Impact of renewables on the Peruvian electricity system

Impact of renewables on the Peruvian electricity system Hugo Fiestas-Chevez a , Juan Manuel Roldan-Fernandez b , * , Angel Luis Trigo-Garcia b , Manuel Burgos-Payan b a Department of Mechanical and Electrical Engineering, Universidad de Piura, Av. Ram´ on Mugica 131, 20009, Peru b Department of Electrical Engineering, Universidad de Sevilla, Camino de los Descubrimientos, 41092, Spain ARTICLE INFO Handling editor: Cecilia Maria Villas Bˆ oas de Almeida Keywords: Electricity markets Renewable energy Energy transition Renewable premium GHG emissions Social cost of carbon ABSTRACT Peru is committed by international agreements such as the Paris Agreement and the UN 2030 Agenda for Sustainable Development to reduce its Green House Gas (GHG) emissions. Although Peru began promoting power plant projects based on renewable energies in 2008, the institutional impulse seems to have ceased today. Nonetheless, adopting renewable energy sources not only aids in fulfilling these commitments but also reduces electricity prices, fostering a more competitive economy. This study aims to provide a thorough analysis and evaluation of the impact that integrating Non-Conventional Renewable Energy Resources (NCRER) has on Peru’s wholesale electricity market. Specifically, it focuses on the net total savings for the system, calculated as the difference between the reduction in energy costs traded and the cost of the renewable premium. To reach that goal a simplified yet analytical tool has been developed based on a multivariable regression of the real data of the market. The volume of natural gas not consumed and the volume of GHG emissions avoided will also be evaluated. The results show that based on the 2021 scenario, a NCRER share of 28.26 % would lead to a net saving for the final users of 316.04 MUSD in the annual energy traded in the market. The reduction of conventional thermal generation prevents the annual burning of 2.04 normal km 3 of natural gas and the emission of 4.99 million tons of CO 2-eq . To the authors’knowledge, this is the first time that the impact of NCRER production in the Peruvian electricity system has been quantitatively evaluated. 1. Introduction Peru, as a signatory country of the Paris Agreement, has the commitment to mitigate its GHG emissions so that global warming by 2030 does not exceed 2 ◦C above the pre-industrial level (UNFCC, 2015). To reach this goal, Peru initially committed to a National Determined Contribution of 30 % reduction, using the Business as Usual (BaU) scenario as a reference. This target was updated in 2020, shifting from relative values to an absolute maximum emission limit of 179 MTCO 2-eq (a 40 % reduction compared to the BaU) (Government of Peru, 2020). Most international authorities and policymakers agree that replacing fuel-based generation by renewables power plants is one of the most cost-effective and cost-competitive ways to reduce GHG emissions. As a result of the design and maintenance over time of global energy policies to promote the integration of renewables (feed-in tariffs and premium to renewables, among others), the cost of renewable technologies has been falling continuously, year after year. According to IRENA (2023), renewable power plants have experienced an unprecedented cost reduction during the period 2010–2021. During that decade, the global weighted average Levelized Cost of Energy (LCOE) of newly commissioned utility-scale PV facilities decreased by 88 %, while that of onshore wind fell by 68 %. As a result, almost two-thirds of renewable power added in 2021 has had lower costs than the cheapest coal-fired options in G20 countries. Those data agree with the Lazard’s report (2023) which presents an average cost reduction of 77 % for utility-scale solar PV and 47 % for onshore wind, during the same period. The report also shows that both the unsubsidized average LCOE of utility-scale PV and onshore wind are lower than for combined-cycle and coal. The data supports the role of renewables as a cost-competitive tool to address the current climate and energy crisis by reducing GHG emissions. Peru began the promotion of generation projects from nonconventional renewable energy resources (NCRER) in 2008 with the approval of Legislative Decree No. 1002 (DL 1002). That was the first * Corresponding author. E-mail addresses: [email protected] (H. Fiestas-Chevez), [email protected] (J.M. Roldan-Fernandez), [email protected] (A.L. Trigo-Garcia), [email protected] (M. Burgos-Payan). Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro https://doi.org/10.1016/j.jclepro.2024.143389 Received 31 December 2023; Received in revised form 9 August 2024; Accepted 10 August 2024 Journal of Cleaner Production 471 (2024) 143389 Available online 11 August 2024 0959-6526/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). Law for the promotion of investment for the generation of electricity from renewable energies (Legislative Decree, 2008) and its corresponding regulation (Supreme Decree, 2011). DL 1002 considers NCRER to wind, solar, biomass, geothermal, tidal and small hydro (up to 20 MW) power plants, and introduced a variety of measures to promote the integration of NCRER in the Interconnected Electrical System (SEIN), including. •Priority in the sale of the NCRER generation considering that its marginal cost is zero, •Guarantee of access of the NCRER power plants to transmission and distribution networks, •Guarantee of a minimum income, above the marginal market price of the SEIN, allowing the NCRER’s promoters to recover the production costs of NCRER generators and, •Integration via biannual auctions. The NCRER generators awarded through auctions under DL 1002 sell their production at the marginal price of the SEIN, like any other type of generator. When the SEIN’s marginal price is below the auction price, NCRER generators receive compensation to cover the difference between the SEIN’s marginal price and the granted auction price. End users are charged in their bills with a supplement in the electricity transmission toll to cover this compensation (renewable premium). As can be seen, the integration of NCRER in the system mainly triggers two opposite economic effects for the end users. First, since the NCRER generation is integrated in the system at a null cost, the resulting SEIN’s marginal price and the corresponding total cost of the traded energy will be pushed down. Since NCRER generation is integrated into the system at zero cost, NCRER production will displace the same amount of the most expensive (and pollutant) thermal generation. These last dispatched generators would be the marginal units, however due to the integration of the NCRER, they are no longer necessary. As a result of the substitution of part of the expensive thermal generation by NCRER at zero cost, both the new resulting marginal price and the total cost of the traded energy in the SEIN will be lower. But now, the end users are charged in their bills with a supplement in the electricity transmission toll to cover the NCRER premium. Calls for NCRER auctions were held in 2009, 2011, 2013 and 2015, which have allowed the execution of 64 projects with a total installed capacity of 1274 MW. Unfortunately, since 2015 there have been no auctions for NCRER projects and the share of NCRER excluding small hydro (NCRER-h), in the annual electricity generation has stagnated at about 5 % since 2017. Table 1, based on data from OSINERGMIN NCRER auction website (OSINERGMIN, 2015), listed the projects awarded in the auctions, indicating their main characteristics, including the plant factor (the ratio between the actual energy produced by the plant and the energy that it would produce at rated power in the same period). It is important to highlight the reduction in the average prices offered for solar FV and wind technologies. It should also be noted that 34 proposals with a total of 2.17 GW of wind power and 48 proposals with a total of 3.09 GW of PV power plants were submitted to the fourth auction. These figures clearly show the potential for both commercial and industrial interest presented by these NCRER. In the regulations for the promotion of NCRER plant projects it was established that the MINEM must set and update every 5 years the target percentage of energy generation with NCRER-h plants. This percentage was initially set at 5 % by 2015 and although it has not officially changed, the MINEM minister, in his presentation to the Congress of the Republic in 2021, stated that the goal for 2030 would be 15 % of energy generated with NCRER-h (MEM, 2021). With the objective of fulfilling Peru’s commitments under the Paris Agreement (Draft Bill, 2020), there is also a draft Law from the Congress of the Republic which indicates that the objective percentage of NCRER-h in the electrical system must not be less than 20 % of total energy generated by 2030 nor less than 50 % by 2040. According to IRENA, the auction mechanisms have been successful due to the transparency of the process and reaching competitive prices. This study recommends maintaining the auction processes with the following improvements: establishing long-term planning, adjusting the design to ensure the participation of experienced developers, encouraging local participation, and updating the potential of NCRER (IRENA, 2014). Mitma (2015) analyzed the design of NCRER auctions in Peru and concludes that the objectives set in the auctions have been achieved. However, it also indicates that there is a need for a clear direction and course on the future of NCRER in Peru, since there is no sustainable energy model with a long-term vision. Mendiola et al. (2020) evaluated the competitive participation of wind projects in the fourth NCRER auction in Peru, using the theory of real options. This study concludes that the regulator can use the real options approach to improve the definition of guarantees of bid security and faithful performance, as well as defining useful guidelines for bidders in order to plan a participation strategy in the auctions. The environmental benefits of the incorporation of NCRER plants have been quantified by Vasquez et al. (2017). They estimated a mitigation of 6.4 million tons of CO 2-eq in the period from 2008 to 2016. The net benefit attributable to the NCRER generation policy, considering both the social value of the mitigated CO 2 and the operating cost of the plants, amounts to 158 million USD for this same period. Table 1 Main characteristics of the NCRER projects awarded in the auctions 2009–2015. Technology Auction Power (MW) Energy (GWh/year) Price (USD/MWh) Plant factor (%) Minimum Maximum