On-site renewable electricity generation is widely used in rural areas, where the electrical grid is weak, or it does not exist. Standalone energy systems usually include a PV (photovoltaic) generator and battery storage. Wind turbines are used in places with high wind speed during the whole year. In many cases, a hybrid system including PV, a fossil fuel genset (diesel or gasoline) and battery storage can be the optimal solution. Auxiliary components as the battery charge controller and inverter are required. The charge controller avoids the overcharge of the battery, preventing from premature failure. It also controls the battery discharge, preventing the over-discharge. Generally, the load is AC (alternating current) type while the battery is DC (direct current), so an inverter (DC/AC converter) is needed. Inverters used in standalone systems include battery over-discharge protection. If the system includes a fossil fuel genset, then a battery charger (rectifier, AC/DC converter) is used to charge the batteries. In many cases, the inverter and the battery charger are a single device called inverter-charger or bi-directional inverter [1
]. Regarding storage in standalone systems, lead-acid batteries have been widely used [2
], although Li-ion batteries can be competitive in some cases [3
] as their cost have been reduced recently. Standalone systems can be DC or AC coupled. DC coupled are usual in low power systems, while AC coupled are used in middle and large power systems.
The sizing of a standalone renewable energy system is a critical process for the stability of the supply. It includes the choice of the suitable components from both technical and economic points of view. Choosing the economically optimal battery is one of the results that can be obtained from a simulation and optimization process. However, the simulation and optimization of hybrid standalone energy systems is a complex task, due to the high number of variables, the time-dependence and the non-linearity of the components [4
]. Heuristic optimization techniques, such as genetic algorithms, particle swarm optimization, etc. have been used by different researchers [5
]. Multiple optimization objectives have been used [4
], but the most common are the minimization of the cost and the minimization of emissions [7
]. Usually, a mono-objective optimization is performed, where the unique objective is the minimization of the net present cost (NPC) or the Levelized Cost of Energy (LCOE). Specific software tools are used for the simulation and optimization of hybrid standalone systems [4
The mono-objective economical optimization is usually done by simulating the performance of different combinations of components and control strategies, looking for the combination of lowest NPC or LCOE. After the simulation of each combination, if it meets all the constraints, NPC is calculated [5
], including the acquisition cost of all the components, the replacement cost at the end of the components lifetime and the O&M (operation and maintenance) cost.
The battery total cost (including its replacement during the system lifetime) is frequently the highest cost in the NPC of the hybrid system. Therefore, in optimization processes for hybrid standalone systems, the accurate estimation of the battery lifetime is one of the most important issues. Great errors in the prediction of battery lifetime would lead to great errors in the estimation of the NPC, making the optimization process unreliable.
The aging factors of lead-acid batteries are charge and discharge rates, charge (Ah) throughput, time between full charges, time at low state of charge (SOC), and partial cycling. Several researchers have analyzed them [9
]. Classical models widely used by researchers to estimate the battery lifetime are the “equivalent full cycles model” and the “rainflow model” [11
]. The first one counts the charge (Ah throughput) cycled by the battery since the start of its lifetime; when this value reaches the charge, the battery can cycle (considering the cycle life shown in the manufacturer datasheet, obtained in standard tests), the end of the battery lifetime is reached. The second model includes the effect of the depth of discharge (DOD). Nevertheless, real operating conditions (current rate, temperature, DOD, SOC, etc.) are commonly different from the laboratory conditions of the cycles shown by the manufacturer datasheet, so a great error in lifetime prediction can be expected. In some of these cases, especially in full PV systems, the lifetime estimation can be two or three times greater than real lifetime [11
]. A much more accurate lead-acid aging model (and also more complex and with higher computational effort) is described by Schiffer et al. [12
], called “weighted Ah
throughput model”. It is used by iHOGA software [13
]. This model includes Shepherd model [14
] to calculate the battery voltage and Lander model [15
] for the estimation of corrosion depending on voltage and temperature. The model is based on the application of weighting factors for the charge throughput of the battery to estimate the lost capacity (considering the different stress factors for cycling and corrosion).
