LCOE production from a PV solar system (charge/discharge) is determined based on the PV power As mentioned above, the building-energy system is de- profile, which determines the charging vs. discharging signed to have a self-sustained energy system that can modes for RFCs and charging/discharging/idle modes produce and store energy when the electricity supplied for LIBs. This factor plays a key role in shaping the final by the PV system exceeds the building demand and can LCOS values. convert it back to the building when needed in the night- Among the flexibilities that the RFC system provides is time or when the PV production cannot cover the building the ability to design the system to act as: (i) an ESS and demand. Fig. 5 summarizes the results obtained from the (ii) a hydrogen-production system to produce hydrogen SAM model for El Paso, TX. The nominal PV-system size for other end uses (e.g. as a fuel for fuel-cell electric ve- is 400  kW , which can produce ≤757  931 MWh in the hicles) [11, 41]. The latter design parameter is discussed DC first year. The peak electricity generation happens in the here as an indicative parameter to guide researchers who month of May in El Paso, TX and the lowest generation are interested in such dual-function systems. Equation (6) occurs in December. The size of the PV system was de- describes the status of the charging state (expressed in kg termined based on the building-energy profile and the of hydrogen stored in the hydrogen tanks): ability of the ESS to store excess energy so that curtailed S = S + H + S (6) t t−1 RFC 0 electricity from the PV system was at a minimum. The where S is the amount of stored hydrogen at time t (kg), profile of the electricity produced by the PV system is S is the amount of stored hydrogen at time t–1 (kg), H shown in Fig. 6 with a maximum electricity production t–1 RFC is the amount of hydrogen produced or consumed by the of 389 kWh. RFC system at time t (kg) and S is the strategic hydrogen- As shown in Fig. 5, the capacity factor of the PV system storage position (kg), which can be set to zero if the RFC is only 21.7%, which seems to be lower than the average system begins with empty tanks. capacity factor for wind turbines in Texas [42]. Even with this low capacity factor, solar energy still supplies low-cost electricity at a cost of 2.21 cents/kWh, which is comparable 1.4  LCOS to that of wind energy [42]. The LCOS is a quantitative metric that is used to evaluate the cost of storing and converting energy into a usable form using an ESS. As shown in Equation (7) below, the 2.2 LCOS in RFCs and LIBs LCOS is a function of several parameters, including capital As discussed previously, Equation (6) was used to calculate cost of the ESS, electricity price, number of cycles between the cumulative amount of hydrogen produced and con- charging/discharging modes, efficiency of the ESS and cap- sumed by the RFC system (Fig. 7). With a strategic starting acity of the ESS. The following formula was adapted from hydrogen-storage position of 250 kg-H at the beginning of [21] and then modified to calculate the LCOS for both LIBs the year, the RFC will not fall to a point at which there is and PEM-RFCs: no hydrogen in the hydrogen tanks—that is, this starting point will ensure that the cumulative hydrogen curve will r P TIC −N 1−(1+r) t=0 not fall below zero. The maximum amount of hydrogen LCOS = + + C(7) O & M t. η RTE stored in the hydrogen tanks reaches 1336  kg-H on 25 kWh 2 dis t=0 June. After that, the hydrogen level fluctuates around this value for several weeks and then starts to decline with time until it reaches 674  kg-H at the end of the year. Careful 2 2 For ESSs, roundtrip efficiency is defined as the ratio of the energy planning is needed here as to whether to sell part of this put in (charging mode) to the energy retrieved from storage in stored hydrogen for other end uses (e.g. for fuel-cell elec- the discharging mode. Inefficiencies include losses in the ESS tric vehicles) or to design the hydrogen-storage system to itself and the losses in transmitting and converting this energy from electricity to other electrochemical energy forms. accommodate all of the hydrogen produced in the whole Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 283 Metric Value Monthly Energy Production 80 000 Annual energy (year 1) 757,931 kwh Capacity factor (year 1) 21.7% Energy