Minimum Maximum Biomass First 2009 27.40 143.30 52.00 110.00 57.08 73.41 Second 2011 2.00 14.02 99.99 99.99 80.00 80.00 Third 2013 – – – – – – Fourth 2015 4.80 29.00 77.00 77.00 68.97 68.97 Wind First 2009 142.00 571.00 65.57 87.00 43.23 52.93 Second 2011 90.00 415.76 89.00 89.00 52.73 52.73 Third 2013 – – – – – – Fourth 2015 162.80 738.60 37.83 51.79 50.25 51.91 Hydro First 2009 179.86 1084.34 55.00 70.00 40.90 85.02 Second 2011 102.00 679.93 47.40 56.45 61.15 85.61 Third 2013 192.80 1171.51 50.50 64.80 48.07 85.46 Fourth 2015 79.66 448.16 40.00 58.20 53.25 76.96 Solar FV First 2009 80.00 172.94 215.00 225.00 21.37 28.92 Second 2011 16.00 2.04 119.90 119.00 30.68 30.68 Third 2013 – – – – – – Fourth 2015 184.50 33.45 47.98 48.50 32.79 30.94 H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 2 More recently, Campod´ onico and Carrera (2021) presented a broad analysis of the challenges posed by the energy transition towards renewables for Peru, studying and explaining the absence of NCRER project auctions since 2015. The work qualitatively identifies the following main criticisms and objections to the NCRER: The affectation of the NCRER premium on the purchasing power of users, since users must cover the premium with charges in electricity transmission rates; the lack of reliability and variability of the NCRER plants and the slowdown in economic growth in Peru, associated with the delay in the start of mining projects. Burgos-Payan et al. (2013),Rold´ an Fern´ andez Juan Manuel et al. (2016) and Roldan-Fernandez et al. (2021) analyzed the impact of renewables on the daily electricity market in Spain. The first study evaluates the influence of renewables on the daily electricity market and concludes that the price reduction offsets the subsidies these technologies receive. The second paper compares this effect with the impact that the introduction of efficiency or demand management measures would have on the day-ahead market. The third analyzed the decarbonization effect of PV self-consumption. Although these studies are developed in a daily energy market model that is different from the Peruvian market model, it is expected that the cost reduction effect will be similar in both cases. Given that the Peruvian wholesale market is based on the economic dispatch of generation units, the dispatching of NCRER generation with a marginal cost equal to zero reduces the marginal price of this market and accordingly the total cost of power generation. Consequently, it is necessary to examine the variations of the yearly average marginal price of the wholesale market and the cost of the energy traded in the market due to the integration of NCRER power plants. These variables play a direct and indirect role in shaping the pricing structure within power generation contracts for end users. As can be seen, none of the studies reviewed have addressed the evaluation of the economic impact that the integration of NCRER plants in the SEIN will cause, both on generators and on end users. To cover this gap, this work proposes a methodology to evaluate the reduction of average marginal price and of the annual total cost of the traded energy in the SEIN due to the integration of NCRER. Then, the reduction in the annual cost of the traded energy will be compared with the annual NCRER premium to determine the net saving for the system. As result of the trade-off between the reduction of the cost of the traded energy and the increase of cost due to the renewable premium, it will be shown how to determine the optimal amount of NCRER to integrate into the SEIN to maximize the annual net saving for the system. Closely related to the main goal of the work mentioned above, there are other secondary objectives that are also addressed, such as the evaluation of the volume of natural gas that no longer needs to be burned and the corresponding volume of GHG emissions avoided due to the integration of NCRER, as well as the evaluation of the avoided social cost due to the GHG abatement. After this introduction, a summary of the actual situation of the Peruvian electricity system will be shortly presented in Section 2, including a description of the wholesale market. Section 3introduces the proposed methodology while Section 4will be devoted to the presentation of results, including a short discussion on the methodology. Finally, the last section summarizes the main conclusions of the work. 2. 2. The Peruvian electricity system Fig. 1 shows the 2009–2021 evolution of the total power by technologies commissioned in the SEIN, based on data from the Economic Committee of the Interconnected Electric System (COES, 2021). According to the annual statistics of COES (2021), the installed power capacity of the Peruvian electrical system reached 13447 MW in 2021, while the maximum power demand of the system was 7173 MW, so the reserve margin of the system is 84 %. During 2021, 53.97 TWh were traded in the wholesale market with a total cost of 988.32 MUSD. The annual arithmetical average marginal price was 18.28 USD/MWh, while the weighted average marginal price was 18.50 USD/MWh. In this work the original Peruvian prices in Nuevos Soles Peruanos (PEN or S/) has been converted to USD considering a rate of 3.88 PEN/USD, which was the average exchange rate for 2021 (BCRP, 2022). As can be seen in Fig. 2, the participation