Hybridization of Distributed Energy Resources in a Steelworks

The steel industry is currently in the midst of a profound transformation aimed at decarbonization. This shift involves the electrification of various production processes that currently depend on coal or gas for power. Additionally, there are plans to integrate wind and photovoltaic generation systems into the grids supplying steelworks to align with global CO2 emission reduction goals. The combination of these two generation systems offers notable benefits, alongside several drawbacks, primarily concerning the loss of inertia characteristic of conventional synchronous generators. In this paper, two main loads are examined because they are responsible for a significant portion of the existing electricity consumption in the steelworks: the hot rolling mill and the electric arc furnace. Concurrently, global demand forecasts are assessed, taking into account the injection of power from a wind farm and a photovoltaic plant connected to the distribution network supplying the steelworks. The analysis of the demand curve enables the projection of substantial power fluctuations capable of affecting frequency control, given the reduced equivalent inertia of the power system following the integration of renewable resources. To mitigate these power variations and enhance frequency control, the advantages of a battery-based storage system and the coordinated management of resources are scrutinized.


Hybridization of Distributed Energy Resources in a Steelworks
Gonzalo Alonso Orcajo , Member, IEEE, José M. Cano , Senior Member, IEEE, Joaquín G. Norniella , Francisco Pedrayes Gonzalez , Carlos H. Rojas , and Josué Rodríguez Diez Abstract-The steel industry is currently in the midst of a profound transformation aimed at decarbonization.This shift involves the electrification of various production processes that currently depend on coal or gas for power.Additionally, there are plans to integrate wind and photovoltaic generation systems into the grids supplying steelworks to align with global CO 2 emission reduction goals.The combination of these two generation systems offers notable benefits, alongside several drawbacks, primarily concerning the loss of inertia characteristic of conventional synchronous generators.In this paper, two main loads are examined because they are responsible for a significant portion of the existing electricity consumption in the steelworks: the hot rolling mill and the electric arc furnace.Concurrently, global demand forecasts are assessed, taking into account the injection of power from a wind farm and a photovoltaic plant connected to the distribution network supplying the steelworks.The analysis of the demand curve enables the projection of substantial power fluctuations capable of affecting frequency control, given the reduced equivalent inertia of the power system following the integration of renewable resources.To mitigate these power variations and enhance frequency control, the advantages of a battery-based storage system and the coordinated management of resources are scrutinized.Index Terms-Electric arc furnace, energy storage, industrial plants, photovoltaic power systems, steel, wind energy.

