Renewables in the spotlight

Neither guilty nor victims: renewables stabilise

When the electricity system experiences periods of increased demand, i.e. periods when maintaining the grid within its safety limits becomes more difficult than normal, public debate often seeks to identify an immediate culprit, and renewables often end up in the spotlight. This is exacerbated if there are also problems or complications that are noticeable to the user, such as voltage fluctuations, occasional disconnections or, in extreme (but already known) cases, large-scale power outages.

In electrical systems, stability is not a matter of opinion, but rather a physical property with a specific definition [1], which must be ensured through a set of measurable performance requirements that are demanded of all resources connected to the system and affect its behaviour. Therefore, rather than assessing whether a technology is ‘more or less stable’ in itself, the relevant debate is to analyse what behaviours are required of that technology, whether it is capable of providing them and how they are coordinated in the operation of the system. With regard to this framework of requirements, a key aspect to be questioned is its consistency with the characteristics of the electricity grid at the time of its application. In other words, it makes no sense to require the same thing in a centralized system dominated by synchronous generation and where renewables represent a minority (as in past decades) as in a more distributed system dominated by renewable energies (the current and future trend, Fig. 1). In both scenarios, the objective is the same: for the grid to operate stably and safely, within defined margins. However, the way in which this objective is achieved depends on the characteristics of the energy source in question and the electrical elements that serve as the interface between the source and the grid connection point.

To develop these aspects in greater detail and assess the capacity of renewable energies to contribute to the objective of stability and security, three levels or approaches can be established to differentiate between conventional (predominantly non-renewable) generation and renewable generation: (i) how the grid is organized and where generation is connected, (ii) how the available power and its regulation capacity behave, and (iii) what type of electrical interface connects the source to the grid and how it is controlled.

In terms of network structure, conventional generation is centralized in large generating nodes, in contrast to a more distributed structure with greater proximity between generation and consumption in the case of renewables. Although this is not an intrinsic characteristic of renewable generation, it is a trend with the increase in its presence. The previous article [1] described how it is necessary to evolve the analysis methods and design criteria for controlling the equipment that regulates renewable generation in this distributed generation framework. Although a distributed electricity system poses new management challenges, it offers a number of benefits highlighted by European network operators in [2] and [3], such as cost reduction, management flexibility, greater adaptability and efficiency, and better resource allocation.

THE VARIABILITY OF THE ENERGY SOURCE

Regarding the energy source and the variability/controllability of the power supplied and its adaptation to demand, it is important to distinguish between two aspects. On the one hand, the availability of the energy resource, which determines a maximum power value Pmax that the energy generator cannot exceed. On the other hand, there is the capacity to regulate power, i.e. to follow a setpoint Pset point that can be increased or decreased as deemed appropriate by higher-level regulators or the network operator. Thus, a generation element, whether renewable or non-renewable, that is at a given moment delivering a power value equal to its maximum power, P=Pmax, can never produce a positive increase in that power. In other words, operating points where the commanded power setpoint is greater than the maximum power are not viable. However, if there is regulation capacity, that same element will be able to produce decreases in its generated power, reducing it below Pmax.

After clarifying these basic aspects, what are the differences between non-renewable or conventional generation and renewable generation? In terms of power regulation, within the range 0<Pset point<Pmax, both have ample capacity, but with nuances. In conventional generation, power ramps are more limited (instructions can be followed with a maximum slope subject to thermal and mechanical restrictions), there may be a technical minimum in the value of power generated, and there may be costs derived from fatigue and maintenance and associated with rapid changes in the power generated [4,5]. Renewable generation (meaning that which is connected to the grid via power electronics-based converters, such as solar or wind power) offers much greater flexibility in power regulation capacity, with faster ramps and much lower or non-existent operating minimums. To clarify, and by way of example, an inverter connected to a photovoltaic panel at a time of high irradiance may be generating Pmax and then switch to generating any value below Pmax (including 0) in a matter of milliseconds if required by the higher control or the grid operator, with this dynamic being determined by the performance of the control algorithm and without negatively affecting system elements, as can occur with conventional generators.