yield (year 1) 1,903 kWh/kW 70 000 Performance ratio (year 1) 0.79 Levelized COE (nominal) 2.77 ¢/kWh 60 000 Levelized COE (real) 2.21 ¢/kWh Electricity bill without system (year 1) $72,758 50 000 Electricity bill with system (year 1) $32,633 Net savings with system (year 1) $40,126 Net present value $155,734 40 000 Simple payback period 12.8 years Discounted payback period NaN 30 000 Net capital cost $651,074 Equity $0 20 000 Debt $651,074 10 000 Jan Feb Mar Apr MayJun JulAug Sep Oct NovDec Fig. 5: Summary report of the PV system from the SAM model PV [kWh] generated by the RFC system is 148.3 kWh on 3 August, which seems to be one of the hottest days in the year in El 400 Paso, TX, so that the energy demand in that day exceeds 350 what the PV can produce. Unlike the RFC system, the duty cycle in the LIB is a bit different. Charging and discharging do not always sum up to 100% because of the limited storage capacity of the LIB. Another issue with the LIB is the ability to supply electricity during peak times for several hours at a discharge rate that is suitable to meet the building-energy profile. For example, a hot summer in El Paso, TX may result in a building that 16-Dec 4-Feb 26-Mar 15-May 4-Jul 23-Aug 12-Oct 1-Dec –50 demands large amounts of electricity for several hours. To Fig. 6: Profile of the electricity generated by the PV system avoid such a case, the storage capacity of the LIB in kWh and the discharge rate (kWh/h) can be increased to accom- modate such a scenario, but this does not automatically Cumulative Hydrogen Production/Consumption eliminate the higher percentage in the idle state. The duty 1,600 1,400 cycles for a 400-kWh LIB for the month of April are shown 1,200 in Fig. 9a. The battery can store electricity until it reaches its 1,000 maximum storage capacity (nominal = 400 kWh). Similarly, the LIB is also limited in terms of the duration over which it can supply the power; e.g. a 400-kWh LIB with 100-kWh lowest DoD can supply 300 kWh during the discharge time. In any case, increasing the size of the LIB system can reduce 16-Dec 4-Feb 26-Mar 15-May 4-Jul 23-Aug 12-Oct 1-Dec the percentage of idle time but, on the other hand, this will Time/Date compromise the total system cost, and thus may increase the capital-cost component in Equation (7). For a larger LIB Fig. 7: Cumulative hydrogen production/consumption in the RFC system (starting H level = 250 kg) size (1000 kWh) in Fig. 9b, one can see that the idle time in April (also applies to other months in the year) can be re- year. The electricity generated by the RFC is shown in Fig. 8 , duced substantially and, on the other hand, the discharge with the total of 170  964 kWh, which contributes ~36.8% time will increase substantially as well. Fig. 10 shows the of the total annual building demand. The peak electricity percentages of time for each operating mode as a function Electricity Generated by PV (kWh) kg-H2 kWh Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 284 | Clean Energy, 2021, Vol. 5, No. 2 kWh from RFC of the LIB size in kWh. The percentages of the charging and discharging modes can be increased from 8.4% and 42.5%, respectively, for a 400-kWh battery to 15.1% and 61.8% for charging and discharging times, respectively, for a 2000- kWh LIB system. The biggest advantage here is the ability to lower the time that the LIB spends in idle mode, which can be reduced from 49.1% for a 400-kWh LIB to 23.1% for a 2000-kWh LIB system. The effect of changing the LIB size on the LCOS and the expected amount of kWh discharged from the LIB during 16-Dec 4-Feb 26-Mar 15-May4-Jul 23-Aug 12-Oct 1-Dec the whole year are shown in Fig. 11. The LCOS increases from 25.5 cents/kWh for a 400-kWh LIB to 65.3cents/kWh Fig. 8: Electricity generated by the RFC system for a 2000-kWh LIB. The main reason behind such an in- crease is that the change in the numerator in the capital- cost component (the first term after the equals sign in Equation (7)) is increasing much more than the expected increase in the denominator, which implies that such a so- lution may not be so competitive from a cost perspective. In terms of the amount of electricity that the LIB system can 200 supply to the building when the PV system is off (night- 150 time) or when it cannot supply the whole amount of the building need at certain times, we can see that a larger LIB system can supply more electricity to the building. A 400- kWh LIB system can supply 97  905 kWh/year (equivalent 31-Mar 10-Apr 20-Apr 30-Apr to 57.3% of the required electricity to be supplied by the Date ESS) and a 2000-kWh LIB system can supply ≤168  832 1,200 kWh/year (98.8% of the required electricity from the ESS). 