Analysis of an Electrical Energy Storage System Based on Flow Batteries

Proyecto Fin de Carrera Ingeniería de Telecomunicación Formato de Publicación de la Escuela Técnica Superior de Ingeniería Autor: F. Javier Payán Somet Tutor: Juan José Murillo Fuentes Dep. Teoría de la Señal y Comunicaciones Escuela Técnica Superior de Ingeniería Universidad de Sevilla Sevilla, 2013 Trabajo Fin de Máster Máster Universitario en Ingeniería Industrial Analysis of an Electrical Energy Storage System Based on Flow Batteries Autor: Elisa Hidalgo Moreda Tutor: José Alfredo Iranzo Paricio Dpto. Ingeniería Energética Escuela Técnica Superior de Ingeniería Universidad de Sevilla Sevilla, 2023 Trabajo Fin de Máster Máster Universitario en Ingeniería Industrial Analysis of an Electrical Energy Storage System Based on Flow Batteries Autor: Elisa Hidalgo Moreda Tutor: José Alfredo Iranzo Paricio Profesor Ayudante Doctor Dpto. Ingeniería Energética Escuela Técnica Superior de Ingeniería Universidad de Sevilla Sevilla, 2023 Trabajo Fin de Máster: Analysis of an Electrical Energy Storage System Based on Flow Batteries Autor: Elisa Hidalgo Moreda Tutor: José Alfredo Iranzo Paricio El tribunal nombrado para juzgar el trabajo arriba indicado, compuesto por los siguientes profesores: Presidente: Vocal/es: Secretario: acuerdan otorgarle la calificación de: El Secretario del Tribunal Fecha: Acknowledgements I pay tribute to my family and friends for the massive motivation and patience that have infinitely helped me stay focused. Special thanks to my tutor for the ongoing support as well as the invaluable amount of trust placed in me. It is only thanks to these people I have managed to accomplish this research project and the many years of intensive work ethic behind it. However, I would also like to express gratitude for the lessons learnt through this project, which I hope to never forget in the future. I have learnt to grasp opportunities behind visible difficulties. I genuinely hope to extend this mindset to other areas in my personal and professional life ahead. Elisa Hidalgo Moreda Sevilla, 2023 I Resumen En el contexto de una transición del modelo eléctrico y energético, este proyecto pretende explorar en detalle una tecnología prometedora: las Baterías de Flujo Redox desde múltiples puntos de vista. En primer lugar se presenta una introducción sobre el interés de estos sistemas: por qué los científicos buscan nuevos sistemas de baterías y qué pueden aportar las RFB en particular al escenario actual. Una vez expuesto esto, se examina el principio de funcionamiento y los componentes de las baterías de flujo redox. Comprender en profundidad cómo funciona el sistema permitirá llevar a cabo una investigación aún más amplia sobre el estado del arte de las baterías de flujo redox. Así, en la sección Líneas de investigación y retos se exponen las diferentes líneas de investigación, los retos encontrados y, en consecuencia, la forma en que se lideran futuras investigaciones. Por último, se realiza un estudio de viabilidad económica para evaluar la viabilidad potencial de este tipo de baterías en el mercado. Este estudio se basa en un sistema ya instalado en Japón, cuya estabilidad a largo plazo ha sido demostrada. Adicionalmente, se evalúa la viabilidad de dicho sistema hipotéticamente conectado a la red eléctrica española mediante un análisis de costes jerárquico e incremental, que tiene en cuenta múltiples factores. Finalmente, este trabajo concluye que existe una amplia variedad de áreas de investigación aún abiertas y que se requiere más investigación para encontrar la combinación ideal de los diferentes elementos en una RFB (electrolitos, electrodos, membranas, etc.) que logren demostrar estabilidad a largo plazo así como una eficiencia alta. Por el momento, la inclusión de las baterías de flujo redox en España no parece viable en términos económicos, debido a las limitaciones de la red y a los elevados costes de inversión inicial para producir todo el sistema. III 2Chapter 1. Introduction Additionally, fossil fuels strongly contribute to contamination and climate change , due to the greenhouse gases emission produced when burning them in order to get energy from them. This, followed by the growing demand of energy pointed out before, makes scientists seek for alternative energy sources. Specifically, renewable energies as they have the potential of being natural energy sources able to produce energy indefinitely. Renewable energy sources have almost no negative impact on the environment, they are "clean energy". Nonetheless, they behave in a very random and unpredictable way. For instance, wind energy depends on the speed of wind as much as photovoltaic sources rely on solar radiation to generate usable energy. This is where it becomes vital to use systems that are able to store energy, like batteries, thereby increasing economic feasibility of energy production through renewable sources. This is a promising scheme capable of balancing energy offer and demand. There are multiple renewable energy sources that can later be stored through battery systems, such as but not limited to [52]: •Solar energy •Wind energy •Hydraulic energy •Seawater energy •Geothermic energy •Energy from biomass and waste Since renewable energy sources are not the study matter of this project, no further emphasis will be made on them. It is also important to note that as much as there are multiple energy sources, there are also a vast variety of ways to store energy like capacitors, compressed air or fuel cells. Again multiple batteries co-exist in the electrochemical storing market, such as lead-acid or lithium-ion. However, what differentiates redox flow batteries from other battery types is: •Their ability to physically increase/decrease power and capacity in an independent manner •Low self-discharge level •Promisingly low degradation level due to the chemical principles they obey Still a lot of research has to be carried out on this battery system type for it to be feasible and useful in its required applications, as will be presented in this work. However, they are a valid candidate to aid decarbonization of the electric sector [52]. 1.2 Possible applications RFBs are low density battery systems, so they have not been considered for mobile systems but rather stationary ones. However, their ability to have energy and power uncoupled is seemingly interesting. Some of the possible applications of conventional batteries likewise apply to RFBs. For instance: • Load leveling. They store energy in periods of excess and consume it in periods when the price of energy is higher. As a result, making the cost of energy more economical. • Smoothing and power regulation. VRFBs are used to smooth the power peaks that occur due to high energy consumption. 1.2 Possible applications 3 • Uninterruptible voltage generators. Due to their short response time and long discharge cycles, these batteries come on line when there is a failure in the generation systems, thus avoiding serious failures in the electrical grid and the cost that this would entail. • Combination with renewable energies. The large number of useful life cycles of VRFBs, up to 20 years in some cases, makes them ideal to compensate for the intermittency and poor stability of clean energies. Already by 2010, Skyllas et. al accounted for multiple live applications of redox flow batteries. [58] Some of them are summarized in Table 1.1. Table 1.1 Example applications of existing RFB systems. Power (kW) Capacity (kWh) Country Year Application 200 800 Japan 1996 Load-levelling 200 1600 Japan 2000 Peak-shaving 250 500 South Africa 2001 UPS back-up power 4000 6000 Japan 2005 Wind energy storage and wind power stabilization Next chapter will dive into the working principle of Redox Flow Batteries. This settles the basis to then understand most common research and development challenges encountered through the literature. 2 Working principle of Redox Flow Batteries Th e research topic of redox flow batteries has been a fairly inviting one over the years. As previously alleged, they hold great promise for addressing the challenges posed by the transition to a renewable energy future along with other battery energy storage systems. This last statement raises the question of what exactly has the potential to put RFBs at an advantage when compared to other battery types. Although finding a concrete answer to this enquiry is complex, it is possible to draw a set of conclusions by merely examining this batteries’ working principle. As suggested by their name, RFBs work thanks to the occurring redox chemical reactions in them. To begin with, their working principle will be outlined with the help of Figure 2.1. Figure 2.1 Working principle of RFBs [67]. While the scheme presented in Figure 2.1 shows an all-vanadium system, the parts or components are equivalent regardless of the compounds taking part in the reactions. As seen in this diagram, the key components that must be present are: •Positive electrolyte (generally also known as catholyte) •Negative electrolyte (generally also known as anolyte) •Two tanks to store both electrolytes •Two pumps integrated in a flow control system 5 6Chapter 2. Working principle of Redox Flow Batteries •Membrane •Two electrodes From Figure 2.1, it can be depicted that the electrolytes are pumped through a battery cell where power can be put into or taken out of the fluids by changing their oxidation states. Since charging and discharging occurs in the stacks, the latter are directly related to the quantity of power that the system is able to supply. This point can be emphasized by affirming that the design is modular [35]: •If more energy is required to be stored in the system, the size of the tanks can be increased, •whereas if there is a higher requirement of power, more stacks can be added Figure 2.2 below shows a visual representation of a cell stack. As it can be seen, cells are stacked in both serial and parallel connections. Piling them up in series increases the voltage, while piling them up in parallel allows the obtention of a greater current. Figure 2.2 Construction of a cell stack [67]. To understand this architecture in detail, this chapter will be broken down into two main sections. Section 2.2 will explore how and where redox reactions occur inside the cell allowing its operation. On the other hand, Section 2.1 will focus on thoroughly analysing the rest of the system. 