Regarding Li-ion batteries [16
], they have higher cycle life, energy density and energy efficiency, and lower maintenance compared to lead-acid batteries. LiFePO4
type is the most used in off-grid systems. The most significant aging external factors of Li-ion batteries are temperature, charge and discharge rate, and DOD [19
]. In optimization of hybrid standalone systems, Li-ion cell level aging models [20
] are commonly used, due to their simplicity. Electrochemical models are usually very complex, even the most simplified ones [19
], thus they cannot be used in optimization due to excessive calculation times [23
]. “Calendar” aging occurs when a battery is not being used while “cycle” aging occurs when the battery is under charge or discharge current [23
]. Cycle aging is affected by the total charge (Ah) throughput from the start of battery lifetime, the current, the ambient temperature, and the SOC. Calendar aging main factors are temperature and SOC [20
]. Arrhenius kinetic-based aging models [20
] are the most used cycle aging models. For example, Wang et al. [24
] model was obtained by doing many accelerated cycling tests to commercial LiFePO4
cells, obtaining a capacity fade model which takes into account the Ah throughput and temperature for different charge/discharge rates. Calendar aging has been modelled by different researchers [20
]. For example, Petit et al. [25
] use an expression based on Arrhenius law, considering temperature, time, and SOC. In this work, researchers consider a condition to switch between calendar and cycle mode, assuming that cycle aging only occurs when the battery is in charge mode and the current is above a given threshold current. A recent study [26
] proposed to use the model of Schiffer et al. [12
] for lead-acid batteries and, for Li-ion batteries, the model of Wang et al. [24
] combined with the calendar aging model of Petit et al. [25
Lifetime aside, there are other characteristics of Li-ion batteries that outperform lead-acid ones, such as lower maintenance costs, higher cycling efficiency, and greater allowed depth of discharge. Against these advantages there is the fact that currently, its acquisition price is higher. Therefore, the question is to determine in which cases it is profitable to pay its highest price. In the context of standalone energy systems, this has been addressed for specific cases in several studies, mainly located on geographical latitudes with very high solar resource. Limiting the review to recent studies, the best results are equally divided between Li-ion and lead-acid batteries, i.e., currently it is a match virtually tied. In [3
], the simulation of a hybrid microgrid located in Colombia obtained lower NPC and LCOE with Li-ion batteries. In [27
], the redesign of an existing rural microgrid located in a small island in Thailand was addressed, proposing and comparing three hybrid solutions with Li-ion batteries and one with lead-acid battery. The lowest cost was obtained with Li-ion batteries, although at the expense of higher fuel consumption. In [28
], the case study is located in Tanzania and in the home solar system, Li-ion batteries obtained a lower LCOE. In [29
], the case corresponds to a small DC load located in India and the study resulted in the use of a Li-ion battery. The four case studies of [30
] are also located in India, where for both types of battery, the proposed systems included the same generation power. Therefore, the effect of battery type on system size was not considered. In [31
], a case study in North Korea offered better economic results with lead-acid batteries, although close to those of Li-ion batteries. In [32
], the case study is located in Malaysia and a lead-acid battery proved to be the most profitable option.
In addition to standalone PV systems, the combined use of PV generation and batteries is frequent in microgrids. A microgrid is a small energy distribution system with generation, storage, loads and usually an EMS (Energy Management System), which depending on its architecture and mode of operation can work connected to the network or disconnected from it [33
]. When it operates in islanded mode, the function and requirements of its batteries are similar to those of standalone systems. Moreover, stationary lead-acid or Li-ion batteries can operate as a buffer in those cases where it is necessary to charge the Li-ion batteries of electric vehicles taking advantage of the stored energy of PV origin [34
]. It is well known that in electric vehicles, the characteristics of Li-ion batteries are much more suitable than those of lead-acid ones but, in stationary systems, lead-acid batteries could be used if its technical and economic performance is considered adequate. Another characteristic that differentiates both types of batteries is the greater management complexity of Li-ion batteries, which require additional electronics and management algorithms [35
]. This can lead to incompatibilities between batteries from different manufacturers and raise the design and maintenance costs of the systems. The additional circuitry also drains small currents that increase self-discharge.