1,000 However, energy engineers should also weigh up the cost of connecting the building to the grid and whether they want to design the building to be a stand-alone 100% green building or design the building to rely on both a renewable- 400 energy source and the power grid when the demand ex- ceeds the total supply from the PV and LIB systems. Such a decision can be extended to include other selection cri- teria and incentives offered to green buildings, which are 31-Mar 10-Apr 20-Apr 30-Apr beyond the scope of this paper. Date Thus far, this model was developed to evaluate the Fig. 9: State of charge of the LIB in April for: (a) a 400-kWh LIB (min- economics and the resiliency of the ESS. It should be imum value = lowest depth of discharge = 100 kWh) (note: zero values highlighted that other metrics can be used in similar com- represent idle states); and (b) a 1000-kWh LIB (minimum value = lowest parisons for energy-efficient buildings. In their compara- depth of discharge = 250 kWh) tive study, Harkouss et al. [ 43] used five metrics to assess the 0.70 180,000 70% 160,000 0.60 60% 140,000 0.50 50% 120,000 0.40 40% 100,000 80,000 30% 0.30 60,000 20% 0.20 40,000 LCOS 10% 0.10 20,000 KWh Discharged 0% 400 600 800 1000 1200 1400 1600 1800 2000 - - 400 600 800 1000 1200 1400 1600 1800 2000 LIB Size (kWh) LIB Size (kWh) % Charge% Discharge % Idle Fig. 11: Effect of changing the LIB size on the LCOS in the LIB storage Fig. 10: Effect of changing the LIB size (in kWh) on the percentage of system (note: the total amount of the electricity that is not directly the time in the operating modes (charge, discharge and  idle) Lowest depth of discharge = 0.25 of the nominal LIB size in kWh. covered by the PV system = 170 964 kWh/year) Percent of time kWh in LIB kWh from RFC kWh in LIB LCOS ($/kWh) kWh Discharged Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 285 performance of the net-zero-energy buildings. These metrics sensitivity-analysis purposes: capital cost and roundtrip include (i) energy-efficiency indicator (total energy consump- efficiency as the key system-design parameters, electricity tion); (ii) energy balance and self-sufficiency indicator, which cost and O&M as the key operating-cost parameters, and is measured using a load matching index; (iii) grid-stress in- lifetime and discount rate as the most important financial dicator using a grid-interaction index; (iv) economic indica- parameters. Tornado charts in Fig. 12a–c show the effect of tors using life-cycle cost, LCOE and simple payback period; changing such parameters on the LCOS in 1000-kWh LIB, and (v) environment indicator using CO-equivalent emis- 2000-kWh LIB and 450 kW RFC systems (equivalent dis- sions. We believe that these performance indicators can be charge rate = 162 kWh/h), respectively. used in a more holistic approach in the future to evaluate the It seems that the LCOS models for both LIB and RFC tech- performance of energy-efficient buildings and different ESSs. nologies are more sensitive to capital cost followed by dis- count rate and expected system lifetime. Operating costs (electricity cost and O&M cost) seem to have a lower effect 2.3 Sensitivity analysis on the final LCOS in both LIBs and RFCs. Another important Sensitivity analysis was performed to assess the effect remark for the LIB system is the impact of changing the of changing some financial and system-design param- nominal storage size in kWh, which will automatically in- eters on the LCOS. Based on the LCOS formula in Equation crease the value of the supplied energy by the system (de- (7), the following parameters were selected for the nominator in the capital-cost component in Equation (7)), which in turn does not seem to outweigh the substantial in- crease in the capital cost in the numerator; thus, increasing LCOS-2,000 KWH LIB (NOMINAL=65.3 CENT/KWH) the LIB size will probably increase the LCOS. Capital Cost Discount Rate 3 Conclusions Lifetime Green-energy buildings that totally or partially rely on renewable-energy sources such as solar or wind energies Rondtrip Efficiency need cost-effective ESSs to