The most obvious weakness of renewables is the value of Pmax and its variability. While the availability of coal, water, gas, etc. is known, irradiance or wind levels can change in a matter of seconds. Therefore, in conventional power plants, Pmax is high and is not normally limited by the resource but by the physical capabilities of the plant, while in photovoltaic or wind power plants, in addition to the physical limits of the inverters and connection lines (which are known and assumed to be well dimensioned), P_max may be reduced when the resource is scarce. This must be added to normal operating conditions. For efficiency reasons, it makes sense to operate renewable energy plants at their maximum available power, since the resource is there and is ‘free’. In conventional power plants, on the other hand, they operate at the minimum power required, with a certain margin with respect to Pmax, since the resource is non-renewable and has costs. Therefore, in a system with an energy mix that includes renewable energy (wind and photovoltaic) and synchronous generation, as is currently the case, the most efficient approach is for the former to operate at 100% (or very close to) its maximum power and for the latter to supply the remaining power to meet demand. This is applicable in general terms, although other scenarios may arise when there are network restrictions, services, etc. In this scenario, any upward regulation in power (increase in generation) will necessarily have to be provided by non-renewable generation, just as it will have to make up for any decrease in renewable resources.

ENERGY STORAGE: FROM VARIABILITY TO OPERATING CAPACITY

The question here is: what happens if we move towards a model with an increasingly lower presence of synchronous generation? There is the option of renewables operating in power reserve mode, i.e. below their maximum power, with a margin that allows for some upward regulation (Fig. 2). In addition to having an energy/economic opportunity cost, this is a partial solution, designed to provide capacity to meet certain operating requirements, such as those related to frequency regulation. The solution in this case is the hybridization of renewables + storage. Storage acts as an energy buffer, absorbing power (charging) when generation exceeds demand and delivering power (discharging) in the opposite case, allowing for upward and downward power regulation, compensating for the variability of renewable resources and replacing the functionality previously provided by conventional/synchronous generation.

Regarding storage technologies, there are many different types, and many of them are highly relevant today: battery-based systems (BESS), with technologies such as lithium-ion [6], lithium LFP [7], lithium NMC [8], sodium-ion [9], vanadium flow [10], etc., hydraulic pumping systems, hydrogen, flywheels, supercaps, etc. In the context of hybridisation with photovoltaic and wind power generation, BESS systems are currently the clear winners in terms of technological and industrial maturity, performance, efficiency, etc. [6]. This is true at least for providing short-term supply capacity (generally set at 4 hours [11]), with other technologies mentioned above being considered for longer-term needs.

HYBRID STRUCTURES: RENEWABLES + BESS

An important aspect of hybridisation is the structure of energy storage integration. Such hybridisation can be carried out as stand-alone assets, where there are renewable generation plants (photovoltaic or wind) and BESS plants, acting as independent resources with different connection points to the electricity grid. In this scenario, it is the grid management mechanisms or, where applicable, the grid operator that is responsible for allocating the power levels of each resource in order to take advantage of the benefits of integrating the BESS. Another option is to hybridise at the plant level, or in other words, what are known as hybrid plants. In this case, each plant, whether solar or wind, integrates storage internally at its grid connection point, enabling it to act as a single resource and self-manage the renewable energy source and the BESS charging/discharging capacity. Within the context of a hybrid plant, there are two options for integrating storage: direct current (DC) or alternating current (AC) coupling. The choice of one or the other scheme will determine the number of converters and their sizing. In the case of DC coupling, which is mainly applicable in photovoltaics (PV), there is a DC bus to which the PV panel and the BESS are connected, usually via a DC/DC power converter. The connection between the DC bus and the electricity grid is made via a DC/AC converter shared by both resources. Broadly speaking, this option offers the advantages of potential efficiency improvements and the ability to oversize the PV resource relative to the DC/AC converter in order to charge batteries during peak generation hours. However, as it is a less modular structure, its control/management is less flexible. In the case of AC coupling, each resource (PV panel or wind turbine) has its own converter with AC output for connection to the common coupling point, and similarly, each BESS has its own PCS (DC/AC converter) with connection to that point, generally at medium voltage. This structure can be considered more flexible both in terms of control/management, with a clearer separation of controls and protections, and in terms of scalability. However, it requires more conversion stages, which can penalise efficiency. In reality, no architecture is universally better: the choice depends on the connection point, the operational objectives (discharges, ramps, service provision), the regulatory framework and the control strategy. In any case, integrating storage adds flexibility and margin of maneuverer to system operation in scenarios with high renewable penetration and lower synchronous generation, converting part of the resource’s variability into operable capacity.

