Nowadays, power systems are under tremendous pressure to be decarbonized to reach different targets imposed by clean energy policies. For example, in the European Green Deal (European Commission, 2019) and Fit for 55 legislative packages, the EU and its member states are committed to cutting net greenhouse gas emissions in the EU by at least 55% by 2030, compared to 1990 levels. In 2030, 40% of the total energy consumption must be generated by renewable energy sources (VRES). One key element for achieving these goals is to increase the share of renewable generation (mainly solar photovoltaics (PV) and wind) for producing electricity. These types of renewables are not controllable or inflexible. Their integration requires other flexible technologies such as combined cycle gas turbines (CCGT), storage hydro, pumped-storage hydro, battery, solar thermal (or concentrated solar power CSP), demand side response (DSR), electric vehicle (EV) with or without a vehicle to grid (V2G) possibility, and power to hydrogen (P2H).

The phasing out of fossil fuel power plants, coupled with the increasing share of renewable energy, stands as a challenge for the System Operator according to the security of supply during peak hours (Denholm et al., 2020) ¹ . Indeed, firm capacity was traditionally provided by thermal and hydroelectric technologies. Given that the energy mix will introduce recent technologies such as batteries, it is necessary to know how the different technologies will respond to the several needs of the power system ² .

Although firmness assessment methods already exist (Madaeni, Sioshansi, and Denholm, 2013), they should be adapted since they are based on the availability of thermal and hydropower technologies. Authors in (Agency for the Cooperation of Energy Regulators, 2020; Ministerio para la Transición Ecológica y el Reto Demográfico, 2020) argue that firm capacity should be assessed jointly with the operational flexibility of power systems.

This paper analyzes how the different generating (e.g., CCGT) and storage technologies (pumped-storage hydro, battery, or solar thermal) play a role in integrating renewables by providing firmness and operational flexibility in 2030 to the Spanish system.

The main contributions of the paper are:

• propose an operational flexibility assessment method in two phases: a) to analyze ex-ante the margin of the different flexibility dimensions and, b) to determine how flexibility dimensions are covered with the various technologies;

The paper is organized as follows. Section 2 presents the metrics for flexibility assessment. Section 3 describes the Spanish case study characteristics for 2030 and the required flexibility needs. Section 4 presents the modeling results and how the flexibility needs are expected to be provided in the Spanish system in 2030. Section 5 concludes.

To analyze the contribution of the storage technologies to the system operation, we must first introduce the definition of operational flexibility as the ability of the system to withstand the uncertainty and variability in generation and electricity demand while maintaining the desired reliability at an affordable cost (Impram, Varbak Nese, and Oral, 2020). We analyze the long-medium-term system operation with a time scope from several weeks to several years representing only the operation of the generation and storage assets (Hadi et al., 2022).

Once assumed this definition is, the next question is to define a framework to measure it: which dimensions of operational flexibility can be defined or how to measure them? Since power systems will be mainly composed of inflexible technologies, the requirements of the power system, its capabilities, and the contributions of the flexible technologies to the operational flexibility of the power system are assessed respectively to the net demand ³ (Heggarty et al., 2020).

(Huclin et al., 2023) propose a conceptual framework for jointly analyzing power systems’ firmness and operational flexibility. They split the analysis among system requirements (Which flexibilities the system needs), capabilities (How much operational flexibility does the system have?), and contributions (Who and in what dimensions is the flexibility provided?). Applying the discrete Fourier transform to the net demand, the authors (Huclin et al., 2023) found that half a day, a day, and a week are the relevant time scopes for analyzing the operational flexibility dimensions in several European countries.The results obtained by (Huclin et al., 2022) are in line with similar studies focused on power system operational flexibility (Heggarty et al., 2020; Saarinen and Tokimatsu, 2021). A similar flexibility assessment framework is presented in (Lannoye, Flynn, and O’Malley, 2012). Some metrics to measure the system flexibility were proposed by (Cochran et al., 2014; Ulbig and Andersson, 2015).

It has been shown that an increased share of RES increases the range of variation of the net demand and, consequently, the flexibility requirement (Heggarty et al., 2019; Saarinen and Tokimatsu, 2021). Flexibility assessment in power systems is essentially based on these three main dimensions that measure the ability of the system to follow and balance generation with demand, see (Ulbig and Andersson, 2015; Heggarty et al., 2019):

a) Power gradient. Ramps [MW/h]

It corresponds to the power variation per unit of time. In systems with enough quick-response generation (e.g., hydropower), the time interval to analyze this metric can be 1 hour. Only systems with no generation of this type may need to deal with shorter time intervals of minutes or seconds. The primary metrics are upward and downward hourly (bi-hourly, tri-hourly) ramps of the net power demand. Analysis of other time intervals (e.g., 6 h) for the ramps for several European countries is found in (Huber, Dimkova, and Hamacher, 2014).

b) Power [MW]

In the short-term, this metric deals with the demand-supply balance at any point, with the procurement of operating reserves for balancing the short-term uncertainty due to forecast errors in generation or demand. In the medium term (weeks to months), the availability of enough generation to supply the demand is defined by the unit firmness ⁴ , which is the contribution of each unit during the critical periods ⁵ , in this paper, approximated as the peak (net) demand hours. In the long term, system adequacy is usually measured by deterministic indexes like the reserve margin (RM) or probabilistic ones like the expected energy not served (EENS), and loss of load probability (LOLP), (Billinton and Li, 1994).

c) Energy [MWh]

Integrating the demand along different time intervals (e.g., day, week, season) defines the system requirements. The net energy variation for those intervals is linked to the system storage needs and flexible generation. It can be easily observed that the demand, hydro, wind, and solar generation have strong seasonal behavior (Red Eléctrica de España, 2021). Therefore, the net energy demand will be affected, as also the flexibility requirements.