2.1 Main components Redox flow batteries consist of several essential components that work together to store and release energy. These components include: 2.1.1 Electrolyte tanks Redox flow batteries use two separate tanks to store the two different electrolytes: anolyte and catholyte. Their design must be done in order to contain and store the electrolyte content safely. As explained below the design of a Redox Flow Battery is modular: •Energy is directly related to the size of the tanks 2.1 Main components 7 •Power can be increased by considering a higher number of stacks The reason why energy is said to be related to the size of the tanks is because having more voluminous tanks allows for more electrolyte to be stored. The active species here are the electrolytes. Thus the capacity of the battery turns out to be directly proportional to the amount of active species that are accessible to take place in the occurring electrochemical reactions. This statement is inherently materialized through real applications. Figure 2.3 shows a 60kWh Redox Flow Battery System (tanks being by the side of the system). The purpose of such allvanadium battery is to integrate it with rooftop photovoltaic panels (PV). [ 7 ] On the other hand Figure 2.4 puts forward one of the Redox Flow Battery systems with highest battery capacity worldwide. This iron-chromium RFB was developed by Enervault and is even connected to the grid in such a way that it’s able to inject energy back to it during times of high demand. Figure 2.3 60 kWhequivalent tanks [7]. Figure 2.4 1 MWhequivalent tanks [48]. Regarding materials, not every single one is valid to build electrolyte tanks. In fact, the material used for the electrolyte tanks depends on multiple factors such as but not limited to chemical compatibility, cost or the ease of production. Some common materials employed may be: 1. High-density polyethylene (HDPE) 2. Polypropylene (PP)- poly tanks 3. Glass-reinforced plastic (GRP)- fiberglass tanks 4. Stainless steel 8Chapter 2. Working principle of Redox Flow Batteries For example, Figure 2.12 implicitly informs that sulphuric acid is stored in the tanks in this allvanadium system equations. Henceforth, the tank must be made of corrosion-resistance materials. Otherwise the security, longevity or stability of the tank may be compromised. [15] [1] If the commonly known Polypropylene and glass-reinforced plastic tanks were to be compared, it is notable that: • GRP tanks are handmade. As a consequence it is subjected to human error in the course of its production. On the other hand, PP tanks’ manufacturing process is automated. •GRP are costly to make, polyethylene/polypropylene tanks are less expensive •Due to the corrosion resistance of poly storage tanks, maintenance is less burdensome In the light of such facts, the market trends are directed towards considering poly tanks as opposed to fiberglass ones. 2.1.2 Electrodes Likewise electrolytes, electrodes of Redox Flow Batteries can be of two types: positive electrode or negative electrode. The positive one (cathode) is liable for the oxidation reaction, while the negative electrode (anode) facilitates the reduction reaction. Note that the type of reaction (oxidation or reduction) is being denoted taking into account the battery being discharged. Refer to Figure 2.9 and Equations 2.1 and 2.2. Regarding sizes, we may encounter that the size of one side of an electrode vary in the order of the tens cm . That is to say that the surface area of the electrodes (magnitude of interest as it will be drawn later on through this Section) will be in the order of the hundreds or thousands cm2 . This will depend on the casuistry and application. Thielke carried out experiments with 125cm2 surface area electrodes [ 64 ], while the IMDEA Energy Institute research group has worked with electrodes as big as 1800cm2each. [18] Anyhow, the distinction of the electrode being "positive" or "negative" only has to do with them being in direct contact with the catholyte or anolyte respectively. In reality, both anode and cathode are generally made from the same materials. This presents multiple advantages such as cost reduction or simplification of the manufacturing process. There can still be exceptions, where the anode and cathode are made of different materials so as to optimize concrete performance characteristics. Some of the most common materials for electrodes can, for instance, be: •Graphite •Carbon felt •Conductive ceramics •Mixedmetal oxides As it can be depicted from the previous list of examples, the electrodes must be highly conductive materials as it is through them that the electrons will flow in order to charge and discharge the battery. Likewise the higher the surface area of the electrodes, the more enhanced the electrochemical reaction as they provide "the reaction sites" [ 10 ]. They accommodate more active material improving the speed of electron transfer. An important point to emphasize is that electrodes are " inert " [ 10 ] and their main function is to be the bridge to enable electron flow. Thus the electrodes must exhibit rapid kinetics of the redox 2.1 Main components 9 couple reactions so that the energy efficiency and voltage of the system is optimized as much as possible. As much as the kinetics for these reactions must be fast, the kinetics of reactions that could produce hydrogen and oxygen at them must be slow, so that undesirable side reactions are inhibited. In view of this there have been research lines focused on intrinsic and catalytic treatments for both anodes and cathodes over time. Intrinsic treatments Intrinsic treatments aim to improve the performance of the electrodes in the whole energy storage system. These may include [21] among others: 1. Thermal treatment 2. Acid treatment 3. Chemical etching 4. Heteroatom doping 5. Microwave treatment Next, this research work will delve into them. Thermal treatment proves to be one of the most appealing methods for improving the electrodes performance, since given optimum conditions of temperature and time exposure functionalization of electrodes takes place. It results in improvement of the cell’s efficiency as oxygen-based functional groups are generated (this is also related to wettability). [ 6 ] In this way, many experiments were carried out. For instance, Skyllas-Kazacos [ 62 ] tested out a thermal treatment upon graphite felt for VRFB, findinf out that its energy efficiency increased by 10% after treating it at 400 degrees Celsius for thirty hours thanks to the formation of C-O-H and C=O functional groups on its surface. This effect accelerated electron transfer, thus increasing energy efficiency. Likewise, [ 20 ] found out the optimum pre-treatment temperature for GF to be 475 degrees Celsius. Acid treatment , on the other hand, has multiple advantages as well like good oxidation effect, simple operation and cost effectiveness. Li et Al ,Sun et Al treated carbon paper with mixed acid under the action of ultrasonic wave, and obtained a hydroxylated electrode [ 77 ]. Increasing treatment time increases the number of functional groups that enable active sites for reaction, yet the mechanical strength is worsened.Besides, excess functional oxygen group can increase resistance decreasing the conductivity of electrons at the electrodes.[ 77 ] Therefore, it becomes crucial for time of these treatments to be controlled. Chemical etching aims to increase specific surface area in carbon-based materials. A successful technique is KOH etching, which has been used in multiple manners to form pores, increasing the number of active sites for fast electrochemical reactions. In [ 79 ] a dual-scale porous electrode was developed with different size pores, presented in Figure 2.5, all thanks to KOH activation of the fibers of carbon papers: 10 Chapter 2. Working principle of Redox Flow Batteries Figure 2.5 The schematic of the conventional carbon-fiber based electrode and the proposed dualscale porous carbon electrodes [79]. Heteroatom doping involves doping organic compounds with nitrogen, phosphorous, or boron among others. Perhaps the most mature technology may be doping N into carbon-based materials. This enhances the electrodes performance in multiple waysimproving electrode activity as well as improving the reversibility of VO2+ / VO+ 2 on the electrode. Researchers have found that co-doping of nitrogen and oxygen futher promotes the redox reaction of vanadium ions than doping with oxygen or nitrogen atoms alone. Diverse studies have been carried out on this regard. As an illustration, Flox et al. [ 32 ] studied the optimum reaction time and temperature for the exposure of GF to NH3 / O2 (1:1) atmosphere revealing that subjecting graphite felt to this environment at 773K and 24 hours made it obtain the highest nitrogen and oxygen content and the best catalytic activity when utilized under an all-vanadium electrolyte (VO2+/VO+ 2). Lastly, microwave treatment prompts the heating process of the electrode and it allows for a precise control on this process. This treatments acts increasing both the functional group content and the specific active surface area; these two effects combined greatly increase the energy efficiency of the RFB. Catalytic treatments For a comprehensive illustration of a catalytic treatment, firstly it is convenient to review the principle behind them. Refer to Figure 2.6. Skyllas-Kazacos studied the effects of adding surface oxygen functional groups on the carbon electrode surface. "Detailed X-ray photoelectron spectroscopy (XPS) analysis confirmed that improved cell performance was associated with an increase in the oxygen functional groups on the surface of the carbon electrode, especially phenolic functional groups, which provide active sites that catalyze both the anolyte and catholyte redox reactions in an all-vanadium system" [30]. 2.1 Main components 11 Figure 2.6 Addition of surface oxygen functional groups to electrodes [30]. Such finding induced further investigations by other researchers. Yan et al. developed a graphite/- graphite oxide composite electrode with several oxygen functional groups in it (on the basal planes and sheet edges of the GO). [33] For instance: •Ether •Ester •Carbonyl 18 Chapter 3. Research lines and challenges Thus commonly electrolyte solutions for VRFBs contain up to several molL−1 of sulfate or hydrogen sulfate ions. These severe conditions rule out many materials that are considered, for instance, for PEM fuel-cell applications. Apart from Nafion, other anion-exchange and cation-exchange membranes were studied by the group of Skyllas-Kazacos at the University of New South Wales in Australia, especially in terms of chemical stability. Refer to [ 61 ] and [ 40 ] for further details, yet their findings will be presented and summarized below: • Cation-exchange membranes. Selemion CMV and N112, where CMV proved to be the least stable one under Vanadium Redox Flow Battery conditions as oppossed to N112 that did show promisingly chemically stable. • Anion-exchange membranes. New Selemion Type 2 and Selemion AMV, where Selemion type 2 showed great stability • Microporous membranes Crosslinking Daramic (a microporous separator) with divinylbenzene (DVB) also showed good chemical stability Nevertheless all membranes showed decrease in area specific resistance except N112 and New Selemion (type 2). While this effect might seem favorable, it’s actually not as it was proven to be a side-effect of membrane degradation. Other membrane properties are then likely to worsen upon the evolving degradation of polymers. This effect was attributed to the oxidation of polymers due to the oxidative power of V5+ ions. Speaking of stability this makes N112 and New Selemion (type 2) take the lead. Further investigations were carried out that drew light to other factors influencing membrane performance. An interesting one was the concentration of ions. However, it is seen that not always the same effect is detected as depicted from the two study cases highlighted below: • Sukkar and Skyllas found out that for a N112E/ H+ membrane, this one presented worse chemical stability when immersed in a 0.1M solution of V5+ ions than in a 1.0M solution of V5+ ions. This effect was attributed to a difference in swelling, when the solution was more dilute it provoked a higher degree of swelling in the