In summary, more than a controversy, currently the choice of one battery type or another depends on each case that is addressed, and the criteria applied to make the decision. In the absence of detailed studies, in practice general criteria or preferences are applied that do not ensure the best choice, since both technologies offer acceptable results in PV standalone systems. In fact, both lead-acid and Li-ion batteries are being installed in new standalone renewable energy systems. The general expectation is that Li-ion batteries will lower their prices and prevail. However, despite the increase towards mass production, the price reduction of Li-ion batteries is not guaranteed. Its characteristics make it the option for electric vehicles, which can dramatically increase its demand and hinder price reductions. For this reason, it cannot be ruled out that in standalone renewable energy systems, lead-acid batteries still have a future. As an alternative, the use of Li-ion batteries recycled for a second life is also being considered in standalone systems [37
] and other stationary applications [38
This work seeks to better understand the technical and, above all, economic implications of using Li-ion batteries instead of lead-acid in standalone renewable energy systems. With this objective, the study presents various novelties. First, it focuses on the simultaneous study of five cases whose main difference is the consumption profile. Unlike previous studies, the cases include agricultural, agro-food, and residential activities located on an intermediate geographical latitude and, therefore, with a solar resource of medium magnitude. Their locations are close together, with similar irradiation and temperature profiles. The results show that currently both types of batteries are competitive and, depending on the case (and therefore the demand profile), the most profitable choice is different. Second, it has been identified how the type of battery affects the sizing of the optimal system, including the capacity of battery, the power of PV generator and, in hybrid systems, fuel consumption. Thirdly, different behaviors are identified depending on the generation system, whether it is full PV or hybrid. Finally, the sensitivity analyses show, among other things, what reduction in the acquisition cost of Li-ion batteries would be necessary to displace lead-acid batteries as the economically optimal solution in all cases studied.
Five real cases with different consumption profiles have been studied, from an economic point of view, through simulations of standalone energy systems. The results show that in both 100% PV and PV-diesel hybrid systems, the use of lead-acid or Li-ion batteries results in different sizing of the economic optimum system. In other words, if the type of battery is changed, to achieve the economic optimum the entire system must be resized. This different size of the system elements affects the capacity of the battery, the power of the PV generator and, in hybrid systems, the fuel consumption. Consequently, for each case, the optimum system (and its NPC) is different for each type of battery. A lower NPC means a lower cost of the system throughout its lifetime, i.e., a higher profitability. Thus, of both optimum systems, with a lead-acid or Li-ion battery, the one that presents the lowest NPC should be considered the absolute optimum and the type of battery that it includes is the best option from an economic point of view.
In all cases of this study, Li-ion batteries have shown a longer lifetime, between +18% and +113% over lead-acid ones. This result is consistent with the estimate [26
] that lithium-ion battery lifetime can double that of lead-acid batteries. The cases in which lead-acid batteries have shown a shorter useful life are both homes (single-family home and second home), in accordance with the results of a previous study focused on their aging [11
]. Consequently, it is in them where the improvement in terms of lifetime is greater when changing to a Li-ion battery.
Furthermore, in all cases, the Li-ion battery has lower capacity than the lead-acid battery in their optimum systems. This result is in accordance with a previous study [2
] based on the charge and discharge characteristics of both battery types. Differences in capacity range between −7% and −92%. Both longer lifetime and less capacity, play in favor of its profitability. However, in the cases studied, it has not been found that the increase in lifetime or the reduction in capacity is enough for Li-ion batteries to displace lead-acid ones. Moreover, its use entails differences in other costs of the system.
The behaviors observed in PV and hybrid systems show similarities, but relevant differences. In PV systems the battery capacity reductions due to incorporating Li-ion are between −7% and −67%, while in hybrid systems they are between −21% and −92%. These battery capacity reductions are accompanied by changes in generation. In PV systems, the power of PV generator changes, resulting in large changes in the amount of excess energy. On the contrary, in hybrid systems the fuel consumption changes, while the power of the PV generator and the excess energy hardly vary. It should be noted that generation is not always higher when using Li-ion batteries, since in some of the study cases reductions in the power of the PV generator or in fuel consumption have been observed. Furthermore, in four of the five case studies, both in PV and hybrid systems, the amount of energy cycled in the battery has barely changed, despite the changes in battery capacity and generation.