THE CONVERTER AND ITS CONTROL: LIMITATIONS OF GRID-FOLLOWING OPERATION

Thus far, we have discussed aspects related to energy resources, such as available power and regulation capacity, in scenarios involving conventional generation, renewables, and renewables plus storage. A crucial and qualitatively different factor remains to be addressed: the grid connection interface, i.e., the synchronous machine vs. the power electronics-based converter or Power Electronic Converter (PEC). The synchronous machine couples a rotating mass to the grid, whose mechanical speed determines the electrical frequency. When there is an imbalance between mechanical and electrical power, kinetic energy exchanges power with the grid according to inertial dynamics. The converter, on the other hand, connects the source using power electronics: its contribution to voltage, current, and power is not imposed by mechanical dynamics, but by the control algorithm and its current limits, which determine how it responds during disturbances.

When renewable energy supply was low, plants connected via converters usually operated in Grid-Following (GFL) mode. It was assumed that the connection point had a “strong” grid with a high short-circuit ratio (SCR), meaning low impedance. The converter was synchronized with the measured voltage (typically using a PLL). The main objective in this scheme was to inject active and reactive power according to setpoints, effectively acting as a controlled current source on a grid whose voltage and frequency were “given” by the system.

As the number of connected converters increases, this assumption becomes less universal. In many places, the grid weakens, resulting in a lower SCR. Consequently, disturbances, load jumps, and variations in power transfer generally generate voltage variations. Furthermore, if synchronous generation is reduced, the system’s inertia will also be reduced, resulting in frequency deviations with steeper slopes (RoCoF) and greater magnitude. Voltage synchronization becomes more delicate. In “weak” networks, studies have been conducted on how the robustness of GFL algorithms is compromised [12]: the signal to which they synchronize is more vulnerable, control interactions emerge, and the behavior of the whole can deteriorate, increasing the risk of disconnections and contributing negatively to overall stability. However, converters in GFL mode remain the majority today, and the functionalities required for system stability are becoming increasingly demanding. These functionalities can be summarized as follows:

• Remain connected when under- or over-voltages occur within certain magnitude and duration limits. Additionally, the converter must provide voltage support by injecting reactive current (capacitive or inductive) in proportion to the fault’s magnitude. These responses require fast dynamics, meaning they must be performed autonomously by each converter based on the voltage measurement at its connection point within the plant.
• Provide voltage support through reactive power inputs. This is generally done at the plant level via Droop Q-V controls, whose response instructs the inverters to provide a level of reactive power proportional to the voltage deviation measured at the connection point.
• Provide frequency support through active power regulation. These functionalities are usually grouped under the concept of Fast Frequency Response, the most typical examples are Droop P-f controls, which respond by varying the active power injection in proportion to deviations in the estimated frequency, or synthetic inertia emulation, which seeks to “imitate” the inertial response of synchronous machines to compensate for the loss of inertia that the system is suffering with the reduction in synchronous generation.

GRID FORMING OPERATION

From the perspective of the converter control algorithm, the massive penetration of solar and wind energy into the electrical system, as well as the future paradigm of 100% renewable generation, calls for Grid-Forming (GFM) operation as a solution. The converter no longer acts as a source of current synchronized with the measured voltage, but as a source of voltage that controls its own voltage and frequency through synchronization based on power transfer. Without going into detail on the different alternatives for GFM control structures (Droops, VSG, synchronverter, VCO, etc.), there are numerous studies in the technical publications on GFM strategies and their comparison with GFL. These studies point to GFM as an operation that offers greater robustness in weak networks. They also motivate its deployment in the electrical system as the number of converters increases.