membrane. This enlargement of pores allows the vanadium ions to enter into them leading to a faster degradation of the membrane [40]. • On the other hand, Soowhan Kim et al. [ 31 ] analysed the stability of sulfonated poly (sulfone) membranes (also known as S-Radel) keeping the molarity of sulfate ions constant and altering the concentration of V5+ ions: 0.1M and 1.7M. Refer to Figure 3.1. They found that higher concentrations of V5+ions accelerated the degradation of the material. On a note, let us take a closer look at Figure 3.1. Figure 3.1 a) is a 0.1M V5+ and 5.0M SO2− 4 stability test solution containing a N117 membrane after 74d at 40 degrees Celsius. Conversely, Figure 3.1 b) is the same stability test solution but containing a S-Radel membrane after 4 days under the same conditions. The difference in color is presumably due to the reduction of V5+ ions to V4+ ones in solution. Figures 3.1 c) and d) are pictures of the S-Radel membrane after 18 days of being soaked in 0.1M V5+ and 5.0M SO2− 4 solution and 1.7M V5+ and 5.0M SO2− 4 solution respectively. Degradation becomes more severe with increasing concentration. The severe mechanical damage is observed without obvious chemical degradation detectable by FTIR spectroscopy[ 31 ], which is why it is believed that this damage was provoked by the decrease in the weight of the polymer as a result of oxidation by the V5+ions. 3.1 Membrane development 19 Figure 3.1 Membrane degradation results [31]. While these presented membranes have been tested in terms of stability, it is important to note that not all membranes are subjected to chemical stability tests. Henceforth, this questions the potential feasibility of multiple innovative and promising technologies for membranes. Many cycles are needed to carry out these tests and chemical stability is a vital factor when it comes to assessing this technology. Research results about stability may not receive as much attention as favorable efficiencies but it is a field that requires further investigation. Table 3.1 presents results of durability tests carrid out on a broad variety of membrane types. 20 Chapter 3. Research lines and challenges Table 3.1 Durability tests for membranes [31]. Another factor of interest in membranes is water transport. The latter is neccessary for proton conduction as it contributes via Grotthuss mechanism along the center of the pores of the membrane. Grotthuss mechanism allows for "an intermolecular transfer of protons between water molecules in the channels penetrating the membrane" as it is the “hopping” of a hydrogen ion through permeation through the network of water molecule. [36] Therefore, by this logic, proton conductivity would be improved by enlarging the membranes’ pore sizes. Yet this will provoke a higher degree of crossover. Investigations shall try to find them the optimum pore size as reducing them reduces both the ion transport of vanadium but also proton conduction across membrane. Another effect to do with water transfer is that net water transfer will cause the electrolyte in one half-cell to become more concentrated and the other electrolyte more diluted in the second half-cell. This imbalance will result in a decrease in capacity and efficiency of VRFBs. Nevertheless a study carried out by Sukkar and Skyllas in [ 60 ] shows that unless VRFBs are over-discharged, water transport within the system is not a major issue during battery operation even if net water transport imbalance occurs throughout the cycles. Another important point on improving membranes is reducing their vanadium ion permeability. Since the diffusion of vanadium ions results in cross-mixing and hence, promotes the self-discharge of the battery it is reasonable for this parameter to be monitored. It is a standard procedure in VRFB membrane research. Normally diffusivity tests are carried out by using solutions of VOSO4, even 3.1 Membrane development 21 though it has been shown that the diffusivity of the different vanadium ions present in VRFBs varies depending on their oxidation state. [56] Associated with thismembrane foulingthat is, vanadium ions getting trapped in the Nafion membrane. The membrane’s properties can be regenerated by soaking the membrane in sulfuric acid, which removes this ion from the pores. Vijayakumar et al. recently reported an NMR spectroscopy study that focused on elucidating the interaction between N117 and the different vanadium ion species in the VRFB electrolyte solutions [ 70 ]. This finding is exemplified in Figure 3.2. Figure 3.2B is after sonication. Sonication removes any potentially surface-adsorbed vanadium species. After this process takes places two wavelength adsorption peaks become noticeablein the range of 625 nm and 760 nm. These two ranges coincide with the values reported in literature for the absorption of [VO(H2O)5]2+ . In this way, study [ 70 ] concludes that membrane fouling is mostly caused by V4+ions. Figure 3.2 UV/Vis absorption of a Nafion membrane after cycling in a vanadium redox flow battery (A) and after sonication of the same membrane in deionized water (B). Spectrum (B) matches that of pure VO2+ solutions. [70]. On the other hand, there exist diverse papers on how Nafion sulfonic groups affect different transport properties. [ 55 ] In this study, N117 membrane was modified with polyaniline (PANI) and polypyrrole (PRP). These composite membranes showed a higher reduction of V4+ ion diffusion, interacted more strongly with sulfonic acid groups. This suggests that V4+ ion diffusion is highly dependent on the availability of sulfonic acid group, but proton conduction is not that much. Even if PANI and PRP were intended to be used as materials for membrane development, it turned out that neither was a suitable material. PANI/N117 degraded quickly. PRP was very quickly reduced from its conductive form to its neutral structurenot useful for application. 22 Chapter 3. Research lines and challenges 3.1.2 Cation-exchange membrane From now on, focus is drawn towards cation exchange membranes. In general, their main disadvantage is that they are not especially selective as to what cations cross it. Thus, the cross-mixing issue arises. [56] Nafion and other fluorocarbon-based membranes Nafion, a fluorocarbon polymer, consists of a hydrophobic Teflon-like backbone with hydrophilic side chains that are decorated with sulfonic acid groups. It is the most utilized material for membrane development. Still, as explored in the future economic analysis of the battery system, it turns out to be quite expensive. Most membrane development research revolves around this material. It was found that incorporating SiO2 particles inside the hydrophillic channels network of Nafion (see Figure 3.3) showed similar conductivity but decreased water uptake. Henceforth, less crossover was taking place as reflected in Figure 3.4. This reflected upon an energy efficiency increase from 73.8 to 79.9 %. [73] Figure 3.3 Schematic of a Nafion membrane modified with SiO2by sol-gel synthesis [73]. 3.1 Membrane development 23 Figure 3.4 Comparison of permeabilities of different vanadium ion species through pure Nafion and Nafion/SiO2[73]. This finding prompted investigation of another type of compound to synthesize Nafion membranes with. A new hybrid membrane with Nafion/ORMOSIL was synthesized. This membrane showed lower water uptake, yet simultaneously, increased proton conductivityit reported a maximum energy efficiency of 87.4%. The same group modified the synthesis of both Nafion/ SiO2 and Nafion/ORMOSIL membranes by adding ammonia solution during the sol–gel reaction. Presumably this treatment increases the particle size of SiO2 inside the Nafion channels. The observed V4+ ion diffusion through their membranes is about half that reported by previous studies. [54] Additionally, other fluorinated polymers around 10-15 years ago were investigated, such as PVDF-g-PSSA-co-PMAc but no further results have been reported. [56] Hydrocarbon-based membranes Other hydrocarbon-based membranes were investigated due to the interest of them being more costeffective than Nafion ones. However, most of them were to be found unstable in acidic environment. [56] Nevertheless, in 1996, Hwang and Ohya [ 26 ] tested out photochemically chloro-sulfonated low-density polyethylene (PE) membranes and asymmetric membranes (PE+chemically bonded to porous polyolefin films containing silica particles). They were crosslinked and sulfonated to different degrees. Authors proposed that asymmetric membranes would have very high coulombic efficiencies (low diffusion coefficient for most vanadium ion species and an area specific resistance comparable to N117). However, this proposal was not demonstrated in any publication afterwards. [26] [56] In an effort to find an alternative to Nafion membranes, SPEEK- sulfonated poly(ether ether ketone) membranes came across as another possible viable option. Luo et al. [ 37 ] findings are summarized in Table 3.2. 24 Chapter 3. Research lines and challenges Table 3.2 Comparison of parameters for N117, Nafion/SPEEK, and SPEEK membranes [37]. Parameter N117 Nafion/SPEEK SPEEK Thickness [mm] 175 100 100 IEC [mmol/g] 0.91 1.67 1.8 Area specific resistance [W/cm2] 1.06 1.61 1.27 Permeability of V4+(10−7cm2/min) 36.55 1.93 2.43 Coulombic efficiency [%]a93.8 97.0 97.1 Voltage efficiency [%]a90.7 85.3 87.3 Energy efficiency [%]a85.0 83.4 84.8 aValues for efficiency parameters are given for comparison purposes. Since SPEEK and Nafion/SPEEK are thinner than N117, this makes the comparison difficult: • area specific resistance of SPEEK is a lot higher yet the membrane is thinner, assigning a disadvantage to SPEEK •The voltage efficiency drops considerably as a result on Nafion/SPEEK •Coulombic efficiency is higher (IEC is higher and decreased ion V4+despite being thinner) An important point to make SPEEK is not very stable in the oxidizing environment of V5+ positive half cell [37]. Coating it with Nafion would solve this issue. In the author’s opinion, [ 37 ], the lower cost and compromise between its capabilities puts the Nafion/SPEEK membrane at an advantage. Overall energy efficiency is still worse. However long-term stability may be problematic because of the layered structure of the composite membrane: interfacial delamination might occur after prolonged exposure to the electrolytes coupled with mechanical stress during operation. [37] [56] Unlike what occurred to fluorinated polymers, research on SPEEK membranes continued. Mai et Al. [ 38 ] tested SPEEK membranes to different degrees of sulfonation .They found out that as sulfonation increases, diffusivity of V4+increases. [38] The V4+ ion permeability was still lower than through N115, which is thicker (127 vs 85 micrometers). Refer to Figure 3.5- this figure puts forward how the selectivity is greater in SPEEK60 than in Nafion115. SPEEK 40 has sulfonation degree of 0.73 and yields the highest coulombic efficiency of the three.It overall yields a higher energy efficiency even if voltage efficiency is lower. 87.5% compared to 84.7% N115.