of NCRER-h in the energy mix, initially set at 5 % by DL 1002, is already fulfilled today (in fact, since 2017), but it is necessary to triple the participation of NCRER-h to reach the 15 % target set by the Ministry’s proposal by 2030 and quadruple the production of NCRER-h to reach the 2050 target of 20 % set out in the Congress proposal. Fig. 3, based on 2021 data (COES, 2021), shows the monthly coverage of the national electricity demand by technology throughout 2021. Peru’s geographical location, slightly below the equator, means that demand does not suffer significant seasonal variations. The rainy period runs from November to April, so hydro production is greater in Fig. 1. Evolution of power installed annually by technologies from 2009 to 2021 in the SEIN. H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 3 those months. As can be seen in Fig. 3, demand is quite constant throughout the year, so renewable (NCRER +conventional hydro) and thermal generations are complementary. When renewable generation grows, part of the thermal production is not necessary. Thus, thermal generation is used to cover the residual demand of the system, that is, the demand not covered by renewable production. In Peru, Decree Law No. 25844 –Electric Concessions Law (LCE) and its Regulations (RLCE) (Decree Law, 1993) establish the conditions for the development of electrical activities. The system is based on a disintegrated model of activities with private participation, divided into generation, transmission, and distribution, including retail marketing within distribution. Generation develops in competition, while transmission and distribution are regulated, natural monopolies. From an institutional point of view, the Ministry of Energy and Mines (MEM) oversees managing electrical concessions, referential planning and approval of regulations. The Supervisory Agency for Private Investment in Energy and Mines (OSINERGMIN) has the role of regulator of the electricity sector. On the other hand, the technical and economic operation of the electrical system is the responsibility of the Economic Operation Committee of the National Interconnected System (COES), whose objective is to coordinate the operation at the minimum cost for the system, guaranteeing the security of the energy supply and the use of available energy resources. End users can be classified, basically, as regulated (tariff, up to 200 kW) and unregulated (agreement), according to their maximum power demand. 2.1. Wholesale market The competitive model in electric power generation in the SEIN is like the wholesale competition model described in Hunt (2002), but in the Peruvian case, generators can also act as energy buyers. The design of the wholesale market as a techno-economical dispatch is also related to that described in Kirschen and Strbac (2019) as an electricity pool, or to that described in Rothwell and G´ omez (2003) as a POOLCO (Power Pool Company) market. Peru’s SEIN operates under the supervision and coordination of COES, a single and centralized coordinator for the operations of both the electricity system (Transport System Operator - TSO) and the wholesale market (Market Operator - MO). The wholesale market in Peru is comparable to the model initially implemented in other countries in the Latin American region (Suding, 1996), such as Chile or Argentina. Bilateral contracts between generators companies (GENCOs) and distribution companies (DISCOs) or with unregulated users support energy withdrawal operations from the wholesale market in the Peruvian SEIN. Fig. 4 outlines the relationships among the different agents and roles in the market. As previously mentioned, the operation of the wholesale market in the SEIN is essentially based on a techno-economic dispatch carried out by COES (MO +TSO). For this, conventional thermal GENCOs must inform COES of their fuel purchases (type, quantity, and price), at the time the transactions were arranged. Since there are no demand agents in the market (asymmetric market), except for the few large users, the COES’s staff oversees forecast the demand on daily and weekly horizons, in 30-min time intervals. Fig. 2. SEIN 2021. Installed capacity (left) and energy generated (right). Fig. 3. Electricity demand coverage in the SEIN by type of technology in 2021. H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 4 After that, COES defines the generation dispatch for the day ahead by means of economic dispatch algorithms. In this way COES establishes the amount of energy to be delivered by each dispatched GENCO that will satisfy the demand and the corresponding marginal prices for all generators dispatched for each 30-min interval of the following day. OSINERGMIN sets the annual tariff for regulated users based on the forecast of the annual average marginal prices of the wholesale market and the information on the energy auctions carried out by DISCOs. Unregulated users negotiate energy prices and other contract terms with a GENCO or a DISCO as they deem appropriate. 2.2. Base scenario The situation of the system by 2021 has been taken as a reference (COES, 2021) to analyze the influence of the integration of NCRER plants in the SEIN. Table 2 shows the generation fleet and the energy generated during 2021, broken down by technology. As can be seen, NCRER-h plants represent 5.69 % of installed power and 5.14 % of annual electricity production. Fig. 5 shows the evolution of the marginal price in the market throughout 2021 (30 min time slots), according to data retrieved from COES (2022). 