I. INTRODUCTION
E NHANCING the efficiency of industrial facilities, curbing their greenhouse gas emissions, and promoting generation based on renewable sources are key priorities advocated by various countries concerning energy usage within these facilities.Achieving these goals without compromising competitiveness and sustainability is essential.
The steel industry is one of the largest industrial electricity consumers, ranking second in terms of energy demand (7% of worldwide consumption).Electricity costs can account for 15-20% of the fixed costs in a medium-sized industrial steelworks (4-5 Mt/year).Moreover, the steel sector ranks first in terms of CO 2 industrial production, being responsible for 5.7% of the EU global greenhouse gas emissions.
This leadership stems from the steel sector being the largest industrial consumer of coal, accounting for approximately 75% of its energy needs.Moreover, the fixed cost associated with emission rights substantially affects the pricing of final products.Since 2015, the cost of emission rights has surged exponentially.With global steel demand projected to rise by over a third by 2050, CO 2 emissions are expected to increase significantly.Consequently, implementing measures to align sustainability with emission reduction regulations has become imperative.As a result, the steel industry is currently undergoing an ambitious renewal process to uphold its competitiveness while striving to achieve the goal of zero net emissions.[1], [2], [3].
Significant changes in steel production processes are required to achieve climate neutrality by 2050.The main initial modifications involve a) the transition from coal-based to natural gas-based furnaces, as precursors to green hydrogen-based ones, in order to produce direct reduced iron (DRI); b) the shift from high-emission fossil fuel-based to low or zero carbon energy sources.Green hydrogen, circular forms of carbon, and carbon capture, utilization, and storage (CCUS) technologies will be used; c) the enhancement in the use of electric arc furnaces and scrap, and d) the prioritization of clean electricity consumption [4].Steel production can follow two distinct pathways: the primary, or integrated route, and the secondary, or electric route.The primary route, which relies on blast and basic oxygen furnaces, dominates global crude production, constituting nearly 71% of unrefined steel manufacturing.In contrast, the secondary route employs an electric arc furnace (EAF) to melt scrap, resulting in substantially lower CO 2 emissions compared to the integrated route.Another strategy to mitigate CO 2 emissions associated with the integrated route involves using a different input for the EAF: direct reduced iron (DRI), which is produced using natural gas.[5], [6].
In the coming decades, the electrification of processes in the interests of decarbonization and the prioritization of clean electricity consumption will lead to steelmaking facilities being incorporated into distribution networks with a high penetration of renewable energy sources, as well as to them demanding a huge amount of electricity mainly due to the use of large EAFs.This scenario presents several technological challenges.
Analyzing and addressing several of these challenges constitute the main focus of this paper.
Inertia for grid frequency control is mainly provided by conventional power plants.In the event of frequency fluctuations due to mismatches between generation and demand, synchronous generators either inject or absorb power to balance such mismatches.However, wind and solar generators, lacking inertia, make grids more susceptible to abrupt frequency fluctuations.This problem is exacerbated in the presence of large loads with demanding dynamics such as EAFs or hot rolling mills (HRM).In this context, the integration of large EAFs into existing grids as part of decarbonization strategies is anticipated.However, the frequency control of these grids may not always meet the stringent demands of electro-intensive steelworks [7], [8].
In this context, a battery energy storage system (BESS) can be used for load leveling, peak shaving, frequency regulation, capacity firming, power quality improvement and spinning reserve purposes [9], [10], [11].Load leveling is an effective method to control storage systems for voltage support in power networks [12].Centralized BESSs are used for load leveling and spinning reserve in countries where the installed capacity of pumped-storage hydroelectric plants is low [13].The correct management of BESSs significantly reduces losses in distribution and transmission networks [14].These systems have also been successfully used in industrial plants [15].Other references describe applications where storage systems jointly perform actions such as frequency control and load leveling [16].
The introduction of new agents into the distribution networks supplying steel plants will bring about substantial changes in power flow and its dynamics within these facilities.To date, there is no documented instance of the interplay between the demand curves of two primary electricity consumers in steel production, HRM and EAF, and the generation curves of renewable sources like wind and photovoltaic systems, which have the potential to reduce carbon emissions.The examination of possible solutions to enhance frequency control within the steel industry, anticipating a significant loss of inertia in the power system alongside increasingly pronounced power variations, is both timely and necessary.Notably, there is a dearth of existing studies that have addressed this specific issue, underscoring the necessity of such research.Examining the potential role of battery energy storage systems (BESS) in mitigating load and frequency fluctuations within the steelmaking context presents another avenue for research that is intended to explore in this study.Fig. 1 shows the research flow diagram.
The primary contributions of this research can be summarized as follows: 1) Conducting an analysis of two major loads within a steelmaking facility, one 60-MW HRM and two 80-MW EAFs.The study establishes production connections between these loads, determining their rated power and periods of demand.Models are proposed to replicate their power demands based on their characteristics and observed power variation levels in a standard and realistic steel plant production environment; 2) Similarly, establishing models for active and reactive power injection from a 30-MW wind farm (WF) and a 30-MW photovoltaic plant (PV) connected to the distribution network.Behavioral patterns of these injection curves are determined based on key variables such as wind speed, irradiance, temperature, and system characteristics involved in energy conversion.Significant power variations are identified in realistic scenarios as well; 3) Defining the configuration of a 20-MW BESS capable of providing support to the previously described generation and demand elements; 4) Analyzing the overall power demand of the steelworks distribution network without the involvement of the storage unit; 5) Establishing a coordinated management procedure for the available power resources of a virtual plant (VP) including the power network of the steelworks, the WF, the PV and the participation of the BESS, in realistic scenarios; 6) Identifying opportunities for improvement in controlling observed power and frequency variations.
The analysis is performed by means of Simulink/Matlab simulation models based on actual production and consumption data [17].
The paper is organized as follows.The HRM and the EAF are respectively described in Section II.The WF, the PV, and the BESS under study are respectively described in Section III.The VP is outlined in Section IV, and the case study is analyzed in Section V. Finally, the conclusion of this work is included in Section VI.