However, the GFM solution is not “magic” compared to GFL. A GFM converter aims to behave like a voltage source. As such, its ability to maintain the voltage in the face of disturbances critically depends on its current margin in both steady states and, more importantly, transients. When a voltage source reaches its current limit (saturation), it loses its nature as a voltage source. It can no longer regulate the voltage within margins, and its behavior is dominated by the limitation. Regarding the first point, to ensure long-term GFM operation, some type of energy buffer is necessary, such as storage. This is not a new concept, as the necessity of storage has already been justified with the integration of renewable energies. Regarding the second point, the main lines of action include careful equipment sizing, detailed current limit characterization (magnitude and time profiles), and control-based current limitation strategy development. The last aspect is probably one of the central focuses of GFM research: designing fast and effective limiters that allow saturation to be overcome quickly while preserving the voltage source behavior as much as possible without resorting to PLL-based synchronization schemes.

Returning to the functionalities required of GFL systems summarized above, many of them are intrinsic to GFM operation, although they depend on the implementation adopted. Requiring specific parameterizations for some voltage and/or frequency regulation functionalities, such as droop values or virtual inertia values, may conflict with the design criteria adopted from the point of view of the control algorithm. For example, in many GFM implementations, the dynamics of power setpoint tracking, the P-f droop value, and emulated inertia are closely related and interdependent, so it is important to understand that their values cannot be adjusted independently, but above all, it is important to consider whether it makes sense to do so. In line with this, given the planned deployment of GFM converters, the operators and technical bodies responsible are working on updating requirements to guide them toward this new operation.

CONCLUSIONS

Ultimately, when analyzing the stability of the system using technical criteria, the relevant question is not whether renewables are “stable” by nature, but rather what measurable benefits they provide under which grid conditions. The transition to a more distributed grid with less synchronous presence highlights two clear limitations: the variability of the resource and the limitations that may arise from the power converter-based interface and its control. Operating renewables with storage converts some of the energy variability into operating capacity, and shifting from grid-following schemes to grid-forming strategies improves the robustness of weak grids and enables active voltage and frequency support. Therefore, rather than being an automatic source of instability, renewables can be part of the solution as long as the formulation of requirements and management of all connected resources and elements are consistent with the grid scenario we are developing.

REFERENCES

1. Javier Serrano Delgado. Stability Problems in Electrical Networks.” Norvento Blog, (February 2026).
2. ENTSO-E. “Distributed Flexibility and the value of TSO/DSO cooperation.” (2017).
3. ENTSO-E. “Research, Development & Innovation Roadmap 2024 – 2034.” (2024)
4. NREL. N. Kumar, P. Besuner, S. Lefton, D. Agan, and D. Hilleman. “Power Plant Cycling Costs.” (2012)
5. Energy Procedia. Yuan-Kang Wu, Yi-Wen Wang. “Literature Review Concerning the Cycling Cost in a Power System with Renewable Power Sources” (2019)
6. Tha’er Jaradat, Tamer Khatib. “A review of battery energy storage system for renewable energy penetration in electrical power system: Environmental impact, sizing methods, market features, and policy frameworks.” (2025)
7. Tao Chen, Man Li and Joonho Bae. “Recent Advances in Lithium Iron Phosphate Battery Technology: A Comprehensive Review” (2024).
8. Farish Irfal Saaid et al. “Ni-rich lithium nickel manganese cobalt oxide cathode materials: A review on the synthesis methods and their electrochemical performances” (2024)
9. Shilin Zhang. “Batteries for Grid-Scale Energy Storage Applications” (2025)
10. Muhammad Shoaib et al. “Advances in Redox Flow Batteries – A Comprehensive Review on Inorganic and Organic Electrolytes and Engineering Perspectives” (2024)
11. Denholm, Paul, et al. “Moving Beyond 4-Hour Li-Ion Batteries: Challenges and Opportunities for Long(er)-Duration Energy Storage.” (2023)
12. Wang, Xiongfei et al. “Grid-Synchronization Stability of Converter-Based Resources – An Overview” (2020)
13. B. Bahrani et al. “Grid-Forming Inverter-Based Resource Research Landscape: Understanding the Key Assets for Renewable-Rich Power Systems,” Modification (2024)
14. L. Meng et al. “Fast Frequency Response From Energy Storage Systems—A Review of Grid Standards, Projects and Technical Issues,” (2020)
15. Red Eléctrica. “Electricity System Report 2024.”