store energy during energy- Electricity Price production periods and the ability to reverse it during non- +50% -%50 O&M Cost production times. RFCs and LIBs were studied in this paper to evaluate their ability to increase green-building resili- –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 Change in LCOS ($/KWH) ency and reduce the cost of energy storage. Results confirm B the attractiveness of both technologies as electricity- LCOS-1,000 KWH LIB (NOMINAL=32.1 CENT/KWH) storage systems that can be connected to commercial Capital Cost buildings. Cost results show that a 400-kW PV system can generate electricity at a cost of 2.21 cents/kWh in El Paso, Discount Rate TX. In terms of energy storage, we found that the average Lifetime LCOS using a 450-kW RFC is ~31.3 cents/kWh, while this could reach as low as 25.5 cents/kWh using a small LIB Rondtrip Efficiency ESS (400-kWh). However, while the RFC provides the flexi- Electricity Price bility required to meet building-energy profiles, LIBs may +50% not be able to meet building needs unless the storage -%50 O&M Cost size is increased substantially, which in turn incurs more –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 Change in LCOS ($/KWH) energy-storage costs, making LIBs less favourable from an economic perspective. Given the nature of the energy pro- LCOS-RFC (NOMINAL=31.3 CENT/KWH) duction and energy consumption of this selected commer - Capital Cost cial building, it seems that LIBs provide a better ESS at a Discount Rate lower cost in $/kWh and can be designed to meet building needs even at peak times. A hydrogen-based ESS is still a Lifetime viable solution for such buildings, but may need to be con- Rondtrip Efficiency sidered when the goal is to store energy for longer periods or when it is are connected to a nearby hydrogen station, Electricity Price where it can be used primarily as a hydrogen-production +50% O&M Cost -%50 system and as an ESS as a secondary goal. Parametric sen- –0.40 –0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40 sitivity analysis of the stand-alone PV-ESS showed that Change in LCOS ($/KWH) the LCOSs of the LIB and RFC systems are sensitive to system capital cost, financial parameters (discount rate Fig. 12: Tornado charts showing the effect of changing certain param- and system lifetime) and, to a lesser extent, the roundtrip eters on the LCOS for: (a) a 1000-kWh LIB, (b) a 2000-kWh LIB and (c) a 450-kW (162 kWh/h) RFC system efficiency and the generated-electricity cost. Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 286 | Clean Energy, 2021, Vol. 5, No. 2 in Thermal Energy Storage Systems. Cambridge: Woodhead Acknowledgement Publishing, 2015, 1–28. The authors gratefully acknowledge the support and funding from [17] Mostafa MH, Abdel Aleem SHE, Ali SGet  , al. Techno-economic Khalifa University - Grant number 8474000238/FSU-2020-14 to con- assessment of energy storage systems using annualized life duct this research. cycle cost of storage (LCCOS) and levelized cost of energy (LCOE) metrics. Journal of Energy Storage, 2020, 29:101345. [18] Nikolaidis P, Poullikkas A. A comparative review of electrical energy storage systems for better sustainability. Journal of Conflict of Interest Power Technologies, 2017, 97:220–245. None declared. [19] Dunn B, Kamath H, Tarascon JM. Electrical energy storage for the grid: a battery of choices. Science, 2011, 334:928–935. [20] Argyrou MC, Christodoulides P, Kalogirou SA. Energy storage References for electricity generation and related processes: technolo- [1] EIA (Energy Information Administration). US Energy Facts gies appraisal and grid scale applications. Renewable and Explained. 2019. https://www.eia.gov/energyexplained/ Sustainable Energy Reviews, 2018, 94:804–21. us-energy-facts/ (20 March 2021, date last accessed). [21] Reznicek  E, Braun  RJ. Techno-economic and off-design ana- [2] Li G, Xuan Q, Akram MW, et al. Building integrated solar con- lysis of stand-alone, distributed-scale reversible solid oxide centrating systems: a review. Applied Energy, 2020, 260:114288. cell energy storage systems. Energy Conversion and Management, [3] The Directive 2010/31/EU of the European Parliament and of the 2018, 175:263–277. Council of 19 May 2010 on the energy performance of buildings, [22] Mostafa MH, Aleem SHA, Ali SG,et  al. Robust energy manage- Official Journal of the European