A. Hot Rolling Mill
The original HRM under study mainly consists of eight rolling stands, two roughing and six finishing, as seen inFig.2. The rated power of the stand motors varies between 7.5 and 8 MW (63 MW overall) and their associated drives are based on 12-pulse cycloconverters.The HRM is connected to a 132-kV distribution network through a 75-MVA transformer.A passive filtering system with four branches is used for harmonic cancellation and displacement factor adjustment, given the drives non-linear nature.This filtering system can provide up to 45.6 Mvar at rated voltage [18].
The various measurements utilized in this paper are based on real records of an actual HRM.In this context, Fig. 3 illustrates the fluctuation in active power demand by the original HRM Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.throughout a rolling campaign.This includes periods of operation exclusively by the roughing mill, solely by the finishing mill, and concurrently by both.A 15-MVA static synchronous compensator (STATCOM) is incorporated into the original HRM exclusively for reactive power management support.The STAT-COM controls the power factor of the installation in conjunction with both the passive filtering system and rolling stands F4, F5, and F6, i.e., those based on circulating-current-mode cycloconverters.
Moreover, the STATCOM can provide voltage support up to +3.6% of the PCC rated voltage when considering a reactive power injection practically equal to its rated power (15 MVA) and a reactance of 0.179 pu between the PCC and the distribution busbars to which the rolling stands are connected.Fig. 4 shows the residual demand of reactive power resulting from the joint action of the passive filtering system and the STATCOM.
The HRM model utilizes averaged models to replicate the power flow within each stand of the roughing and finishing mills throughout a rolling campaign.These models accurately reproduce the current demand, including its harmonic components.Furthermore, the model includes detailed models of the passive filtering system and the STATCOM.Validation of the model accuracy is conducted by comparing its current and power responses with real-world measurements.

