Evaluation of the Levelized Cost of Energy with New Costs for Concentrating Solar Power Tower Plants in Northern Chile and Impact of Green Taxes

. The world is undergoing an energy transformation, from a system based on fossil fuels to a system based on renewable energy, in order to reach the Paris Agreement target. Chile is no exception, and through the Asociación de Concentración Solar de Potencia (ACSP) has generated a new cost structure for tower-based Concentrated Solar Power (CSP) technology in the framework of the Long-Term Energy Plan, which updates the costs reported in previous studies. The changes experienced in this cost structure make it necessary to study the impact of this cost structure on the levelized cost of energy (LCOE). This work develops simulations of CSP tower plants analyzing their generation and the LCOE based on the new cost structure mentioned, for the north of Chile between the Arica and Parinacota region and the Coquimbo region. The most relevant results show a significant reduction in the LCOE, compared to studies from 2020, reaching a minimum LCOE value of 52,6 USD/MWh. On the other hand, the impact on the net social benefit of including the green tax to the merit order of each of the generating plants that make up the electric park until 2021 is studied, which yields a negative impact under current legislation.


Introduction
Chile is recognized worldwide for having privileged conditions for the availability of renewable resources.In terms of solar energy for electricity generation, northern Chile has a very low atmospheric attenuation compared to other locations in the world [1], in conjunction with the highest annual average direct solar radiation in the world.Especially in the Atacama Desert high direct solar radiation and clear skies exist most of the year [2].Given the above and that the prices of electricity production by CSP technology in the last bidding round for regulated customers in Chile went below 34 USD/MWh it makes sense to study where this type of technology could be located and what levelized cost of energy (LCOE) could be achieved in those locations.

Methodology
The Association for Concentrating Solar Power (ACSP) has published a report on CSP technology costs updated to the year 2021 so that these costs can be considered in long-term strategic planning of the country [3].Three configurations are studied depending on the target demand to be supplied: • CSP for Peak demand and part of the night, SM = 2,0 and TES=9,0 hrs.
• CSP for Peak and all-night demand, SM=2,5 and TES=13,0 hrs.For these configurations a tower plant with molten salts (sodium nitrate) as heat transfer fluid with a gross capacity of 111.2 MWe, turbogenerator efficiency of 42,5% and a heliostat of 144.4 m2 area are chosen.
Simulations are performed using the System Advisor Model (SAM) [4] software for different locations in Chile.The sites chosen are those which were submitted by Fraunhofer Chile Research to the Ministry of Energy with cloudiness index less than 5%.Thus, 63 sites have been qualified in the north of Chile, ranging from the Region of Arica and Parinacota to the Region of Coquimbo.

Cost Structure
Table 1 shows the cost structure for the different configurations analyzed.It was assumed that the operating cost structure considers a variable cost of 3,5 USD/MWh and a fixed capacity cost of 66 USD/kW per year.Additionally, according to soiling estimates for the Atacama Desert [5] it has been considered that it is necessary to clean the heliostat field 35 times every year, so that the reflectivity of the field does not drop below 0,90.This has an impact on water consumption associated with field cleaning.For this purpose, it has been considered that water has a price of 5,0 USD/m 3 and that consumption is 07 l/m 2 , of aperture area.Additionally, the water consumption associated with the power cycle was considered, which is calculated for each of the locations according to the result of each simulation by SAM at the same price indicated above.Finally, the parameters that are not mentioned in the table have the default number of the SAM Software itself.

LCOE Calculation
The levelized cost of energy (LCOE) represents the cost of generating a unit of electrical energy.This parameter is widely used to compare different technologies in terms of economic viability.Equation 1shows the mathematical formula to calculate the LCOE as reported by Hernandez-Moro and Martinez-Duart [6].The discount rate (t) is assumed to be 5 [7] and 7% [8] depending on the case, and lifetime (i) of the CSP tower is 30 years [8] [9]. (1)