3. Methodology and data As indicated above, the main objective of this work is the evaluation of the two main and antagonistic economic effects of the integration of NCRER in the SEIN market. To reach that goal it is necessary first to develop a methodology for the evaluation of the economic effects of the integration of the NCRER in the SEIN market. On the one hand, it is needed to evaluate the reduction of the annual average marginal price and, therefore, of the total annual cost of the energy traded in the wholesale market, due to the integration of the production of NCRER. On the other hand, it is also necessary to consider the cost of the integration of the NCRER plants, the annual renewable premium that end users have to pay in their bills. The difference between the reduction of the total cost of energy traded on the market and the renewable premium will allow the net effect of integrating renewables to be assessed. After that, the amount of renewables that maximize the net reduction of cost of the energy traded in the market will be calculated. To replicate the optimal economic dispatch that the market operator should carry out after the integration of a certain amount of NCRER energy in a certain time slot of the market, it would be necessary to know the marginal costs of each of the generators that wants to sell production in the market. Once the market was replicated (economic dispatch) for all 17520 30-min time slots of the year, the annual average value of the marginal price and the yearly total cost of energy traded in the replicated market, as well as the corresponding renewable premium and the total saving, could be determined. Unfortunately, the necessary information on the marginal cost of different types of generation units is not publicly available, so the problem needs to be approached in another way. It is important to precise that the objective of the work is not to determine the marginal price or the total cost of the traded energy in each of the 17520 30-min time slots of the market for an entire year. The objective of the work is rather to determine the value of the annual average marginal price and the annual total cost of the energy traded in the market after integrating the considered amount of NCRER energy. Thus, the calculation of the marginal prices of each of the 15,720 time slots of the year is only the way to estimate the annual average value of the marginal price and the total cost of the annual energy traded in the market. In this work, a multiple linear regression will be used to describe the marginal market price of the as a function of the demand and the energy traded by technology. The market model will be parametrized based on the information of the operation (30-min time intervals) of the wholesale market (marginal prices and technologies dispatched) through 2021. 3.1. Wholesale market price The marginal price of the market is a function of the amount of energy produced by the different generation technologies programmed in Fig. 4. Diagram of the short-term wholesale market in Peru. Table 2 Situation of the electrical system in 2021. Type of power plants Installed power (MW) Power (%) Generated energy (GWh) Energy (%) Hydro 4907.9 36.50 28343.74 52.52 Thermal 7419.6 55.18 20537.29 38.05 Hydro NCRER 354.0 2.63 2319.35 4.30 Thermal NCRER 71.0 0.53 177.55 0.33 Solar PV 282.3 2.10 796.99 1.48 Wind 412.2 3.07 1797.04 3.33 Total conventional 12327.5 91.67 48881.03 90.57 Total NCRER 1119.5 8.33 5090.93 9.42 Total SEIN 13447 100.00 53971.96 100.00 Total NCRER-h 765.5 5.69 2771.58 5.14 Maximum power demand 7173.00 MW Reserve margin 87.47 % H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 5 the dispatch of energy. Fig. 6 is elaborated based on COES (2021) operation data, showing the relation between the marginal price and the amount of energy of every technology dispatched in the market in 2021. As can be seen, the marginal price is strongly dependent on both the thermal generation (natural gas), with a positive correlation factor of CF G = + 0.7826 ((USD/MWh)/MWh), and conventional hydro, with a negative correlation factor of CF H =–0.6255 ((USD/MWh)/MWh). Although the correlation factor of NCRER generation is negative and rather low, CF NCRER =–0.2282 ((USD/MWh)/MWh), it has a significant influence on the marginal price because it is inversely associated with the thermal generation. This negative cross-coupling is due to the fact that hydro power, due to its low cost, is being dispatched. This means that the residual energy demand must be covered by the rest of the technologies (gas +NCRER). Hence, if the NCRER generation decreases the thermal generation must increase to satisfy the demand. This observation leads to proposing that the marginal price, MP i , for a time interval i, can be explained as a function of the amount of the total energy demand of the system, E Di , the amount of dispatched conventional hydro, E Hi , and the residual dispatched energy, E ri , that is, the difference between the total demand and the conventional hydro energy, E ri =E Di -E Hi . Since that difference must be covered by the sum of the energy