[38] 3.1 Membrane development 25 Figure 3.5 Selectivity in permeability of V4+ions and protons [38]. Stability is very promising in this experiment as well, as can be depicted from Figure ??. Chen et Al. [ 12 ] synthesized other types are SPTK (Sulfonated poly arylene thioether ketone) and SPTKK (arylene thioether ketone ketone) membranes, which showed high coulombic efficiency. See Table 3.3. Table 3.3 Proton conductivity and VO 2+ permeability of SPTK, SPTKK, and Nafion 117 membranes at room temperature [12]. Membrane Proton Conductivity (×10−2S cm−1) VO2+ Permeability (m2s−1) SPTK 1.05 1.2×10−13 SPTKK 1.36 3.1×10−13 Nafion 117 2.03 4.9×10−12 3.1.3 Anion-exchange membranes One of the membrane types left to discuss is anion-exchange membranes. They could be promising technology as they would circumvent cross-mixing of all vanadium cations. In this way, efficiencies close to 100% could be achieved. Nevertheless, since it is an anion-exchange membrane, as well as reducing permeation of vanadium ions, it also hinders the conduction of protons. This latter effect sacrifices voltage efficiency. [56] Selemion AMV and New Selemion Type 2 were investigated through sulfonation. The effect it had was reduction of water transport, and hence, reduced capacity loss as an advantage. However, it is already known that energy efficiency is also reduced due to this effect. [39] Interesting studies on custom-made poly ether sulfone [ 75 ], poly(ether sulfone ketone) [ 27 ] were conducted. About the poly(ether sulfone ketone) membrane, compared to N117 membrane it was reported that [27]: •Similar chemical stability in solutions with V5+ions •Higher energy efficiency (88.3 vs. 82.9%) due to: –Similar voltage efficiencies (91.6 vs. 90.6% for N117) 26 Chapter 3. Research lines and challenges –Increased coulombic efficiency (96.4% vs. 91.5% for N117) Note that if the energy efficiency is higher, maintaining the voltage efficiencies at a similar or even greater level, it implies that the membrane is more selective. On the contrary, the poly(ether sulfone) membrane did report a lower voltage efficiency, even if the overall energy efficiency was higher due to a greatly increased coulombic efficiencythe membrane was thus less selective. It reduced the permeation of vanadium ions but also limited proton conduction. Increased amounts of amine functional groups resulted in more enhanced ion exchange capabilities. [75] Also chemical stability and higher water transport are additional issues. They are being studied and this research may lead to the construction of more efficient and complex hybrid membranes leading to the mitigation of these problems. 3.1.4 Amphoteric ion-exchange membranes An interesting field of study is that of amphoteric membranes. They consist of membranes with different types of functional groups that may aid or prevent ion transport-cationic and anionic. For example, in [ 74 ], varying layers of polyanion (PPDA) and polycation (PSS) were included to the surface of Nafion N117 membranes. The nomenclature followed is , where n is the number of multilayers. This model was tested for n=3,5 and 7. Refer to Figure 3.6. Figure 3.6 Layer by layer self-assembly of Nafion-[PDDA −PSS]n[74]. This configuration embodied reduced vanadium ion permeability (by 6-10 times less when compared to pure Nafion). Moreover, proton conductivity was slightly lower than for pure Nafionaround 50mScm−1, for n=3,5,7. [74] This configuration was also stable for 50 cycles, which is promising for a first design. In terms of efficiencies, higher values were achieved when compared to N117 membrane’ efficiencies. [74] 3.1.5 Nonionic separators On another note, nonionic separators have been evaluated for use in redox flow batteries. Interest is drawn towards them because of their cheap price and the fact that they are already in use in other battery types (such as lithium ion or nickel cadmium batteries). For this application, pore size was found to be too big. As a result, the big degree of vanadium ion permeability was unacceptable. [56] Nevertheless, inserting modifications that reduce the pore size of this material makes it apt to use it as a redox flow battery membrane. For example, Daramic/Amberlite400CG later crosslinked with DVB is a composite membrane that has yielded an energy efficiency of 83%with a coulombic efficiency of 95% and a voltage efficiency of 85%. [40] 3.2 Electrolyte Chemistry and Optimization 27 A charge/discharge cycling test for 8000 h (1650cycles) was performed on another Daramic/Amberlite 400CG/DVB composite membrane; this membrane had an energy efficiency of 72.7% and showed excellent stability over the whole cycling time. [13] On another note, membrane swelling is a clear obstacle to elaborate commercially feasible membranes. [ 56 ]. So is the need for materials that resist a very acidic environment, since otherwise V5+ions cannot be kept and will eventually undergo precipitation to oxoanions. 3.1.6 Membrane-less Apart from all of these studies, some scientists have pursued the idea of going membrane-less- Refer to Figure 3.7. [9] However, aqueous electrolytes are limited for output potential, namely water electrolysis happens at 1.23 V in standard condition, which limits the energy density.[9] Introducing inorganic nanoparticles has been proved effective to modify the porous membranes.[ 9 ] Figure 3.7 Schematic of reactant flow within the HBLFB [9]. 3.2 Electrolyte Chemistry and Optimization Different redox couples have been tested out over time searching for the most efficient chemistry. Most industrialized redox flow batteries are vanadium redox flow batteries (VRFBs) and zincbromine flow battery (ZBFBs). Nevertheless more species have been investigated seeking the best candidates that will provide higher energy efficiency, that is, high values for both; coulombic and voltage efficiencies. This section will highlight some of them. 3.2.1 Simple inorganic ions All-vanadium redox flow battery First proposed by Skyllas-Kazacos, it only employs vanadium in both catholyte and anolyte so contamination due to crossover is not an obstacle. Therefore, a long system life can be achievedapprox 5-20 years. [47] Fast electrochemical kinetics with a short response time is a plus. [53] As disadvantages, it is worth noting that: •Strong corrosive strength •Less energy density than lithium-ion batteries, for example Iron-chromium This configuration was primarily developed by NASA [ 17 ], delivering a cell voltage of 1.18V. As per the reactions Equation 3.1 shows the cathodic one and Equation 3.2 shows the anodic one. 34 Chapter 4. Case study: Economic feasibility Figure 4.3 Tariff 6.1 TD [4]. Hereinafter the following approximate values are used after considering the above inputs: •0.2 EUR/kWh for peak hours •0.1 EUR/kWh for off-peak hours These two values have been estimated from data in P1 and P5. 4.2 Main characteristics of system proposed The goal of this study is to examine the economical feasibility of a realistic system in a lifelike environment. The environment has already been picked (Spain and its tariffs) and now it’s time to define the system. The purpose of a realistic system further complicates the study, since not that many big-scale industrial systems have been implemented. Research carried out reaches the conclusions that: • Many investigations (e.g. membrane, ions selection, electrode selection) have been carried out at a small scale •Regarding the few industrial systems, two main issues arise: – Detailed information on the system is not readily available, which is required to carry out the study – Not many systems have been checked in terms of stability . This point turns out to be crucial as it is only a long-term analysis that can ensure the economic feasibility of the solution Considering these factors, finally the approach is to pick a real-life system, whose stability has been already proven over a substantial period of time. Obviously the available information and data must be as much as possible, so that the conclusions drawn are mostly based on facts and the system’s characteristics as opposed to assumptions. The chosen system is a 450-kW system which was installed in 1996 in Tatsumi Substation of the Kansai Electric Power Co. Ltd. Refer to [ 66 ] and [ 65 ]. This battery is an all-vanadium one. A block diagram of the system is shown below in Figure 4.4 4.2 Main characteristics of system proposed 35 Figure 4.4 450-kW system concerned figure [66]. It is noticeable there are additional systems that support a safe connection to the network, a transformer and AC/DC converterjoint by a power line. Signal lines are added to account for the pump system, protection equipment and watching equipment. The analysis presented in this Section will estimate the costs of all the subsystems joint by power lines as well as the pump system. The readily available data is presented and calculated when not directly given through the article. The 450kW battery is made from battery cell stacks of 18.8 kW each, with the following characteristics: Table 4.2 20-kW battery system specification [66]. DC output 18.8 kW DC voltage 75.2 V DC current 250 A Area of electrode 5000 cm2 Cell stack 60 cells (series) Electrolyte Vanadium 1M Sulfuric acid solution Therefore, the battery system is built by piling up stacks of this sort in series and parallel. Specifically, one module is achieved by piling up cell stacks in a 4P x 2S configuration, meaning that one module obeys: VDC =2×75.2V=150.4V(4.1) IDC =4×250A=1000A(4.2) In turn, modules are piled up in series of three, so that: VDC =3×150.4V=451.2V≈450V(4.3) These calculations exemplify the fact that piling up cell stacks in series increases the voltage, whereas piling them up in parallel increases the current. Thus, the readily available specifications of the whole system are: 36 Chapter 4. Case study: Economic feasibility Table 4.3 450-kW battery system specification. DC output 450 kW Battery capacity 900 kWh (450 kW x 2 hours) DC voltage 450 V (av.) DC current 1000 A (av.) Cell module 1 module: Cell stack 4P x 2S Series of 3 modules Electrolyte Vanadium 1M Sulfuric acid solution AC/DC converter PWM control The current density of current collectors is 50 mA cm2[66]. On another note, as mentioned before, the stability of the system is mandatory. The 20-kW battery system cycle life test presents the following results in Figure 4.5: Figure 4.5 20-kW battery system cycle life test [66]. There is no substantial downfall through 12000 cycles of the three key efficiencies (coulombic, voltage, energy). This test was carried out in half an hour cycles. This proves the stability of a single stack. The efficiencies of the 450-kW battery system were estimated to be: Table 4.4 Result of 450-kW battery system Charge/Discharge test. Means of operation Constant current Charge/Discharge current 1000 ATest items Coulombic efficiency 96.7 % Voltage efficiency 85.1% Battery efficiency Energy efficiency 82.3% 4.3 Methodology 4.3.1 Summary The methodology proposed is developed by Moore et Al. [ 42 ] and inspired by a hierarchical step procedure devised by Douglas [14]. Douglas in [14] proposes a methodology of several steps that include heuristics for the design of a chemical system. This method is designed in an additive 4.3 Methodology 37 manner, so that if economic infeasibility of the system is detected in one of steps (so called levels), the procedure is stopped and the conclusion of infeasibility reached. The same philosophy is used in Moore’s approach. Since the costs can be classified into three areas [42]: •Costs that scale in proportion to the power capacity •Costs that scale in proportion to the energy capacity •Costs that do not scale with size (the control system