In this paper, we show that concentrated solar power (CSP) with thermal storage is an economically attractive technology to achieve high solar penetration levels. To this end, we utilize an alternative framework of net levelized cost of electricity (net-LCOE), which captures the projected curtailment rate, to economically compare PV with batteries to CSP in power systems with high levels of solar penetration. We also explore potential benefits enabled by hybrid PV-CSP configuration, which are unattainable by either technology independently. We show the existence of a turning point in power systems with 20%–30% solar penetration levels, after which it is cheaper to supply electricity by CSP with thermal storage by up to 16%. It is demonstrated that storing excess PV electricity in low-cost thermal storage is valuable, enabling CSP configuration with solar multiple as low as 0.5 to operate with a high capacity factor. Furthermore, we show that converting green hydrogen to electricity using CSP power block is cost-effective when seasonal storage is required, thus enabling deep decarbonization based on solar electricity. In these high solar penetration levels, using CSP resulted in a reduction of up to 65% in the net-LCOE. The results may enable researchers and policymakers to evaluate CSP with thermal energy storage as a cost-effective solution for achieving high penetration levels of solar electricity.
As countries aspire to meet global obligations for limiting global mean temperature to well below 2
compared to the pre-industrial level, implied by the Paris agreement [1], economies should reach net-zero greenhouse gas emissions by 2050 [2]. Recent studies have shown the technical feasibility and paved pathways for achieving net-zero greenhouse gas emissions goal in this time frame [3], [4], [5], [6].
Photovoltaic (PV) solar cells are the most widespread means to harness solar energy, due to their low cost [7], and are expected to remain an important factor in the transition to a low-carbon economy. PV cells convert solar radiation directly to electricity and thus are extremely valuable for supplying demand when solar radiation is sufficient. However, since PV electricity generation and demand temporal profiles do not match well, overgeneration from PV cells may be significant in power systems with high solar penetration levels, depending on the grid’s flexibility to absorb this surplus electricity. One of the most common methods to overcome the overgeneration obstacle when PV penetration is low is by curtailment, i.e., dumping energy that cannot be consumed, which may become unacceptably expensive in power systems with high solar penetration levels. Energy storage systems such as battery energy storage systems (BESS) provide another method to overcome overgeneration. As utility-scale BESS prices had plummeted significantly in recent years [8], such systems are now used as flexible means to maintain power balance within power grids and to provide other grid services [9]. However, enabling high solar penetration levels using energy storage systems is still an expensive solution [10], [11].
In addition to solar cells, Concentrated Solar Power (CSP) plants, such as parabolic troughs and solar power tower plants, may be used to harness solar energy [12]. In contrast to PV cells, these technologies convert solar radiation to heat, which is used to generate electricity by a power block. The heat can be stored efficiently and used when needed in thermal energy storage (TES) tanks. Between these two technologies, solar power tower technology has the advantage of high working temperature as higher solar concentration can be achieved, which enables a higher efficiency Rankine cycle [13], [14]. In these plants, solar energy is concentrated using a field of mirrors (heliostats) to a receiver which is held on top of a central tower. The high energy concentration (500–1000 suns) heats a heat transfer fluid to a high temperature up to 565
, which is used by a steam turbine to generate electricity. To further increase efficiency, recent studies have suggested boosting working temperatures even higher with particle-driven CSP receivers [12], [15], [16]. High-efficiency TES may allow 15 hours or more of full-capacity operation, which enables 24 hours of energy generation [17].
Numerous hybrid approaches of CSP and PV have been suggested to leverage the advantages of each, namely the low cost of PV cells with the dispatchability that is offered by CSP plants [18]. Particularly, a configuration in which electric heaters are used to store PV surplus electricity in TES has been suggested as a readily available and feasible solution [19]. One of the core advantages of such technology is that the PV cells and TES can be operated independently in different locations, meaning that already existing PV cells can be used.
In addition, recent works on solar-aided hydrogen production give rise to the production of hydrogen with CSP heat and electricity [20], [21], [22], [23]. This is a promising opportunity to support high levels of solar penetration levels by storing energy for months [24].
One of the main barriers to CSP deployment is its high costs [25], which are commonly analyzed using the levelized cost of electricity (LCOE) merit. Although CSP LCOE is much higher than for PV cells, CSP with TES should be compared to PV cells combined with electricity storage to compare similar functionalities. As such, recent studies have compared CSP with TES to PV with BESS or TES, and inferred that the LCOE of CSP with TES is lower than PV plus BESS for storage duration greater than 4–6 hours [26], [27], [28]. This implies that for power systems with high PV penetration levels CSP with TES would be more economically viable than PV cells with BESS.
Nevertheless, the LCOE calculation captures the total produced electricity, and thus, does not reflect any electricity curtailment. Since high curtailment rates have been shown to be economic in power systems with high solar penetration levels [29], the LCOE merit is not suitable to determine the most cost-effective solar-based solution in such scenarios. To increase solar penetration level, it may be cheaper to incorporate some curtailment instead of installing expensive storage solutions, particularly for PV with BESS technology, in which PV cells are much cheaper than BESS per delivered power. On the other hand, when solar penetration level is high and excess electricity is abundant, energy storage may become the dominant factor in the collective cost of electricity generation and storage favoring the use of low-cost TES. These factors are not well represented by the conventional LCOE merit. Furthermore, these factors change for any given level of solar penetration; thus, there is no fixed value for a wide range of solar penetration. Showing the possible error of the LCOE measure, a recent study [30] evaluated the LCOE sensitivity to the capacity factor, a measure of the annual number of hours a technology has been used, showing that this sensitivity is high. For example, the corrected LCOE when considering the actual capacity factor value for a coal power plant during 2020 is
4 times higher than the anticipated conventional LCOE.