generated from thermal plants (mainly gas), E Gi , and the total energy NCRER, E Ri (E ri =E Di -E Hi =E Gi +E Ri ), the amount of thermal energy, E Gi , can be eliminated by expressing it in terms of the residual energy, E ri =E Di -E Hi , and the NCRER energy, E Ri , as: EGi =Eri −ERi =EDi −EHi −ERi (1) As a result, the marginal price can be expressed as: Fig. 5. Marginal prices (USD/MWh) of the SEIN during 2021. Fig. 6. Marginal price correlations with energy generated by technology and demand. H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 6 MPi=f(EDi,EHi,ERi) ≈ k0+kʹ H⋅EHi +kR⋅(EDi −EHi −ERi) =k0−(kR−kʹ H)⋅EHi +kR⋅(EDi −ERi) =k0−kH⋅EHi +kR⋅(EDi −ERi)(2) where k 0 ,k’ H ,k H and k R are the coefficients of the multiple linear regression. Using real market energy data (MWh) and price (USD/MWh) corresponding to the 17520 30-min time slots of 2021, the marginal market price can be expressed as: MPi=f(EDi,EHi,ERi) ≈ k0−kH⋅EHi +kR⋅(EDi −ERi) = − 8.1691 −0.0108 ⋅EHi +1.0977 ⋅10−2⋅(EDi −ERi)(3) (R2=0.6127 ) Now, the annual total cost of energy traded in the wholesale market, C T , can be expressed as: CT=∑ 17520 i=1 CTi =∑ 17520 i=1 EDi ⋅MPi≈∑ 17520 i=1 EDi ⋅(k0−kH⋅EHi +kR⋅(EDi −ERi)) =∑ 17520 i=1 EDi ⋅(−8.1691 −0.0108⋅EHi +1.0977⋅10−2⋅(EDi −ERi))(4) the average (arithmetic) marginal price, MP av , as: MPav =1 17520⋅∑ 17520 i=1 MPi ≈1 17520⋅∑ 17520 i=1 EDi ⋅(k0−kH⋅EHi +kR⋅(EDi −ERi)) (5) and the annual weighted average marginal price, MP wav , as: MPwav =CT ED =∑ 17520 i=1 EDi⋅MPi ∑ 17520 i=1 EDi ≈∑ 17520 i=1 EDi⋅(k0−kH⋅EHi +kR⋅(EDi −ENRi)) ∑ 17520 i=1 EDi (6) Table 3 shows a comparison of the values of the total energy traded in the market, the total cost of the energy traded, as well as the annual average and weighted averaged marginal prices in 2021, according to COES (2021), and those obtained by the proposed market model (3)–(6). As can be seen, the results of the estimates obtained with the proposed model are in very well agreement with the real market data. On the other hand, the annual energy generated from thermal plants, E G , can be expressed from the annual amounts of demand, E D , conventional hydro generation, E H , and the NCRER energy, E R , as: EG(ED,EH,ER) = ∑ 17520 i=1 EGi =∑ 17520 i=1 EDi −∑ 17520 i=1 EHi −∑ 17520 i=1 ERi =ED−EH−ER (7) Finally, the reduction in thermal production, ΔE G , can be determined by the difference between the 2021 thermal production, E G2021 , and the thermal production of the considered scenario, E G . Then, using (7) the reduction in thermal production is expressed in terms of the annual amounts of demand, E D , conventional hydro generation, E H , and NCRER energy, E R , as: ΔEG(ED,EH,ER) = EG2021 −EG(ED,EH,ER) = EG2021 −ED+EH+ER(8) 4. Results and discussion As mentioned, the real situation of the system in 2021 has been taken as a reference for what refers to the NCRER energy production profiles, the operating patterns of the different NCRER generation technologies, as well as for the proportion between each of the NCRER technologies. First, the impact of the integration of a varying percentage of NCRER in the market will be addressed. Scenarios both with greater and lesser participation of NCRER than the reference situation of the SEIN are considered. In this way, both the downward pressure on the average marginal price and total cost of the energy traded in the market due to the integration of additional NCRER can be evaluated. Table 4 summarizes the main characteristics of the 21 NCRER scenarios considered. Taking the NCRER capacity of the actual situation of the SEIN in 2021 as the reference (100 % of the installed NCRER power capacity), the scenarios have been chosen so that the installed NCRER varies in intervals of 25 % of the capacity with respect to the base scenario. As can be seen, Scenario 4 (100 % of NCRER) corresponds to the reference, the real situation of 2021. Scenarios 0 to 3 correspond to hypothetical) reductions in the NCRER production compared with the baseline scenario. The remaining sixteen scenarios (5–20) correspond to potential future NCRER percentages higher than the share corresponding to the baseline scenario. Since large hydro power plants already produce 52.52 % (2021) of the energy demand of the SEIN, going further Scenario 20 would mean that new NCRER generation would replace part of the conventional hydro which would not make much sense. It has been considered that each scenario reproduces the situation of the base scenario (2021), expanded or reduced in scale to match the Table 3 Comparison of actual 2021 annual market data with those obtained by the market model. Magnitude COES Proposed Model Error MP av , average marginal price 18.2793 (USD/ MWh) 18.2848 (USD/ MWh) 0.03 (%) MP wav , weighted average marginal price 18.4971 (USD/ MWh) 18.4941 USD/ MWh) 0.02 (%) E D , total energy traded in the market 53971.96 (GWh) 53971.96 (GWh) 0.00 (%) C T , total cost of the energy traded 998.32 (MUSD) 998.16 (MUSD) 0.02 (%) Table 4 Main characteristics of the scenarios considered. Scenario NCRER power share a NCRER Installed power capacity NCRER generated Energy Participation in energy generated by NCRER in the SEIN a Participation in energy generated by NCRER-H in the SEIN a P R% P R E R E R% E R-H% (%) (MW) (MWh) (%) (%) 0 0 0.00 0.00 0.00 0.00 1 25 279.88 12.73 2.36 1.28 2 50 559.75 25.45 