and balance of plant) The method proposed by Moore studies six different levels [42]: •Level 1. Input information for the VRB •Level 2. Input-output analysis •Level 3. Power capacity considerations •Level 4. Energy capacity considerations •Level 5. Control system and balance of plant •Level 6. Total capital investment estimate 4.3.2 Level 1: Input information for the VRB 4.3.3 Level 2: Input- Output Analysis Level two aims to obtain the maximum economic potential of the VRB. The philosophy behind is for it to be the difference between the product value and the raw material cost, not taking into account processing costs. The equation that embodies this concept is given below: EP2= (ED× e kW hD −Ec× e kW hc )×cycles year (4.4) Ecand EDare defined by: Ec=Esc ξc (4.5) Ed=Esc ×ξD(4.6) The number of cycles per year is the variable to be plotted in the x-axis, while the economic potential will be in the y-axis. For simplicity, it is assumed that ξc=ξd . As the cycle efficiency is, as its name implies, the efficiency of the entire cycle, it is estimated to be the product of te following efficiencies: •Energy efficiency •Transformer Efficiency •PCS efficiency The energy efficiency is retrieved from the experiments carried out in the actual system and it is 0.823. Refer to Table 4.4. On the other hand, the transformer and PCS efficiencies are estimated to be 0.995 and 0.97 respectively as taken from [44]. In this way: ξc=ξD=0.891 (4.7) since it is assumed that: 38 Chapter 4. Case study: Economic feasibility ξc×ξD=ηenergy ×ηtrans ×ηPCS (4.8) The battery capacity is taken to be 900 kWh as indicated in Table 4.3. Also, it is induced from Section 4.1 that: e kW hD =0.2(4.9) e kW hC =0.1(4.10) Therefore, the following curve for the Level 2 Economic Potential (EP2) is obtained. Figure 4.6 Level 2 Economic Potential. A cycle is considered when both charging and discharging have occurred. Since it takes 2 hours for its mid-cycle to complete, it is estimated that the maximum achievable economic potential is given when 6 cycles occur per day, that is, the battery is always charging and discharging-never stops working. In a year this is 2190 cycles. Following the equation that the Figure 4.6 obeys, the maximum achievable economic potential is 130149 euros, that is approximately 130ke. 4.3.4 Level 3: Power Capacity Considerations The next step to follow is accounting for the costs of the power capacity considerations. The materials and equipment that will be taken into account, are therefore: •The heat exchangers •The pumps •The cells themselves • The PCS that converts electricity in AC form (as it’s given from the network) to DC, so as to charge the battery, and viceversa, when discharging takes place 4.3 Methodology 39 Heat exchangers To start with, let us examine whether heat exchangers are required in this design. The amount of heat that may increase the system’s temperature, and thus, need dissipation is the energy lost due to inefficiencies. This energy is calculated as follows: qC=P ξC ×(1−ξC)(4.11) qD=P×(1−ξD)(4.12) It turns out to be logical that qC≥qD, as more heat is dissipated during charging. In this case: qC=26.68kW (4.13) qD=25.18kW (4.14) One assumption made within this procedure is that this heat is shared equitatively between the anode and cathode solution with the temperature change of the vanadium ion solutions dependent on the flow rate of the vanadium solution though the stack. [ 42 ] To calculate this flow rate, the moles of vanadium ions oxidized per second is divided by the molarity of the vanadium ions in the solution: FM=Is×Nc×Ns F×Cv (4.15) In Equation 4.15 [42]: •FM stands for the minimum flow rate assuming all flowing vanadium ions are oxidized, that is full conversion in l/s •ISstands for the current through a stack in A •NCstands for the number of cells per stack •NSstands for the number of stacks in the battery •Fstands for Faraday’s constant in C/mol •CVstands for the concentration of vanadium ions in mol/L Breaking Equation down: •IS=250Aas shown in Table 4.2 •NC=60 as shown in Table 4.2 •NS=24 since there is a total of twenty-four modules in the 3 x (4P x 2S) configuration •CVis 1M as exemplified in article [66] Henceforth, FM=3.73L s However, not all vanadium ions that flow through the cells are oxidized. Transport loss is a phenomenon that occurs in RFBs and it is intrinsically associated with the incomplete conversion of all posible V ions flowing through the battery. Being aware of this effect, it will be assumed that the minimum flow rate represents 10% of the greater bulk rate, so that: FA=FM XV,P (4.16) , where FA represents the actual flow rate of vanadium ion solution and XV,P the fraction of vanadium ions converted per pass. 40 Chapter 4. Case study: Economic feasibility Once the flow rate has been obtained, the change in temperature of the solution is calculated by [42]: ∆T=q 2×Cp×FA (4.17) Note that here in Equation 4.17 the specific heat capacity of the electrolyte is required. It is retrieved to be 4.2J gK from [ 50 ]. Refer to Table 4.5 in [ 50 ], where the heat capacity of commercial electrolyte (CE) is given. This same article establishes that the specific heat capacity is practically insensitive to the molarity of the vanadium solution. Additionally, in other studies such as [59] this same value is used as reference for the specific heat capacity. Table 4.5 Composition and thermodynamic properties for commercial electrolyte and mixed-acid electrolyte. Heat capacity [JmL−1K−1] Thermal coefficient [mVK−1] Equilibrium voltage Thermoneutral voltage CE 4.2 -1.16 1.38 1.72 MAE 3.9 -0.80 1.44 1.67 Using the value of heat dissipated while the battery charges, the result yields 0.085K . This difference is considered negligible and thus, no heat exchanger will be considered in the design. The literature considers heat exchangers when the temperature rise during the pass of fluid through the stacks is less than 25◦C. Pumps The correct design for pumps is mandatory as it is a main component in the solution, as it has previously been explained. The shaft power of the pumps can be calculated using the following Equations: W=FA×∆p εi (4.18) εi= (1−0.12F−0.27 A)(1−µ0.8)(4.19) εi is the intrinsic efficiency of the pump and depends on the viscosity of the electrolyte, µ , and the actual flow rate of vanadium ion solution, FA. The viscosity of the electrolyte is not easy to estimate, as data is not readily available. Besides,the concentration of sulphide-bisulphide ions is not given in Tokuda’s articles [ 66 ] nor [ 65 ]. However, its effect on the calculation of efficiency is not that relevant. Studies in [ 49 ] estimate a viscosity of 8cP =8mPa ·sfor 2M Vanadium solution. On another note, refer to Figure 4.7 from [76]. [76] carries out an intensive study on the effect of SOC on the electrolyte viscosity. In this paper the value of the viscosity ranges from 1.2mPA ·sto 1.4mPA ·s. 4.3 Methodology 41 Figure 4.7 Variations in the electrolyte viscosity with SOC in the positive and negative half-cells. Regardless of these observations, this difference in value only shows a difference of one percentage point on the efficiency. Therefore, a value of 1.3cP is taken from [ 76 ] as [ 49 ] deals with supersaturated vanadium electrolyte. On another note, a value of of 2 bar for the pressure drop is assumed, that is 2·105Pa. Lastly, it’s relevant to note that in these equations FAmust be in m3 s. In turn this yields: ε=70.5% (4.20) W=10.6kW (4.21) In order to calculate the costs of this pump, A guide to chemical engineering design and economics [ 68 ] by Ulrich will be used. Specifically refer to Appendix B.2, where the graphs are attached. In this way, a nickel alloy centrifugal pump and a ratio dollar-euro of 0.94e $is assumed, giving: Cp =0.94e $×7000$ =6580e(4.22) FM=3.5(4.23) Fp=1(4.24) FBM =7(4.25) Therefore, for one pump: CBM =46060e(4.26) Cell stack components Here, the following costs will be contemplated: •Ion-exchange membrane 42 Chapter 4. Case study: Economic feasibility •Current collectors •Carbon felt electrodes In order to derive an estimation of these costs, the cost per squared meter is considered for each component. Each 20kW battery system has 60 cells in series, and 24 of these systems are arranged to form the 450 kW battery. So, firstly keeping in mind the 20-kW battery system: •60 ion-exchange membranes must be considered (one per cell) • 61 current collectors must be considered, due to adjacent cells sharing one common current collector •120 carbon felt electrodes (two per cell) Multiplying this by 24, as there are 24 of these system yields the number of elements to be considered, that is: •1440 ion-exchange membranes must be considered (one per cell) • 1464 current collectors must be considered, due to adjacent cells sharing one common current collector •2880 carbon felt electrodes (two per cell) Since [ 66 ] explicitly accounts for the area of one electrode to be 5000cm2 , it will be assumed that the area of one IEM and one current collector is the same due to the way RFBs are designed. On another note, for this type of equipment, CBM =1.2×Cp. [68] In summary, the following data is used: Table 4.6 Calculation of cost of items in cell stacks. Element Cost in eper m2Number m2required Total cost in e(Cp) Total cost in e(CBM) IEM 282 1440 720 203040 243648 Current collectors 75.2 1464 732 55046.4 66055.7 Carbon felt 47 2880 1440 67680 81216 The costs per squared meter come from [42] and [41]. Additional equipment This subsection considers the cost of additional equipment also part of the system: Power Conversion System (PCS), power transformer, and breakers, contacts and cabling. The costs are considered per kW and obtained from [ 41 ]. These are estimates from the The Electric Power Research Institute (EPRI) projected onto 2025. Carrying out the calculations results in: Table 4.7 Calculation of PCS and associated items’ costs. Element Cost in eper kW Total cost in e(Cp) Total cost in eCBM PCS 188 84600 101520 Transformer 33.84 15228 18273.6 Breakers, contacts and cabling 26.45 11903.22 14283.86 Total PCS and associated items 134077.46 Annualized cost Starting from Level 3 Analysis, the capital costs are annualized. This includes annual expenses, the cost of capital, and equipment depreciation. The annual expenses are directly proportional to fixed capital as indicated in Table 4.8. 4.3 Methodology 43 Table 4.8 Annual expenses Proportional to Fixed Capital. Capital-related cost item Fractions of fixed capital Maintenance and repairs 0.06 Operating supplies 0.01 Overhead, etc. 0.03 Taxes and insurance 0.03 General 0.01 Total 0.14 The cost of capital considers the required return on investment for a given capital outlay. The required return on investment will vary among companies, and it is assumed for this case to be 10%. Annualized interest on invested capital expressed as a fraction of the initial capital investment is calculated with Equation 4.27. fRI =n(i(1+i)n (1+i)n−1)−1 n(4.27) The service life of the components n, is assumed to be 15 years. The annualized cost of components may be calculated by: AC =Ccost(0.14 +fRI +fD) == Ccost ×0.27 (4.28) In summary, these are the results: Table 4.9 Level 3 Capital Costs. Item Total annualized cost Stacks 106 124 EUR Pumps 25 008 EUR PCS and associated items 36 399 EUR Total 167 531 EUR Unluckily this value is greater than the maximum yearly economic potential reached in Level 2, i.e 167 531 EUR is greater than the previously calculated 130 000 EUR. Therefore, already in Level 3 this battery system is unfeasible. Refer to Figure 4.8. 