Union. https://eur-lex.europa. ment and economic analysis of microgrids considering dif- eu/LexUriServ/LexUriServ.do?uri=OJ:L:2010:153:0013:0035:en: ferent battery characteristics. IEEE Access, 2020, 8:54751–54775. PDF (20 March 2021, date last accessed). [23] Hammond GP, Hazeldine T. Indicative energy technology as- [4] Aeleneia  L, Gonçalves  H. From solar building design to net sessment of advanced rechargeable batteries. Applied Energy, zero energy buildings: performance insights of an office 2015, 138:559–571. building. Energy Procedia, 2014, 48:1236–1243. [24] Ito H, Miyazaki N, Ishida M, et al. Efficiency of unitized revers- [5] Balaban O, de Oliveira JAP. Sustainable buildings for healthier ible fuel cell systems. International Journal of Hydrogen Energy, cities: assessing the co-benefits of green buildings in Japan. 2016, 41:5803–5815. Journal of Cleaner Production, 2017, 163:S68–S78. [25] Hur  W, Moon  Y, Shin  K, et  al. Economic value of Li-ion en- [6] Allen JG,et  al. Green buildings and health. Current Environmental ergy storage system in frequency regulation application from Health Reports, 2015, 2:250–258. utility firm’s perspective in Korea. Energies, 2015, 8:5000–5017. [7] Bakos GC, Soursos M, Tsagas NF. Technoeconomic assessment [26] IRENA. Battery storage technology improvements and cost of a building-integrated PV system for electrical energy saving reductions to 2030: a deep dive. In: International Renewable in residential sector. Energy and Buildings, 2003, 35:757–762. Energy Agency Workshop Dusseldorf, Kai-Philipp Kairies, [8] Tribioli L, Cozzolino R. Techno-economic analysis of a stand- ISEA/RWTH Aachen, 17 March 2017. https://www.irena. alone microgrid for a commercial building in eight different org/-/media/Files/IRENA/Agency/Events/2017/Mar/15/2017_ climate zones. Energy Conversion and Management, 2019, Kairies_Battery_Cost_and_Performance_01.pdf (20 March 1:58–71. 2021, date last accessed). [9] Chatzivasileiadi  A, Ampatzi  E, Knight  I. Characteristics of [27] EnergyPlus. https://energyplus.net/ (10 May 2021, date last electrical energy storage technologies and their applications accessed). in buildings. Renewable and Sustainable Energy Reviews, 2013, [28] Pezzutto S, Croce S, Zambotti S, et al. Assessment of the space 25:814–830. heating and domestic hot water market in Europe—open data [10] Lucchese FC, Canha LN, Brignol WS,et  al. Energy storage sys- and results. Energies, 2019, 12:1760. tems role in supporting renewable resources: global overview. [29] SAM (System Advisor Model, 2020.2.29). National Renewable In: 2019 54th International Universities Power Engineering Energy Laboratory. https://sam.nrel.gov/ (10 May 2021, date last Conference (UPEC). IEEE. Bucharest, Romania, 3–6 September accessed). 2019, 1–6. doi: 10.1109/UPEC.2019.8893470. [30] Mayyas A, Wei M, Chan SH, et al. Fuel cell forklift deployment [11] Mayyas A, Wei M, Levis G. Hydrogen as a long-term, large-scale in the USA. In: Stolten  D, Samsun  RC, Garland  N (eds). Fuels energy storage solution when coupled with renewable energy Cells: Data, Facts and Figures. Weinheim: Wiley-VCH, 2016, sources or grids with dynamic electricity pricing schemes. 334–342. International Journal of Hydrogen Energy, 2020, 45:16311–16325. [31] Wei  M, Chan  SH, Mayyas  A, et  al. Deployment and cap- [12] Kharel S, Shabani B. Hydrogen as a long-term large-scale energy acity trends for stationary fuel cell systems in the USA. In: storage solution to support renewables. Energies, 2018, 11:2825. Stolten D, Samsun RC, Garland N (eds). Fuels Cells: Data, Facts [13] Guney  MS, Tepe  Y. Classification and assessment of energy and Figures. Weinheim: Wiley-VCH, 2016, 257–269. storage systems. Renewable and Sustainable Energy Reviews, [32] Ayers KE, Capuano CB, Anderson EB. Recent advances in cell 2017, 75:1187–1197. cost and efficiency for PEM-based water electrolysis. ECS [14] Chen H, Cong TN, Yang W, et  al. Progress in electrical energy Transactions, 2012, 41:15–22. storage system: a critical review. Progress in Natural Science, [33] Mayyas  AT, Ruth  MF, Pivovar  BS, et  al. Manufacturing Cost 2009, 19:291–312. Analysis for Proton Exchange Membrane Water Electrolyzers (No. [15] Wagner L. Overview of Energy Storage Methods, Research Report, NREL/TP-6A20-72740). Golden, CO: National Renewable Energy 2007. http://www.moraassociates.com/reports/0712%20 Laboratory (NREL), 