Results
The main results obtained in this work are summarized below.The LCOE results obtained for the different configurations and different discount rates are shown in Figure 1 and 2; in blue the configuration for peak demand is shown (SM 1,7 with 6,0 hours of storage), in yellow the configuration for peak demand and all night (SM 2,0 with 9,0 hours of storage) is shown, and in orange the configuration for the CSP plant for peak and overnight demand (SM 2,5 and 13,0 hours of TES) is shown.
Figure 1 shows the values for a discount rate of 7%.The SM 2,5 configuration and 13,0-hour storage (red color), presents a minimum of 60,0 USD/MWh, a maximum of 91,7 USD/MWh, an average of 66,7 USD/MWh and a mode in the lower class of LCOE, that is between 60 and 62,5 USD/MWh.On the other hand, the CSP configuration with SM 2,0 and TES=9,0 hrs (yellow color) gives a minimum LCOE of 65,0 USD/MWh, a maximum of 99,6 USD/MWh and an average of 72,4 USD/MWh.Finally, the CSP configuration for peak demand (SM 1,7 and TES=6,0 hrs, blue color) at a rate of 7% presents a minimum of 69,0 USD/MWh, a maximum of 105,8 USD/MWh and an average of 77,0 USD/MWh.When evaluating the three configurations of the CSP plant at a discount rate of 5%, it is obtained that for the configuration with SM of 2,5, the minimum LCOE reaches 52,6 USD/MWh, with an average of 58,3 USD/MWh and a maximum of 80,1 USD/MWh, and its mode is in the class of 52,5 and 55 USD/MWh.On the other hand, the CSP configuration for peak demand and part of the night (SM 2,0 and TES=9,0 hrs, yellow color) evaluated at a rate of 5% obtains a minimum LCOE of 57,1 USD/MWh, an average LCOE of 63,5 USD/MWh and a maximum LCOE of USD 87,2 USD/MWh.Finally, the CSP configuration for peak demand (SM 1,7 and TES=6,0 hrs, blue color) at a rate of 5% presents a minimum of 60,8 USD/MWh, an average of 67,8 USD/MWh and a maximum LCOE of 92,9 USD /MWh.In all cases the LCOE is reduced by approximately 12% with respect to the 7% rate.In both figures it is possible to observe a downward trend for configurations with many hours of storage with a higher solar multiple and a lower discount rate.The results geographically for the 63 locations evaluated are shown in figure 3 and figure 4.This figure shows that as the solar multiple increases along with the hours of storage of the CSP plant the LCOE decreases so the color intensity of the points becomes stronger.Also, the locations within the Arica and Parinacota Region and the Antofagasta Region present lower LCOE compared to the Coquimbo Region.
Finally, a comparison was developed between the results obtained in this study with respect to the results of the work [5].The comparison shows a reduction in the LCOE, reaching values around 16% below the values previously reported.This difference represents a significant reduction in energy cost compared to the minimum LCOE reported in [5], near Copiapó using the same cost structure (without transmission costs).

Evaluation Tax CO2
As of July 2022, Chile's generating park is made up of 29 332 MW, where 57% of the installed capacity comes from renewable sources, while 43% is from traditional thermal sources, as indicated in Table 2.In Chile, carbon price instruments correspond to an environmental policy which emerges as a response to minimize environmental degradation at a lower social cost.These instruments respond to the damage generated, either by reducing emissions, offsetting them, or paying the price for the social cost.There are two types of instruments: carbon taxes and tradable emission permits.Carbon taxes can be a direct regulation on emissions, or indirect through the regulation of fossil fuels regulating the price through a charge on carbon content using emission factors.In both cases they oblige the generator to internalize the cost of the issues, but no limits are established.The emphasis is on price due to the social cost of polluting.In general terms, the rate of the tax should be equal to the marginal social damage of producing an additional unit of CO2eq.
In the case of Chile, in September 2014, Law 20 780 [12] was approved, establishing three environmental taxes or green taxes, see Table 3.