5 Conclusion Th is work puts forward the issues and challenges in the development and investigation of Redox Flow Batteries. This pioneering technology opens a way to deal with problems that require a battery system to be solved, such as load leveling or frequency stabilization. Above all, it appears as promising as renewable energy sources must enter the market to counteract the effects of burning fossil fuels as energy source (the latter are a cause of pollution and ceaselessly used cause irreversible climate change effects). Continuity of supply is never guaranteed using renewable resources of energy though. In this way, batteries provide a mean of storing energy and employing it when a renewable energy source is not producing it (e.g. at night in case of PV panels). As it was already discussed, many battery models exist to compete against redox flow batteries. In general terms, these battery types consist of more mature technology and have a higher energy density, making them more desirable for a vast range of applications. Efforts have been made to improve Redox Flow Batteries’ performance. Most interesting research lines have been drawn towards improving energy efficiency (both coulombic efficiency and voltage efficiency), yet few research lines investigate on the stability of a battery system. Even if improved efficiency result may seem more appealing from a research point of view, the latter point is crucial to make the systems commercially and industrially viable. However, a considerable amount of time is required to prove long-term stability of a system. Most research lines then focus on improving one of the crucial parts that conform a redox flow battery to prove improved efficiency. This is the case of functional groups being added to electrodes, or the grand amount of treatments that diverse membrane types have been subjected to. Multiple configurations exist and most of them are alive research topics. However, all-vanadium batteries seem to be the most commercially available technology, with big industrial systems being installed. This is due to no cross-mixing contamination being possible, as both electrolytes contain the same species. There is still big room for improving the energy density of these systems as well as the costs associated to it. While the achievements in membrane and electrolyte research are indeed remarkable, it is important to acknowledge that several significant challenges remain. The integration of these developments into practical RFB systems requires careful consideration of system design, engineering, and optimization. Moreover, safety aspects and environmental impact assessments will need to be addressed rigorously to ensure the technology’s viability and acceptance on a large scale. 51 52 Chapter 5. Conclusion In order to check viability when considering a real-life application, an analysis was carried out on a technically feasible and stable system already installed in Japan. The results on this analysis prove that they are not economically feasible. In countries like the United States, there is a higher difference between the electricity cost at peak and off-peak hours (one being around ten times greater than the other, as oppossed to Spain, where it is only double). Therefore, installing a redox flow battery system in this network may result in a positive result for feasibility. On the other hand, the percentage distribution of capital costs shows that the membrane and electrolyte solutions are the most expensive items of an all-vanadium system. Definitely, more or less percentage cost is attributed to the electrolyte volume depending on the battery duration. When comparing the estimated cost distribution of the case study with other studies carried out by other authors, the outcomes are therefore consistent and coherent with the existing trends. Appendix A Tariff 6.1 TD Tarifa Óptima 6.1TD Muchas gracias por interesarte en nuestras ofertas. Contratando la Tarifa Óptima disfrutarás de un precio ajustado y estable para que ahorres y no tengas que preocuparte de las oscilaciones del precio de la energía. 7% Descuento indefinido sobre el término de energía. Precios Precios con descuentos incluidos Término de Energía (€/kWh) P1 P2 P3 P4 P5 P6 0,195622 0,175803 0,152045 0,141073 0,116955 0,130619 La potencia a facturar se calculará según corresponda a su tarifa de acceso y potencia contratada, aplicando la facturación por maxímetro o por curva de carga cuarto horaria. La energía será calculada mediante curva de carga horaria. Las variaciones que se den en los componentes regulados que le son de aplicación, así como nuevos que puedan aparecer, se trasladarán al cliente, tanto al alza como a la baja, sobre el precio base. También se trasladará al cliente cualquier otro cargo que pudiera hacer la distribuidora como energía reactiva, alquiler de equipos de medida, derechos, etc. Asimismo, los términos de potencia y energía no se actualizan con el valor del IPC. Horario: Los periodos horarios corresponderán en todo momento con los de su tarifa de acceso, mes y ubicación geográfica. Para mayor detalle puede consultar los calendarios en el siguiente enlace, dependiendo de si el suministro está ubicado en Península, Canarias o Baleares: https://www.endesa.com/es/comunicacion/cartera/cliente-empresa Con permanencia durante 1 año: En caso de resolución unilateral del contrato por parte del Cliente antes de iniciada la primera prórroga, podrá ser aplicada una penalización. Para los suministros con tarifas de acceso 6.1TD, esta cantidad será la equivalente al 10% del precio del Término de Energía del contrato o del precio del Periodo 1 en caso de existir varios periodos, multiplicado por la energía pendiente de suministro (esta energía pendiente de suministro será calculada como el producto del número de días pendientes del Contrato desde la fecha de baja del suministro por la máxima potencia contratada y por un factor 7,1). Esta energía se facturará al precio de la energía, sin descuentos, que tenga el contrato en el momento de la resolución del mismo. En caso de suministro en alta tensión medido en baja tensión, se aplicarán los mismos recargos por pérdidas que aplique la distribuidora. El exceso de reactiva se facturará al cliente del mismo modo y con el mismo precio que para su tarifa de acceso. Precio sin impuestos: Impuesto de electricidad (luz - 0,5%) y al importe resultante de lo anterior se aplican los impuestos indirectos: IVA (21%) en Península y Baleares, IGIC en Canarias (0% consumo viviendas con potencia menor o igual a 10kw, resto 3%) e IPSI en Ceuta y Melilla (1%). Tras la publicación del RDL 11/2022 y del RDL 20/2022, temporalmente y si se dieran las circunstancias indicadas en la normativa, se le aplicaría un IVA del 5%. Tarifa de acceso: 6.1TD hasta 450kW. Oferta válida para contrataciones desde el 07/09/2023 hasta el 15/09/2023. Precio base del Término de Potencia sin descuento €/kW y año €/kW y mes €/kW y día P1 23,365215 1,947101 0,064014 P2 20,241178 1,686765 0,055455 P3 10,727582 0,893965 0,029391 P4 8,960662 0,746722 0,024550 P5 2,308583 0,192382 0,006325 P6 1,548958 0,129080 0,004244 El precio base del Término de Energía sin descuento (€/kWh) P1 P2 P3 P4 P5 P6 0,210346 0,189036 0,163489 0,151691 0,125758 0,140450 Término de Potencia €/kW y mes Término de Potencia P1: 23,365215 P2: 20,241178 P3: 10,727582 P4: 8,960662 P5: 2,308583 P6: 1,548958 €/kW y año 1,947101 1,686765 0,893965 0,746722 0,192382 0,129080 €/kW y mes Figure A.1 Tariff 6.1 TD by Endesa [4]. 53 Appendix B Graphs for cost estimation by Ulrich B.1 Heat exchangers Figure B.1 Purchased equipment cost for heat exchangers [68]. 55 56 Chapter B. Graphs for cost estimation by Ulrich Figure B.2 Pressure factors for high pressure heat exchangers [68]. Figure B.3 Bare module factors for heat exchangers [68]. B.2 Pumps 57 B.2 Pumps Figure B.4 Purchased equipment cost for pumps [68]. Figure B.5 Bare module factors for pumps [68]. 58 Chapter B. Graphs for cost estimation by Ulrich B.3 Tanks Figure B.6 Purchased equipment costs for storage vessels [68]. List of Figures 1.1 Global primary energy consumption by source [51] 1 2.1 Working principle of RFBs [67] 5 2.2 Construction of a cell stack [67] 6 2.3 60 kWhequivalent tanks [7] 7 2.4 1 MWhequivalent tanks [48] 7 2.5 The schematic of the conventional carbon-fiber based electrode and the proposed dualscale porous carbon electrodes [79] 10 2.6 Addition of surface oxygen functional groups to electrodes [30] 11 2.7 Example of enhanced mechanism due to catalytic treatment [30] 12 2.8 Membrane classification [34] 13 2.9 All-vanadium system [35] 14 2.10 Cross section of single cell [35] 15 2.11 Diagram of RFB cell [72] 15 2.12 Illustration of the full ionic equations of the VRB during the charge [8] 16 3.1 Membrane degradation results [31] 19 3.2 UV/Vis absorption of a Nafion membrane after cycling in a vanadium redox flow battery (A) and after sonication of the same membrane in deionized water (B). Spectrum (B) matches that of pure VO2+ solutions. [70] 21 3.3 Schematic of a Nafion membrane modified with SiO2by sol-gel synthesis [73] 22 3.4 Comparison of permeabilities of different vanadium ion species through pure Nafion and Nafion/SiO2[73] 23 3.5 Selectivity in permeability of V4+ions and protons [38] 25 3.6 Layer by layer self-assembly of Nafion-[PDDA −PSS]n[74] 26 3.7 Schematic of reactant flow within the HBLFB [9] 27 3.8 Scheme of a kW-class hybrid Zn-Br system [53] 28 4.1 Tariff 2.0 TD, Evolution in time [45] 32 4.2 Percentages of prices of energy [19] 33 4.3 Tariff 6.1 TD [4] 34 4.4 450-kW system concerned figure [66] 35 4.5 20-kW battery system cycle life test [66] 36 4.6 Level 2 Economic Potential 38 4.7 Variations in the electrolyte viscosity with SOC in the positive and negative half-cells 41 4.8 Level 3 Economic Potential 44 4.9 Current price of vanadium pentoxide V2O5 [69] 45 59 66 Bibliography [41] Kendall Mongird, Vilayanur V. Viswanathan, Patrick J. Balducci, Md Jan E. Alam, Vanshika Fotedar, V S. Koritarov, and Boualem Hadjerioua, Energy storage technology and cost characterization report, (2019). [42] Mark Moore, Robert Counce, Jack Watson, T.A. Zawodzinski, and H. Kamath, A step-by-step design methodology for a base case vanadium redox-flow battery,46 (2012), 239–250. [43] Mark Alan Moore, nbsp; a base case design and capital cost analysis of an all vanadium redox-flow battery, Ph.D. thesis, Chancellor’s Honors Program Projects, 2013. [44] Elizabeth Hope Mouron, nbsp; the economic potential of the all-vanadium redox flow battery with a focus on state of charge, Ph.D. thesis, Chancellor’s Honors Program Projects, 2011. [45] OCU, El precio de la luz en septiembre,https://www.ocu.org/vivienda-y-energia/gas-luz/ informe/precio-luz, Jun 2023. [46] V.M. Ortiz-Martínez, L. Gómez-Coma, G. Pérez, A. Ortiz, and I. Ortiz, The roles of ionic liquids as new electrolytes in redox flow batteries, Separation and Purification Technology 252 (2020), 117436. [47] Feng Pan and Qing Wang, Redox species of redox flow batteries: A review, Molecules 20 (2015), no. 11, 20499–20517. [48] Sonal Patel, Grid-scale iron-chromium redox flow battery connected, POWER (2014). [49] Faizur Rahman, Optimization of supersaturated vanadium electrolyte