2019. Energy%20storage.pdf (20 March 2021, date last accessed). [34] Carmo M, Fritz D, Maier W, et al. Alkaline water electrolysis vs. [16] Cabeza  LF, Martorell  I, Miró  L, et  al. Introduction to thermal PEM water electrolysis—exploring their full performance. In: energy storage (TES) systems. In: Cabeza  LF (ed). Advances 227th ECS Meeting. Chicago, 24–28 May 2015. Downloaded from https://academic.oup.com/ce/article/5/2/273/6284344 by DeepDyve user on 01 June 2021 Mayyas et al. | 287 [35] Schmidt  O, Gambhir  A, Staffell  I, et  al. Future cost and [44] DOE (Department of Energy). DOE Technical performance of water electrolysis: an expert elicit- Targets for Hydrogen Production from Electrolysis. ation study. International Journal of Hydrogen Energy, 2017, https://www .ener gy .go v/eer e/fuelcells/ 42:30470–30492. doe-technical-targets-hydrogen-production-electrolysis [36] Wei M, Mayyas A, Lipman TE, et al. Fuel cell systems: total cost [45] DOE (Department of Energy). DOE Technical Targets for Fuel Cell of ownership. In: Lipman  TE, Weber  AZ (eds). Fuel Cells and Systems for Stationary (Combined Heat and Power) Applications. Hydrogen Production, A Volume in the Encyclopedia of Sustainability https://www.energy.gov/eere/fuelcells/doe-technical-targets- Science and Technology, Second Edition. New York City: Springer fuel-cell-systems-stationary-combined-heat-and-power (20 Nature, 2019, 27–81. March 2021, date last accessed). [37] IRENA. Utility-scale Batteries Innovation Landscape Brief, [46] Mayyas  A, Mann  M. Manufacturing competitiveness ana- 2019. https://www.irena.org/-/media/Files/IRENA/Agency/ lysis for hydrogen refueling stations. International Journal of Publication/2019/Sep/IRENA_Utility-scale-batteries_2019.pdf Hydrogen Energy, 2019, 44:9121–9142. (10 May 2021, date last accessed). [47] Ahluwalia RK, Papadias DD, Peng JK, et  al. System level ana- [38] Regmi  YN, Peng  X, Fornaciari  JC, et  al. A low tempera- lysis of hydrogen storage options. In: U.S. DOE Hydrogen and ture unitized regenerative fuel cell realizing 60% round Fuel Cells Program 2019 Annual Merit Review and Peer Evaluation trip efficiency and 10000 cycles of durability for energy Meeting, Washington, DC, 29 April–1 May 2019. https://www. storage applications. Energy & Environmental Science, 2020, hydrogen.energy.gov/pdfs/review19/st001_ahluwalia_2019_o. 13:2096–2105. pdf (10 May 2021, date last accessed). [39] Wu J, Yuan XZ, Martin JJ, et  al. A review of PEM fuel cell dur - [48] Bruijn FD. PEMFC Lifetime and Durability: An Overview. ability: degradation mechanisms and mitigation strategies. Thessaloniki, 21 September 2011. https://www.sintef.no/ Journal of Power Sources 2008, 184:104–119. globalassets/upload/materialer_kjemi/energikonvertering- [40] Dufo-López  R, Bernal-Agustín  JL. Techno-economic ana- og-materialer/bilder/pemfc-overview-thessaloniki-fa- lysis of grid-connected battery storage. Energy Conversion and debruijn---f.pdf (20 March 2021, date last accessed). Management, 2015, 91:394–404. [49] Tsiropoulos I, Tarvydas D, Lebedeva N. Li-Ion Batteries [41] Nguyen VN, Blum L. Reversible fuel cells. In: Barbir F, Basile A, for Mobility and Stationary Storage Applications. European Veziroğlu  TN (eds). Compendium of Hydrogen Energy, Volume 3: Union’s Joint Research Center (JRC). EUR 29440 EN, 2018. Hydrogen Energy Conversion. Cambridge: Woodhead Publishing, https://ec.europa.eu/jrc/en/publication/eur-scientific-and- 2016, 115–145. technical-research-reports/Li-ion-batteries-mobility-and- [42] U.S. Energy Information Administration. Wind Is a Growing Part of stationary-storage-applications (20 March 2021, date last the Electricity Mix in Texas. 15 October 2020. https://www.eia.gov/ accessed). todayinenergy/detail.php?id=45476#:~:text=In%20April%20 [50] Larsson P, Börjesson, P. Cost Models for Battery Energy 2019%2C%20wind%20generators,United%20States%20as%20 Storage Systems. KTH School of Industrial Engineering a%20whole. (22 March 2021, date last accessed). and Management. Energy Technology EGI-2018, TRITA- [43] Harkouss  F, Fardoun  F, Biwole  PH. Optimal design of renew- ITM-EX 2018:428. http://www.diva-portal.org/smash/get/ able energy solution sets for net zero energy buildings. Ener gy, diva2:1254196/FULLTEXT01.pdf (20 March 2021, date last 2019, 179:1155–1175. accessed).

Journal

Clean EnergyOxford University Press

Published: Jun 1, 2021

References