With respect to the economic comparison of the optimum systems with both types of battery in each case, differences are found between the studied cases. In PV systems the differences in NPC range between −16% and + 22%, with the Li-ion battery being the most profitable in one case. This corresponds to the second home. In hybrid systems, the differences in NPC range from −7.2% to + 5.9%, with the Li-ion battery being the most profitable in two cases. These correspond to the two homes studied, the second home and the single-family home. In this sense, new studies are needed on how the demand profile influences the economic performance of both types of battery. To sum up, in terms of cost effectiveness, no single option has been found, as the two types of batteries currently compete closely in standalone renewable energy systems. This is in accordance with other recent studies, taken as a whole, refs. [3
] and it poses tougher competition than just three years ago, when Li-ion batteries were too expensive for power supply systems [17
Regarding differences between PV and hybrid systems, in every one of the study cases, in the hybrid system a more favorable bias has been obtained for the Li-ion battery than in the PV system.
One of the questions of interest is how possible variations in the price of Li-ion batteries are going to affect their profitability, displacing or not lead-acid batteries in standalone renewable energy systems. With the base prices of the study, which correspond to the current market, Li-ion batteries are the option with the lowest NPC in one of the five cases of the PV systems and in two of the hybrid systems, all of them being houses. The analysis has shown a strong sensitivity to the acquisition cost of Li-ion batteries, which with a variation of −21% would be the economically optimal choice (in terms of system NPC) in all hybrid systems studied and with a −36% in all PV systems.
Regarding how variations in the acquisition cost of Li-ion batteries affect the sizing of the economic optimum system, some changes have been observed. In both PV and hybrid systems, the lower the price of Li-ion batteries, the greater the battery capacity. In addition, in optimal PV systems, the lower the price of Li-ion batteries, the lower the power of the PV generator. On the contrary, in optimal hybrid systems, faced with reductions in the price of Li-ion batteries, the power of the PV generator increases slightly while the fuel consumption decreases strongly.
Sensitivity analysis on the acquisition cost of the PV generator elements has shown that its reduction favors profitability of Li-ion batteries in PV systems and practically has no effect over hybrid systems. In these, lower inflation in the fuel price favors profitability of Li-ion batteries. Both reductions in the cost of the source of energy (PV or diesel) cause slight reductions in battery capacity of the optimum systems.
This study has addressed the comparison of economic performance of lead-acid and Li-ion batteries in standalone renewable energy systems. For five real case studies, their energy supply has been simulated with a 100% PV system and with another PV-diesel hybrid. The economic optimum in terms of NPC have been sought and sensitivity analyses have been carried out on various parameters.
The results show that for each case, the use of a lead-acid or Li-ion battery results in different dimensions of the economic optimum system. Battery capacity, PV generator power and fuel consumption (in hybrid systems) are affected. As these changes also affect the NPC of the system, the economic studies must consider them.
Regarding the economic comparison, of the ten systems studied (five PV and five hybrids), in three of them the absolute optimum is obtained with Li-ion batteries and in seven with lead-acid batteries. In two of the latter, the differences in NPC are less than or equal to 2%. The differences in system NPC range from −16% to +22% in PV systems and from −7.2% to +5.9% in hybrid systems. On the one hand, the result of the comparison as a whole is virtually a tie. On the other hand, the five case studies produce different results. Furthermore, the results of each case keep similarities between both systems, PV and hybrid, although attenuated by the presence in the latter of a dispatchable generation (the diesel genset), i.e., currently, both lead-acid and Li-ion batteries compete for economic profitability in standalone renewable energy systems. The best choice depends on each case. Possible changes in the acquisition price can cause one or the other type to displace the other in terms of profitability, in a generalized way.
The economic performance of Li-ion batteries, compared to lead-acid ones, is relatively better in hybrid systems than in PV. Greater solar irradiation favors Li-ion batteries in PV systems, but harms them in hybrid systems. In these, it would be favored by lower inflation in fuel prices. Finally, a 21% reduction in the price of Li-ion batteries would make them the economically optimal option in all hybrid cases and if 36% is reached in all PV cases studied.
The limitations of the present study derive from the number of cases studied and their nearby locations. Sensitivity analysis to the magnitude of solar irradiation can be useful in this regard, but other effects of change of location are lacking, such as different temperature data.
More studies are needed to advance in the identification and quantification of the characteristics of each case that influence the economic performance of both types of battery. Models should also be studied to in depth determine the maximum lifetime of Li-ion batteries in standalone renewable energy systems, beyond the design life indicated by the manufacturers.