for high energy density vanadium redox battery, 10 2000. [50] Danick Reynard, C.R. Dennison, Alberto Battistel, and Hubert H. Girault, Efficiency improvement of an all-vanadium redox flow battery by harvesting low-grade heat, Journal of Power Sources 390 (2018), 30–37. [51] Hannah Ritchie, Pablo Rosado, and Max Roser, Energy production and consumption, Our World in Data (2020), https://ourworldindata.org/energy-production-consumption. [52] David Rivero Bermúdez, Sistemas de almacenamiento de energía. baterías de flujo, (2021). [53] Eduardo Sanchez-Diez, Edgar Ventosa, Massimo Guarnieri, Andrea Trovò, Cristina Flox, Rebeca Marcilla, F. Soavi, Petr Mazúr, Estibaliz Aranzabe, and Raquel Ferret, Redox flow batteries: Status and perspective towards sustainable stationary energy storage, Journal of Power Sources 481 (2021), 228804. [54] Dominik Schulte, J Drillkensb, BSchulteb, and Dirk Uwe Sauer, Nafion hybrid membranes for use in redox flow batteries, ECS Transactions 25 (2010), 247–255. [55] Birgit Schwenzer, Soowhan Kim, M. Vijayakumar, Zhenguo Yang, and Jun Liu, Correlation of structural differences between nafion/polyaniline and nafion/polypyrrole composite membranes and observed transport properties, Journal of Membrane Science 372 (2011), no. 1, 11–19. [56] Birgit Schwenzer, Jianlu Zhang, Soowhan Kim, Liyu Li, William Liu, and Zhenguo Yang, Membrane development for vanadium redox flow batteries, ChemSusChem 4(2011), 1388– 406. [57] Joan I. Senyk, Darrell D. Ebbing, and Larry K. Krannich, Study guide: General chemistry, ebbing, Houghton Mifflin, 1996. Bibliography 67 [58] Maria Skyllas-Kazacos, George Kazacos, Grace Poon, and Hugh Verseema, Recent advances with unsw vanadium-based redox flow batteries, International Journal of Energy Research 34 (2010), no. 2, 182–189. [59] Kevin Spellman, Kendrick Stiles, and Ian Little, nbsp; economic report on vanadium redox flow battery with optimization of flow rate, Ph.D. thesis, Chancellor’s Honors Program Projects, 2013. [60] Theresa Sukkar and Maria Skyllas-Kazacos, Water transfer behaviour across cation exchange membranes in the vanadium redox battery, Journal of Membrane Science 222 (2003), no. 1, 235–247. [61] Theresa Sukkar and Maria Skyllas-Kazacos, Membrane stability studies for vanadium redox cell applications, Journal of Applied Electrochemistry 34 (2004), no. 2, 137–145. [62] B. Sun and M. Skyllas-Kazacos, Modification of graphite electrode materials for vanadium redox flow battery application—i. thermal treatment, Electrochimica Acta 37 (1992), no. 7, 1253–1260. [63] Chenxi Sun, Jian Chen, Huamin Zhang, Xi Han, and Qingtao Luo, Investigations on transfer of water and vanadium ions across nafion membrane in an operating vanadium redox flow battery, Journal of Power Sources 195 (2010), no. 3, 890–897. [64] Michael W Thielke, Gengyu Tian, and Ana Jorge Sobrido, Sustainable electrodes for the next generation of redox flow batteries, Journal of Physics: Materials 5(2022), no. 2, 024004. [65] N. Tokuda, M. Furuya, Y. Kikuoko, Y. Tsutui, T. Kumamoto, and T. Kanno, Development of a redox flow (rf) battery for energy storage, Proceedings of the Power Conversion Conference- Osaka 2002 (Cat. No.02TH8579), vol. 3, 2002, pp. 1144–1149 vol.3. [66] N. Tokuda, Toru Kanno, T. Hara, T. Shigematsu, Y. Tsutsui, A. Ikeuchi, T. Itou, and T. Kumamoto, Development of redox flow battery system, SEI Technical Review 50 (2000), 88–94. [67] Nobuyuki Tokuda, Takahiro Kumamoto, Toshio Shigematsu, Hiroshige Deguchi, Takefumi Ito, Noriyasu Yoshikawa, and Takushi Hara, Development of a redox flow battery system, SUMITOMO ELECTRIC TECHNICAL REVIEW-ENGLISH EDITION- (1998), 88–94. [68] Gael D. Ulrich, P. T. Vasudevan, and Gael D. Ulrich, A guide to chemical engineering process design and economics: A practical guide, CRC, 2003. [69] VanadiumPrice.com, https://vanadiumprice.com/. [70] M. Vijayakumar, M.S. Bhuvaneswari, P. Nachimuthu, Birgit Schwenzer, Soowhan Kim, Zhenguo Yang, Jun Liu, Gordon L. Graff, S. Thevuthasan, and Jianzhi Hu, Spectroscopic investigations of the fouling process on nafion membranes in vanadium redox flow batteries, Journal of Membrane Science 366 (2011), no. 1, 325–334. [71] Vilayanur Viswanathan, Alasdair Crawford, David Stephenson, Soowhan Kim, Wei Wang, Bin Li, Greg Coffey, Ed Thomsen, Gordon Graff, Patrick Balducci, Michael Kintner-Meyer, and Vincent Sprenkle, Cost and performance model for redox flow batteries, Journal of Power Sources 247 (2014), 1040–1051. [72] Francisco Javier Vázquez Galván, Redox flow batteries: From vanadium to earth abundant organic molecules (quinones), Ph.D. thesis, 2019. 68 Bibliography [73] Jingyu Xi, Zenghua Wu, Xinping Qiu, and Liquan Chen, Nafion/sio2 hybrid membrane for vanadium redox flow battery, Journal of Power Sources 166 (2007), no. 2, 531–536. [74] Jingyu Xi, Zenghua Wu, Xiangguo Teng, Yongtao Zhao, Liquan Chen, and Xinping Qiu, Selfassembled polyelectrolyte multilayer modified nafion membrane with suppressed vanadium ion crossover for vanadium redox flow batteries, J. Mater. Chem. 18 (2008), 1232–1238. [75] Dongbo Xing, Shouhai Zhang, Chunxiang Yin, Bengui Zhang, and Xigao Jian, Effect of amination agent on the properties of quaternized poly(phthalazinone ether sulfone) anion exchange membrane for vanadium redox flow battery application, Journal of Membrane Science 354 (2010), no. 1, 68–73. [76] Q. Xu, T.S. Zhao, and C. Zhang, Effects of soc-dependent electrolyte viscosity on performance of vanadium redox flow batteries, Applied Energy 130 (2014), 139–147. [77] Lu Yue, Weishan Li, Fengqiang Sun, Lingzhi Zhao, and Lidan Xing, Highly hydroxylated carbon fibres as electrode materials of all-vanadium redox flow battery, Carbon 48 (2010), no. 11, 3079–3090. [78] Mengqi Zhang, Mark Moore, J. S. Watson, Thomas A. Zawodzinski, and Robert M. Counce, Capital cost sensitivity analysis of an all-vanadium redox-flow battery, Journal of The Electrochemical Society 159 (2012), no. 8, A1183. [79] X.L. Zhou, Y.K. Zeng, X.B. Zhu, L. Wei, and T.S. Zhao, A high-performance dual-scale porous electrode for vanadium redox flow batteries, Journal of Power Sources 325 (2016), 329–336. 1. Introducción La energía es la materia prima que sustenta la vida humana y contribuye al desarrollo de la industrialización y estilo de vida. A medida que la digitalización se impone junto con el desarrollo tecnológico y la electrificación, se registra un mayor consumo de energía a lo largo de los años. El petróleo y el carbón son las fuentes que reportan mayor consumo de energía primaria. El aumento del consumo de energía en el siglo XIX es bastante insignificante si se compara con el aumento registrado en el siglo XX. El hecho de que el petróleo y el carbón sean las fuentes principales pone de manifiesto varios obstáculos [52]: • Los precios de los combustibles fósiles están subiendo mucho debido a la escasez de esta fuente y a la complicación que conlleva extraer energía de ellos. • Los combustibles fósiles contribuyen fuertemente a la contaminación y al cambio climático, debido a la emisión de gases de efecto invernadero que se produce al quemarlos para obtener energía de los mismos •La demanda de energía es casi exponencialmente creciente Esto hace que la investigación se centre en encontrar fuentes de energía alternativas. Existen múltiples fuentes de energía renovable que posteriormente pueden almacenarse mediante sistemas de baterías, tales como, pero no limitadas a la energía solar, la energía eólica, la energía hidráulica, la energía del agua de mar, la energía geotérmica y la energía de la biomasa y los residuos. Muchos sistemas de almacenamiento de energía, entre ellos múltiples tipos de baterías coexisten en el mercado de almacenamiento energético. Entre las baterías, se pueden destacar las de plomoácido o las de ión-litio. Sin embargo, lo que diferencia a las baterías de flujo redox de otros tipos de baterías es: • Su capacidad para aumentar/disminuir físicamente la potencia y la capacidad de forma independiente (son dos magnitudes desacopladas) •Bajo nivel de autodescarga • Nivel de degradación prometedoramente bajo debido a los principios químicos a los que obedecen. Aún queda mucho por investigar sobre este tipo de sistema de baterías para que sea viable y útil en las aplicaciones requeridas, como se verá más adelante. Sin embargo, son las baterías de flujo redox son candidatas a contribuir a la descarbonización del sector eléctrico. 69 2. Principio de funcionamiento de las Baterías de Flujo Redox Co mo sugiere su nombre, las RFB funcionan gracias a las reacciones químicas redox que se producen en ellas. Para empezar, describiremos su principio de funcionamiento con ayuda de la Figura 7. Figure 7 Principio de funcionamiento de RFBs [67]. Aunque el esquema presentado en la Figura 7 muestra un sistema totalmente de vanadio, las partes o componentes son equivalentes independientemente de los compuestos que participen en las reacciones. Como se aprecia en este esquema los componentes clave que deben estar presentes son: •Electrolito positivo (generalmente también conocido como catolito) •Electrolito negativo (generalmente también conocido como anolito) •Dos tanques para almacenar ambos electrolitos •Dos bombas integradas en un sistema de control de caudal •Membrana •Dos electrodos A partir de la Figura 7,se puede representar que los electrolitos se bombean a través de una pila donde se puede poner o quitar energía de los fluidos cambiando sus estados de oxidación. Dado que 71 72 la carga y la descarga se producen en las pilas, estas últimas están directamente relacionadas con la cantidad de energía que el sistema es capaz de suministrar. Este punto puede enfatizarse afirmando que el diseño es modular [35]. •Si más energía almacenada se requiere, el tamaño de los depósitos puede aumentarse, •mientras que si hay un mayor requerimiento de potencia, más pilas puedn añadirse La Figura 8 muestra una representación visual de un conjunto de celdas. Como puede verse, las células se apilan tanto en serie como en paralelo. Apilarlas en serie aumenta la tensión, mientras que apilarlas en paralelo permite obtener una mayor corriente. Figure 8 Construcción de un conjunto de celdas [67]. En relación a los tanques, contemplarlos más voluminiosos permite almacenar mayor cantidad de electrolito. Las especies activas en este proceso no son otras que los electrolitos. Por tanto, la capacidad de la batería resulta ser directamente proporcional a la cantidad de especies activas que son capaces de someterse a reacciones electroquímicas produciendo energía. Algunos materiales comunes empleados pueden ser [15]: 1. Polietileno de alta densidad (HDPE) 2. Polipropileno (PP)- tanques poly 3. Plástico reforzado con fibra de vidrio (PRFV) - depósitos de fibra de vidrio 4. Acero inoxidable Los tanques de PRFV son artesanales y costosos de fabricar. Además los tanques de polipropileno tienen alta resistencia a la corrosión, facilitando el mantenimiento. La tendencia del mercado es hacia los tanques de polipropileno. En cuanto a los electrodos, algunos de los materiales más comunes pueden ser, por ejemplo: - Grafito - Fieltro de carbono - Cerámica conductora - Óxidos metálicos mixtos Como se desprende de la lista de ejemplos anterior, los electrodos deben ser materiales altamente conductores, ya que es a través de ellos por donde fluyen los electrones para cargar y descargar la batería. Asimismo, 73 cuanto mayor sea la superficie de los electrodos, mejor será la reacción electroquímica, ya que proporcionan "los lugares de reacción". Acogen más material activo mejorando la velocidad de transferencia de electrones. Un punto importante a destacar es que los electrodos son "inertes"- no participan en la reacción [ 10 ]. Para mejorar la eficiencia de los electrodos, se enumeran diferentes tratamientos investigados: térmico, por ácido, grabado químico, dopaje de heteroátomos, etc. [ 21 ] En cuanto a la membrana, una membrana selectiva de iones ocupa su lugar entre los electrodos positivo y negativo en una pila de flujo Redox. Esta membrana permite que los electrones fluyan a través de ella a la vez que evita la mezcla de los dos electrolitos (anolito y catolito). La membrana selectiva de iones desempeña un papel crucial en el mantenimiento de la separación de cargas y en la prevención de la contaminación cruzada de los electrolitos. Hay elementos adicionales como los captadores de corriente, sistemas de bombeo y posibles intercambiadores de calor, entre otros. El principio electroquímico se ejemplifica a continuación mediante las ecuaciones químicas de la batería de vanadio. [11] VO2++2H++e−−−⇀ ↽−− VO2++H2O(5.1) V2+−−⇀ ↽−− V3++e−(5.2) 3. Líneas de investigación y retos Du rante este capítulo se abordan las principales líneas de investigación y hallazgos, que se resumen muy brevemente a continuación. 