4.71 2.56 3 75 839.63 38.18 7.07 3.84 4 100 1119.50 50.91 9.42 5.13 5 125 1399.38 63.64 11.78 6.41 6 150 1679.25 76.36 14.13 7.69 7 175 1959.13 89.09 16.49 8.97 8 200 2239.00 101.82 18.84 10.25 9 225 2518.88 114.55 21.20 11.53 10 250 2798.75 127.27 23.56 12.81 11 275 3078.63 140.00 25.91 14.09 12 300 3358.50 152.73 28.27 15.38 13 325 3638.38 165.45 30.62 16.66 14 350 3918.25 178.18 32.98 17.94 15 375 4198.13 190.91 35.33 19.22 16 400 4478.00 203.64 37.69 20.50 17 425 4757.88 216.36 40.04 21.78 18 450 5037.75 229.09 42.40 23.06 19 475 5317.63 241.82 44.76 24.34 20 500 5597.50 254.55 47.11 25.63 a Based on the year 2021. H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 7 specified NCRER share. This includes maintaining the power of conventional hydro and thermal plants without variations. For instance, Scenario 8 considers that the installed NCRER power is twice that of the base scenario. This means that the installed power of each of the NCRER technologies (PV, wind, small hydro, biomass thermal) is double (200 %) that actually installed in 2021 (100 %), and that the annual production of each NCRER technologies will also be double that of the base scenario. Using (3), the marginal prices are evaluated for each of the 17520 30 min time slots of the year corresponding to each of the scenarios considered. Next, expressions (4)–(6) are used to evaluate the total annual cost of energy traded in the wholesale market, C T , the annual average (arithmetic) marginal price, MP av , and the annual weighted average marginal price, MP wav , corresponding to the share of NCRER, E R % , of each considered scenario. Table 5 shows the main results. Here, the relative reduction in the total cost of the energy traded corresponding to the integration of a certain amount of NCRER production, ΔC T2021 (E R% ), is calculated as the difference between the total cost of the energy traded in the base scenario, C T2021 , and the total cost of the energy traded corresponding to the integration of a certain amount of NCRER production, C T (E R% ): ΔCT2021(ER%) = CT2021 −CT(ER%)(9) As can be seen, as the participation of NCRER, E R% , integrated into the market grows, both the marginal price, MP av , the cost of the energy traded in the market, C T , decreases. On the contrary, the reduction of the cost of the energy, ΔC T2021 , and the number of hours with zero price, t (MP i =0), increases with the growing participation of NCRER. Starting with the results of Scenario 4, the base scenario, it is worth looking at the results of scenarios 0 to 3. These scenarios consider a full or partial reduction of the current NCRER production (2021). As can be seen, as renewable generation is reduced from Scenario 4 to 0 (100 % reduction). •The average price increases from 18.49 USD/MEM (2021) to 24.94 USD/MWh (Scenario 0). •The total cost of energy negotiated in the market increases from 998.16 MUSD to 1345.87 MUSD •The number of hours with zero marginal price is reduced from 23.5 h to 0. Therefore, a 9.42 % reduction of the 2021 NCRER production integrated into the market would cause a price increase of 34.88 % (1.35 times higher) in both the marginal price and the total cost of the energy traded in the market. The opposing effects become evident in scenarios with increasing shares of NCRER production. When comparing the results of the base scenario with those of Scenario 8, it becomes clear that doubling the NCRER generation integrated into the market has the following effects. •The average price decreases from 18.49 USD/MEM (2021) to 12.41 USD/MWh (Scenario 8). •The total cost of energy traded in the market decreases from 998.16 MUSD to 669.53 MUSD •The number of hours with zero marginal price grows from 23.5 h to 1221.5 h. Doubling (100 % increase) of NCRER production integrated into the market would result in a price reduction of 32.88 %, making both the marginal price and the total cost of the energy traded in the market 0.67 times lower. Using the results of Table 5,Fig. 7 shows the variation of the annual weighted average marginal price of the SEIN market, MP wav , as a function of the percentage of NCRER generation integrated into the system, E R% . The fitting by regression of the points (E R% ,MP wav ) leads to a second-degree polynomial such as (10): MPwav(ER%) ≃ 6.868 ⋅10−3⋅E2 R%−0.8104⋅ER%+25.28 (10) (R2=0.9997) Similarly, the adjust of the data of cost of the energy traded in the market, C T (MUSD), relative reduction of the total cost of the energy traded, ΔC T2021 (MUSD), and number of hours with zero price, t MP_0 (h), with the share of NCRER integrated into the market, E R% (%), lead to expressions (11)–(13): CT(ER%) ≃ 0.3707 ⋅E2 NR%−43.737⋅ER%+1.3647⋅103(11) (R2=0.9997) ΔCT2021(ER%) ≃ 0.3706 ⋅E2 R%−43.7308⋅ER%+3.6654⋅102(12) (R2=0.9997) tMP0(ER%) ≈ 1.9320 ⋅10−4⋅E5 R%−2.0555⋅10−2⋅E4 R%+0.6324⋅E3 R% −1.0805⋅E2 R%−29.6763⋅ER%+46.0972 (13) (R2=0.9992) But the integration of the NCRER into the market is not free for end users since they have to pay an additional cost in their bills: the renewable premium (transmission toll). Consequently, in order to determine the net effect of the integration of the NCRER in the system, it is necessary to evaluate the renewable premium. According to COES, in 2021 (base scenario), where the share of NCRER was, E R2021% =9.42 %, the total compensation paid to NCRER