Source type Category
Taxes on emissions from fixed sources have been in effect since January 1, 2017, regarding the emission report and the first payment of recorded emissions was in 2018 and taxes emissions from establishments with boilers and/or turbines with power equal to or greater than 50 MWt.Today, Chile has a CO2 tax of 5 USD/tCO2.The value was determined based on the social price of carbon estimated by the Ministry of Social Development.In 2017, the ministry adjusted its methodology for calculating the social cost of carbon and currently places it at around 32 USD.
The current government program proposes to gradually increase the CO2 tax to 40 USD/tCO2 and expand its scope with different emission sources.
The green tax adopted in Chile is a downstream tax, that means it does not tax the carbon content of the fuels used in the production processes, so the objective is distorted [13].The determination of subparagraph 17 of article 8 of Law 20 780 excludes the tax from the calculation of the variable cost of each generating unit, which establishes a problem since from the environmental perspective it limits the theoretical compliance of the green tax for three main reasons (i) It does not establish a corrected value of the tax due to modifications by the social tax (ii) It does not alter the merit order dispatch (iii) It requires a mechanism for calculating and paying compensation for those generating units whose total unit cost (being the variable cost considered in dispatch plus the unit value of the tax) is greater than or equal to the marginal cost, which establishes the payment of compensation even to renewables generators, who do not emit pollutants subject to the tax .In other words, if the generating unit that determines the marginal cost must pay the green tax and its total costs exceed the marginal cost, the difference is transferred to the rest of the generating plants that extract energy from the system.Therefore, the tax does not charge 100% of the CO2 emissions to the emitting sources since it is being compensated by all the generating plants (including renewables technologies) [14].
During the first year of execution of Law 20 780, 191,2 MMUSD were collected, and 94,8% are from the generation companies' segment.Of the total collected money, the renewable generation plants paid 0,9% of the total compensation.For the year 2021, 183,8 MMUSD were cashed, of which 95,7% are associated with payments made by generating companies.
One proposal to be evaluated is to consider the green tax within the merit order of the generating plants, thus the thesis work of Rodrigo Bórquez [15] evaluated the social and environmental impact and analyzes the cost-benefit of four different scenarios under the PLP modeling to include the green tax in the marginal pricing of the system under a dry hydrology.Among the main results, the scenario of considering the green tax within the marginal cost obtained an incremental cost of the system close to MMUSD 120,1 (13% additional to the base scenario).The determination of the Net Social Benefit for the scenario of incorporation of the Green Tax in the marginal pricing system poses a negative result of MMUSD -126,8.Which means that the evaluation of this measure does not meet the economic viability criteria to proceed with its recommendation.However, under this scenario, it is possible to see benefits due to the reduction of local pollutants that constitute concrete evidence of the positive impact of this instrument, especially in provinces with a higher rate of impact due to atmospheric pollution, such as the provinces of Huasco and Conception.
Therefore, for the closure of coal-fired power plants to materialize within a period that does not affect the security of the National Electric System and without increasing the consumption of other fuels, it is necessary to identify enabling and conditioning measures to maintain the system economically efficient and adapted on demand, which also ensure the security and quality of the service.In this context, the study carried out by ACERA [16] identifies the need to develop new generation, storage, and transmission reinforcement capacity to carry out this replacement efficiently, maintaining the supply of demand at a minimum cost.In addition to the expected 10 GW of projects under construction and development expected to enter the SEN, an additional 8 GW of emerging technologies such as CSP, pumped hydro, and Battery Systems (BESS) are required to replace coal units between the period 2021 and 2025, and 1,1 GW is required in storage systems.On the other hand, regardless of the year of retirement of the coal-fired plants, the importance and convenience of developing solar energy near the sources of consumption and a strong wind development throughout the country is observed, concentrated in the Taltal area.Therefore, in case the system by 2025 is not adapted in terms of generation and additional storage, the system will increase generation based on fossil fuels (mainly diesel) reaching six times what is used in the year 2020, increasing the cost of the system close to 100 USD/MWh and consequently, increasing the level of total CO 2 emissions even higher than the case where not all coal plants are withdrawn.
Consequently, there is still nothing in sight that satisfies with the "polluter pays" principle of responsibility and that seeks to generate behavioral changes in those responsible for emissions and reduce their impacts.

Conclusion
This work has identified that the current LCOE has a reduction of approximately 16% compared to 2020 studies for the same location, which is mainly due to improved local knowledge about the technology and the entry of new players into the CSP market.
The second major finding was that including the green tax in the merit order does not produce the desired effect, since it produces a negative net social benefit, making the national electricity system more expensive, Unaligned with the principle of least-cost operation of the power supply system.A further study could assess the effect of measures such as: determining a percentage per hourly block of participation of long-term energy storage (LDES) technologies, such that it results in a positive net social benefit, fulfilling the "polluter pays" principle.

Figure 1 .
Figure 1.Histogram of LCOE values for different CSP plant configurations evaluated at a discount rate of 7%.

Figure 2 .
Figure 2. Histogram of LCOE values for CSP plant configurations evaluated at a discount rate of 5%.