3.1. Desarrollo de membranas El estudio se centra en las membranas utilizadas en baterías RFB acuosas. Se descubren y se somete a pruebas diferentes tipos de membranas: •Membranas de intercambio de cationes •Membranas de intercambio de aniones •Membranas microporosas La membrana de Nafion es la más popular y se debe principalmente a su capacidad de soportan las condiciones acídicas requeridas. El principal problema es la falta de testeo en cuanto a la estabilidad a largo plazo. La mayoría de las investigaciones se centran en contribuir a mejorar la eficiencia del sistema. Diferentes factores se comprueban en las investigaciones, tales como la resistencia específica o la concentración de iones. No obstante, los hallazgos dependen del tipo de membrana. Por ejemplo, en el caso de la membrana S-Radel, tiene lugar mayor degradación de la misma a mayor concentración de iones de vanadio. Se puede ver Figura 9 c) 0.1M de iones de vanadio y d) 1.7M de iones de vanadio. 75 82 Chapter 5. Figure 13 Sistema de 450 kW [66]. Se observa que hay sistemas adicionales que dan soporte a una conexión segura a la red, un transformador y un convertidor CA/CC unidos por una línea de alimentación. Se añaden líneas de señalización para dar cuenta del sistema de bombeo, el equipo de protección y el equipo de vigilancia. El análisis presentado en este apartado estimará los costes de todos los subsistemas unidos por líneas de potencia, así como del sistema de bombeo. Se presentan y calculan los datos disponibles cuando no se facilitan directamente en el artículo. La batería de 450 kW está formada por pilas de celdas de batería de 18,8 kW cada una, con las siguientes características [66]: Table 1 Especificación del sistema de 20 kW [66]. Salida DC 18.8 kW Voltaje DC 75.2 V Intensidad DC 250 A Área del electrodo 5000 cm2 Conjunto de celdas 60 celdas (serie) Electrolito Vanadio 1M Solución de ácido sulfúrico Por tanto, el sistema de baterías se construye apilando pilas de este tipo en serie y en paralelo. En concreto, un módulo se consigue apilando pilas de celdas en una configuración 4P x 2S, lo que significa que un módulo obedece: VDC =2×75.2V=150.4V(5.3) IDC =4×250A=1000A(5.4) Los módulos se apilan en series de tres: VDC =3×150.4V=451.2V≈450V(5.5) Estos cálculos ejemplifican el hecho de que apilar pilas de células en serie aumenta la tensión, mientras que apilarlas en paralelo aumenta la corriente. Así, las especificaciones fácilmente disponibles de todo el sistema son: 83 Table 2 Especificación del sistema de 450 kW. Salida DC 450 kW Capacidad de la batería 900 kWh (450 kW x 2 horas) Voltaje DC 450 V (av.) Intensidad DC 1000 A (av.) Módulo de celda 1 módulo: Conjunto de celdas 4P x 2S Series de 3 módulos Electrolito Vanadio 1M Solución de ácido sulfúrico AC/DC Convertidor Control PWM La densidad de corriente es 50 mA cm2[66]. Por otra parte, como ya se ha mencionado, la estabilidad del sistema es obligatoria. La prueba de ciclo de vida del sistema de baterías de 20 kW presenta los siguientes resultados en la Figura 14: Figure 14 Pruebas de ciclos de vida [66]. A lo largo de 12.000 ciclos no se observa ninguna disminución sustancial de las tres eficiencias clave (coulómbica, tensión y energía). Esta prueba se realizó en ciclos de media hora. En relación a la metodología: La metodología propuesta ha sido desarrollada por Moore et Al. e inspirada en un procedimiento jerárquico por pasos ideado por Douglas. Douglas [ 14 ] propone una metodología de varios pasos que incluyen heurísticas para el diseño de un sistema químico. Este método se diseña de forma aditiva de manera que si se detecta la inviabilidad económica del sistema en uno de los pasos (los denominados niveles) el procedimiento se detiene y se llega a la conclusión de la inviabilidad. La misma filosofía se utiliza en el enfoque de Moore [ 42 ]. Los costes pueden clasificarse en tres áreas: [14] •Costes proporcionales a la capacidad energética •Costes proporcionales a la capacidad energética •Costes que no varían con el tamaño (el sistema de control y el equilibrio de la planta). El método propuesto por Moore [42] estudia seis niveles diferentes: •Nivel 1. Información de entrada para el VRB •Nivel 2. Análisis entrada-salida 84 Chapter 5. •Nivel 3. Consideraciones sobre la capacidad de potencia •Nivel 4. Consideraciones sobre la capacidad energética Durante el procedimiento se anualizan los costes. A continuación, se presentan los resultados de acuerdo a los niveles estudiados: Figure 15 Potencial económico de Nivel 2. Desafortunadamente, ya en el Nivel 3 este sistema deja de ser económicamente viable. Figure 16 Potencial económico de nivel 3. De ella se desprende que el punto en el que la batería se vuelve viable es a partir del ciclo número 3413. 85 No obstante, se calculan los costes del resto de equipamiento relacionado (Ánálisis de nivel 4) y se obtiene la siguiente distribución de costes, que se compara con otras distribuciones derivándose las siguientes conclusiones: Figure 17 Distribución de costes para el caso de estudio. Como se muestra a continuación, el coste distribuido del electrolito oscila entre el 8% para un cuarto de hora de duración de la batería y el 43% para cuatro horas de duración de la batería. En el estudio que se ha realizado, la bateríaes de 2 horas y el porcentaje de coste atribuido es del 22%, situándose dentro de este rango. Del mismo modo, el coste de los depósitos es en ambos análisis el coste porcentual más bajo de todo el sistema. Por tanto, se puede concluir que las pilas y el electrolito de la batería suponen la mayor parte del coste en ambos casos. El PCS, los depósitos y las bombas contribuyen al coste total, pero no suponen una gran cantidad de dinero en la inversión. El porcentaje atribuido al pentóxido de vanadio es proporcional a la duración de la batería. Además, también cabe señalar que la pila de celdas constituye en el análisis de Viswanathan [ 71 ] un porcentaje del 60% y el 73% se debe al separador en el sistema de baterías de 0,25 MWh. Del mismo modo, para el sistema de 4 MWh, el 38% del coste pertenece a la pila de celdas y el 71% a al separador. En el caso de estudio presentado a lo largo del análisis, las cifras siguen siendo similares, con un El 45% se atribuye a la pila de celdas y el 62,3% se debe al IEM. Estas conclusiones comparten descubrimientos sobre la distribución de costes de los componentes entre diferentes sistemas. 5. Conclusión Es te trabajo expone los problemas y retos que plantea el desarrollo y la investigación de las baterías de flujo redox. Esta tecnología pionera abre una vía para resolver problemas que requieren un sistema de baterías, como la nivelación de la carga o la estabilización de la frecuencia. Sobre todo, parece prometedora y necesaria dada la casi obligatoria entrada en el mercado de fuentes de energía renovables para contrarrestar los efectos de la quema de combustibles fósiles como fuente de energía (estos últimos son causa de contaminación y su uso incesante provoca efectos irreversibles en el cambio climático). Sin embargo, la continuidad del suministro nunca está garantizada utilizando recursos energéticos renovables. De este modo, las baterías proporcionan un medio de almacenar energía y emplearla cuando una fuente de energía renovable no la está produciendo (por ejemplo, por la noche en el caso de los paneles fotovoltaicos). Como ya se ha comentado, existen muchos modelos de baterías que compiten con las baterías de flujo redox. En términos generales, estos tipos de baterías consisten en una tecnología más madura y tienen una mayor densidad energética, lo que las hace más deseables para una amplia gama de aplicaciones. Se han realizado esfuerzos para mejorar el rendimiento de las baterías de flujo redox. La mayoría de las líneas de investigación interesantes se han dirigido a mejorar la eficiencia energética (tanto la eficiencia coulómbica como la eficiencia de voltaje), pero pocas líneas de investigación investigan la estabilidad de un sistema de baterías. Aunque la mejora de la eficiencia pueda parecer más atractiva desde el punto de vista de la investigación, este último punto es crucial para que los sistemas sean comercial e industrialmente viables. Sin embargo, se necesita un tiempo considerable para demostrar la estabilidad a largo plazo de un sistema. La mayoría de las líneas de investigación se centran entonces en mejorar una de las partes cruciales que conforman una batería de flujo redox para demostrar una mayor eficiencia. Es el caso de los grupos funcionales que se añaden a los electrodos o la gran cantidad de tratamientos a los que se han sometido diversos tipos de membranas. Existen múltiples configuraciones y la mayoría de ellas son temas vivos de investigación. Sin embargo, las baterías de todo vanadio parecen ser la tecnología más comercializada, con la instalación de grandes sistemas industriales. Esto se debe a que no es posible la contaminación por mezcla cruzada, ya que ambos electrolitos contienen las mismas especies. Aún queda mucho por hacer para mejorar la densidad energética de estos sistemas, así como los costes asociados. Para comprobar este último punto, se ha realizado un análisis de un sistema técnicamente viable y estable que ya se está instalando en Japón. Los resultados de este análisis demuestran que no 87 88 Chapter 5. son económicamente viables. En países como Estados Unidos, existe una mayor diferencia entre el coste de la electricidad en horas punta y en horas valle (una es unas diez veces mayor que la otra, a diferencia de España, donde es sólo el doble). Por lo tanto, la instalación de un sistema de baterías de flujo redox en esta red puede suponer un resultado positivo para la viabilidad. Por otra parte, la distribución porcentual de los costes de capital muestra que la membrana y las soluciones electrolíticas son los elementos más caros de un sistema totalmente de vanadio. Definitivamente, el volumen del electrolito representa un porcentaje mayor o menor de los costes en función de la duración de la batería.