generators amounted RP 2021 = 203.46 MUSD (OSINERGMIN, 2022). In this work the renewable premium, RP, corresponding to any other scenario has been calculated proportionally to the corresponding NCRER quota, E R% : Table 5 Total cost of the system and annual average marginal price according to percentage of participation in energy generated by NCRER in SEIN (annual energy traded: 53.97 TWh). Scenario Share of NCRER energy generated in the SEIN Annual weighted average marginal Total cost of the traded energy Relative reduction of the total cost Time with marginal price equal to 0 E R% MP wav C T ΔC T2021 t(MP i =0) (%) (USD/ MWh) (MUSD) (MUSD) (h) 0 0.00 24.94 1345.87 −347.71 0.0 1 2.36 23.32 1258.89 −260.73 0.5 2 4.71 21.71 1171.9 −173.74 4.0 3 7.07 20.10 1084.99 −86.83 19.5 4 9.42 18.49 998.16 0.00 23.5 5 11.78 16.90 911.91 86.25 159.5 6 14.13 15.33 827.5 170.66 396.5 7 16.49 13.82 746.13 252.03 770.0 8 18.84 12.41 669.53 328.63 1221.5 9 21.20 11.09 598.52 399.64 1688.5 10 23.55 9.89 533.75 464.41 2173.0 11 25.91 8.79 474.15 524.01 2593 12 28.26 7.77 419.45 578.71 3001.0 13 30.62 6.84 369.34 628.82 3398.5 14 32.97 5.99 323.46 674.70 3740.0 15 35.33 5.21 281.29 716.87 4054.5 16 37.68 4.50 242.89 755.27 4413.0 17 40.04 3.87 208.85 789.31 4800.0 18 42.39 3.31 178.80 819.36 5172.0 19 44.75 2.83 152.57 845.59 5525.5 20 47.10 2.41 129.89 868.27 5884.5 H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 8 RP(ER%) = ER% ER2021% ⋅RP2021(ER2021%)(14) As can be seen, the renewable premium is null for the scenario without any renewable energy, RP(E R% =0) =0. In order to have a common reference for the reduction of the cost of the energy traded in the market, the absolute reduction in the total cost of the energy traded corresponding to the integration of a certain amount of NCRER production, ΔC T (E R% ), is calculated now as the difference between the total cost of the energy traded in the scenario without any NCRER, C T0 = C T (E R% =0), and the total cost of the energy traded corresponding to the integration of a certain amount of NCRER production, C T (E R% ): ΔCT(ER%) = CT(ER%=0) − CT(ER%) = CT0−CT(ER%)(15) Now the net annual saving in cost of the energy traded in the market, S(Table 6), can be calculated as the difference between reduction of the total cost of energy traded in the market, ΔC T , and the corresponding renewable premium, RP: S(ER%) = ΔCT(ER%) − RP(ER%)(16) The points (E R% ,S) shown in Fig. 8 illustrate the variation of the net saving with the share of NCRER integrated into the market. The fitting of the points (E R% ,S) leads to: S(ER%) ≃ − 0.3706⋅E2 R%+22.1376⋅ER%−18.8614 R2=0.9945 (17) As can be seen, there is an optimum share of renewables that maximizes the net saving of the system. That optimum point corresponds to the Scenario 12, where a share of renewables, E ROP% =28.26 %, would lead to a net saving of S=316.04 MUSD. It is worth noting that the maximum of the net saving curve is fairly flat. Table 7 shows the composition of the generating fleet and its participation in power generation corresponding to the optimum case with a 28.26 % share of NCRER energy integrated into the market. 4.1. 4.1. The volume of gas replaced and avoided GHG emissions The integration of renewable production in the SEIN displaces thermal generation, which leads to a reduction in the consumption of fossil fuels and, consequently, GHG emissions. But not only CO 2 emissions are avoided, since other gases that pollute the environment and are even harmful to health are also avoided. Given that the SEIN thermal generation fleet uses mainly natural gas and some residual diesel, the integration of renewables would first replace the diesel power plants (more expensive and polluting production) and then the natural gas plants. A simplified worst-case approach has been adopted to estimate the amount of GHG emissions, considering all the thermal fleet as if they were gas-fired power plants. This leads to a conservative estimation of the reduction of the GHG emissions. According to the daily operation evaluation reports for the month of December 2021 (COES, 2022), an average consumption rate of v G =200.32 m 3 /MWh of natural gas for a combined cycle power plant has been considered. As a result, the avoided volume of natural gas (respect to 2021) replaced by NCRER Fig. 7. Variation Average Marginal Price of the system (USD/MWh) based on the percentage of NCRER integrated in the SEIN. Table 6 Annual savings in the total cost of the energy traded in the wholesale market. Scenario NCRER’s share of energy generated in SEIN Reduction of the Total Cost Renewable Premium Net saving E R% ΔC T RP S (%) (MUSD) (MUSD) (MUSD) 0 0.00 0.00 0.00 0.00 1 2.36 86.98 50.97 36.01 2 4.71 173.97 101.73 72.24 3 7.07 260.88 152.70 108.18 4 9.42 347.71 203.46 144.25 5 11.78 433.96 254.43 179.53 6 14.13 518.37 305.19 213.18 7 16.49 599.74 356.16 243.58 8 18.84 676.34 406.92 269.42 9 21.20 747.35 457.89 289.46 10 23.55 812.12 508.65 303.47 11 25.91 871.72 559.62 312.10 12 28.26 926.42 610.38 316.04 13 30.62 976.53 661.35 315.18 14 32.97 1022.41 712.11 310.30 15 35.33 1064.58 763.08 301.50 16 37.68 1102.98 813.84 289.14 17 40.04 1137.02 864.81 272.21 18 42.39 1167.07 915.57 251.50 19 44.75 1193.30 966.54 226.76 20 47.10 1215.98 1017.30 198.68 H. Fiestas-Chevez et al. Journal of Cleaner Production 471 (2024) 143389 9