To overcome the flaws of the LCOE measure, other metrics have been suggested, including levelized avoided cost of electricity (LACE) [31], system profitability [32], cost of valued energy (COVE) [33], and system LCOE [34], [35], [36], [37]. In addition to a more thorough analysis of the total costs, these metrics incorporate expected revenues, which are dependent on the generation mix, fuel prices, and energy policies, and therefore, are complex and case-specific [32]. In contrast, another metric, the net-LCOE, similar to the conventional LCOE does not include revenues. It introduces a simple improvement compared to conventional LCOE by considering the curtailment caused by overgeneration and is calculated by net-LCOE
LCOE/(1–curtailment rate) [38], i.e. it captures only the useful produced electricity. The advantage of the net-LCOE metric is the combination of its simplicity on the one hand, and its ability to manifest supply and demand mismatch on the other. For example, it has been shown that increasing PV penetration in California from 10% to 20% annual solar energy penetration in a limited flexibility power grid would increase PV cell’s marginal net-LCOE from 6 ¢/kWh to 11 ¢/kWh while adding CSP with TES to the grid should decrease PV electricity curtailment, and thus, decrease its marginal cost significantly [39]. Thereby, the net-LCOE can be used to deduce a threshold of solar penetration from which PV cells become more expensive than alternative energy sources. Nevertheless, to the best of our knowledge, there is no study that used the net-LCOE merit to compare solar technologies, specifically PV cells with BESS to CSP with TES.
To fill this gap, in this work, we use the measure of net-LCOE to study the cost competitiveness of CSP with thermal energy storage compared to PV with battery systems. To this end, we propose a greedy algorithm, which is used to find minimum net-LCOE for a wide range of solar penetration levels. In contrast to previous studies, CSP capacity and configuration, and thermal energy storage capacity are all jointly optimized. We enable the use of a PV-CSP hybrid configuration, which utilizes to the fullest the low-cost electricity generated by PV cells and low-cost thermal energy storage. We also evaluate the benefit of CSP power block when it is used to convert green hydrogen into electricity, which is meaningful when seasonal storage is required to avoid high costs. Our analysis provides a framework for comparing electricity costs from solar technologies in power systems with high solar penetration levels. It emphasizes the significance of CSP with TES in power systems with high solar penetration levels. Overall, it is found that CSP with thermal energy storage has a positive economic impact on the marginal net-LCOE for a wide range of solar penetration levels. We also show that although a high solar multiple CSP configuration generally obtains lower LCOE [26], low solar multiple, as low as 0.5, achieves the lowest net-LCOE using a hybrid configuration. We find that CSP power block can be used to supply electricity using hydrogen as a solar fuel, which reduces the need for a hydrogen-fired combined cycle gas turbine (CCGT), thus providing a route for deep decarbonization based on solar technologies.
A greedy algorithm was developed to obtain the least cost marginal net-LCOE solar-based solution for each solar penetration level. The simulation was used to compare a CSP scenario, in which CSP with TES is considered, to a no-CSP scenario, in which CSP is not considered. In this way, we uncover the impact of CSP with TES on the marginal net-LCOE of solar power generation. Starting from zero solar penetration, for each additional 0.25% solar penetration three options were examined for the CSP
The marginal net-LCOE of increasing solar penetration levels in steps of 0.25% are shown for 2030 costs in Fig. 2 for low and high grid flexibilities and for the CSP and no-CSP scenarios. At low solar penetration levels, up to 10% and 20% for low and high flexibility, respectively, there is no overgeneration, and PV is the lowest-cost technology. Therefore, net-LCOE of solar electricity is constant and equal to the LCOE of PV. As the solar penetration increases so does the PV overgeneration.
In this work, we propose the net-LCOE framework to show that CSP with TES is an economically attractive technology to reach high solar penetration. This method, in contrast to the conventional LCOE method, captures the large curtailment rates accompanied by high solar penetration levels. We show that while PV cells with BESS achieve the minimal net-LCOE in low solar penetration levels, CSP with TES becomes cost-competitive afterward (
20%–30%) for the California case study. This result
Dror Miron: Conceptualization, Data curation, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Aviad Navon: Writing – review & editing. Yoash Levron: Supervision, Writing – review & editing. Juri Belikov: Visualization, Writing – review & editing. Carmel Rotschild: Conceptualization, Supervision, Writing – original draft, Writing – review & editing.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
The work of D. Miron was supported by the Israel Department of Energy (PhD fellowship) and by the Nancy and Stephen Grand Technion Energy Program (GTEP), Israel. The work of A. Navon was partially supported by the Israel Department of Energy (PhD fellowship). The work of Y. Levron was partially supported by the Israel Science Foundation (ISF) , grant No. 1227/18, and the Ministry of Energy, grant No. 22111051. The work of J. Belikov was partially supported by Estonian Research Council, Estonia
, , , ,
https://www.sciencedirect.com/science/article/abs/pii/S2352152X23018613