From a more practical point of view, ENTSO-e (European Network of Transmission System and Operators for Electricity, 2022a) suggests two flexibility metrics (ramps and scarcity periods) as the starting point to analyze at a European scale. They mention that ramps can be especially critical at sunset in regions with large PV generation and simultaneous demand increases. Besides, they also propose the analysis of 5-day scarcity periods (e.g., dunkelflaute, an anticyclonic gloom where almost no wind and solar energy is generated) to analyze extended periods with low weather-dependent generation.

Other potential flexibility metrics associated with real-time system operation (e.g., inertia, rate of change of frequency, area control error, etc.) or related to transfer capacities and congestion management among areas, voltage control, and power quality and other system services are out of the scope of this paper.

In this paper, we propose these two phases for assessing the operational flexibility in an electric system:

a) Margin analysis

This phase answers the question: Does the system have enough operational flexibility? For that purpose and any dimension, a margin based on the system availability of the product (i.e., net ramp, net load, net energy) and the system requirements is computed: d i m e n s i o n   m a r g i n = a v a i l a b i l i t y r e q u i r e m e n t

For example, the upward ramp margin will be the ratio between the sum of the available upward ramps of the flexible technologies and the maximum upward ramp of the net demand.

b) Flexibility in system operation

How much is the contribution of each technology to each flexibility product? The system operation is simulated by an operational model that determines the optimal operation of the system, i.e., the use of the generation and storage resources to satisfy the demand considering all the operating constraints of the generation and storage units. The model considers the limitations of thermal units (ramp up/down, minimum up/down time, minimum load, must run, etc.), hydro scheduling of hydropower plants and reservoirs, battery management constraints (e.g., state of charge and charging, and discharging processes), and operating reserve requirements. It is genuinely relevant to consider all these constraints, given that the model must represent the system operation as realistically as possible.

The Spanish National and Climate Plan NECP (Ministerio para la Transición Ecológica y el Reto Demográfico, 2020) proposes the pathway to reach the emission reduction and increase in renewable production required for achieving the European energy policies. A similar consistent exercise is done in the National Trends scenario of the Ten-Year Network Development Plan 2022 by ENTSO-e (European Network of Transmission System and Operators for Electricity, 2022a). These studies analyze horizons 2030, 2040, and 2050. For the paper’s case study, we have selected the first horizon, 2030, as the more realistic. The case study’s wind, solar, and demand data have been taken from the TYNDP 2022 (European Network of Transmission System and Operators for Electricity, 2021). We have updated the CO2 price to 140 €/tCO2 according to the last estimations made by ENTSO-e in July 2022 (European Network of Transmission System and Operators for Electricity, 2022b).

In Table 1, we present a summary of the installed capacity and production for the different technologies according to the TYNDP 2022 and the objective scenario of the NECP. The operation in the TYNDP is represented for three scenarios (called climate years). The last row of the table shows the peak demand (left) and the installed capacity (right).

We have considered for the Spanish power system 3 nuclear power plants, 50 CCGTs, 50 storage hydro programming units, three open-loop pumped-hydro storage (OL-PHS), and ten closed-loop pumped-hydro storage (CL-PHS). Solar PV and solar thermal, and wind are aggregated for all units.

Now, we assess the deployment of the system flexibility by simulating the system operation from an economic point of view. We use the openTEPES model (Ramos, Quispe, and Lumbreras, 2022). This optimization-based model determines the hourly dispatch of the different generating units with all the detailed operating constraints to minimize the total system variable cost.

Table 3 and Figure 3 summarize the output of each technology in 2030. Wind generation has the highest energy share, followed by solar PV, CCGT, and nuclear. Then, several storage technologies, such as storage hydro, pumped-hydro storage, and solar thermal, also have essential production. According to these numbers, hydro, wind, and solar renewable generation satisfies 75% of the demand.

Figure 4 shows the capacity factor of each technology, which is the energy produced divided by the installed capacity time and the year hours. The capacity factor of the PHS is higher than that of the battery, which means that the PHS’s significant storage capability overcomes the battery’s higher efficiency ¹⁴ . Similar observations were made by (Huclin et al., 2022).

In the paper, we have proposed three dimensions relevant to the flexibility assessment: power gradient, power, and energy available at different timescales. A two-phase procedure will analyze an electric system’s flexibility to cope with renewables’ integration. The first step, ex-ante, determines the margin on any dimension, and the second runs a cost-based operation model to determine how these dimensions are covered.

Upward and downward ramps of the net demand will increase dramatically in the Spanish system in 2030 due to high wind and solar share. These high ramps introduce new challenges to the operation of flexible technologies (CCGT, storage hydro, open-loop and closed-loop pumped-hydro storage, and batteries). A ramp margin of 120% in 2030 is enough to consider that the system will cope with the high ramp due to the decrease in solar PV generation at sunset or a reduction in the generation to cope with the sunrise.

Net demand ramps are evenly provided by different flexible technologies, with batteries and new CL-PHS being the main contributors.

Although the annual capacity factors of the hydro storage technologies barely exceed 20%, they enormously increase to +50% in the critical net demand hours, showing their high contribution to the system firmness. On the contrary, batteries can only play a minor role in system firmness due to their limited storage capacity.

Storage hydro and open-loop pumped-hydro storage provide weekly variation of the net demand, while other technologies contribute slightly.

The markets and the remuneration of flexibility services should incentivize the investments and the proper operations of different technologies to provide the flexibility needs of power systems highly dominated by non-dispatchable resources.