B. Electrical Arc Furnace
Electric arc furnaces are key elements in the transformation that the steel industry must undergo in the coming years [19], [20], [21].
The rated power of EAFs must be adapted to production targets.The capacity of the HRM under study is 470-530 slabs/day with an average weight of 22-26 t/slab, which means a maximum production of 5 Mt/year, 13.78 kt/day or 574.16 t/h of steel.This production requires 5.2 Mt/year of liquid steel if a conversion factor of 1.0417 from liquid to crude steel is considered.Therefore, 14.246 kt/day of liquid steel must be produced, almost 50% of which (i.e., 7.2 kt/day) is proposed to be obtained by using EAFs.For this purpose, two 180-t/heat EAFs are considered, which means that each EAF must manage roughly 20 heat/day and, therefore, that their associated tap-totap time is 72 minutes.Because 2 or 35-minute charging buckets are needed a 60-minute ON time and a 12-minute OFF time are assumed.
The estimated energy consumption of each EAF is 420 kWh/t, i.e., 75.6 MWh/heat.Therefore, the estimated effective power is 75.6 MW/heat for a 60-minute ON time, which means an 80-MW/heat active power if a performance ratio close to 0.9 is assumed.Consequently, the rated power of each EAF can be 100 MVA if a power factor of 0.8 is considered.Finally, a transformer with a rated power of 150 MVA is selected for the connection of the EAFs to the 132-kV distribution network and two 80-MW, 180-t EAF is connected to the grid under study.Fig. 5 shows the evolution of the active and reactive power demanded by the EAF throughout a complete melting and refining cycle of three scrap baskets.
The EAF is modelled including the HV/MV (132/34.5 kV, 150 MVA) and MV/LV (34.5/0.7 kV, 150 MVA) transformers, as well as the electrodes, which are replicated as an RL impedance.The EAF is defined as a current-controlled voltage source based on models previously described in the literature [22].
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.The active and reactive power demand is approximated using evolution masks, which accurately replicate the average active power value at the fundamental frequency during the phases of boring, melting, and refining.The influence of harmonic currents is also considered as it depends on the length of the electric arc, which is determined by the condition of the scrap metal, the specific cycle stage, and whether the initial or final minutes of the cycle are considered [23].
Figs. 6-8 show details of the voltage, current and power demand involved in the operation of the EAF.The 132-kV power network feeding the EAF is supported by a dynamic reactive power compensation system based on a Static Var Compensator (SVC) and a STATCOM to prevent the transfer of reactive power variations upstream and the consequent disruption of the normal operation of other processes in the steelworks.

III. RENEWABLE ENERGY SOURCES AND STORAGE UNIT
The hybridization of the WF and the PV, along with the coordinated management of the storage unit, offer the opportunity to optimize available energy resources.

A. Wind Farm
In order to provide the VP with generation systems based on renewable sources, a 30-MW WF is connected to the grid feeding the steelworks.This WF comprises 20 1.5-MW turbines (30 MW overall) based on a doubly-fed induction generator (DFIG), allowing for autonomous control of both active and reactive power within the specified limits determined by the generator rated power and the back-to-back converter.An average wind speed of 7.5 m/s is projected for the chosen location.From this average wind speed, the power-speed curve of the wind turbine, and the Weibull wind speed distribution curve, the average annual electrical energy production of each wind turbine can be estimated (see Fig. 9, including the wind power duration curve as well, which shows that there is no significant wind at the chosen location for approximately 1900 hours per year).Accordingly, each turbine can generate up to 4.16 GWh/year, in which case the WF production is 83.38 GWh/year, i.e., 22.8% of the annual energy consumed by the HRM [24], [25], [26], [27], [28].CO 2 -free power generation from the WF is complemented by that generated from the PV.
The simulation model for the WF relies on a detailed representation of a 1.5-MW wind turbine equipped with a DFIG.Its primary input variable is the forecasted wind speed at the designated site.Subsequently, the 30-MW WF model is constructed based on this detailed turbine model, employing an averaged approach that scales proportionally in power.This WF model incorporates the distribution network, integrating line and transformer models.
The WF has an energy reserve to supply reactive power either capacitive or inductive depending on the corresponding requirements.This energy can be provided by both the induction machine and the grid-side converter (GSC).The GSC, which normally manages between 20 and 30% of the wind turbine rated power, can be oversized to increase its energy reserve.Fig. 10 shows the evolution of wind speed over the course of half a day at the site under study, whereas Fig. 11 shows the evolution of the active and reactive power injected by the WF into the grid over the same time interval.

B. Photovoltaic Plant
The values of the PV rated and peak power are, respectively, 30 MW and 38.57MWp.The PV system comprises twenty 1.5-MW, 60-Hz inverters, each containing 4704 solar panels rated Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.at 410 Wp and four maximum power point trackers (MPPTs).The panels are arranged in a configuration where each MPPT manages 49 strings (as shown in Fig. 12), distributed across five parallelized inputs.Each input handles 9 or 10 strings, resulting in each string consisting of 24 panels in series.The rated output voltage of each inverter is 690 V.
The daily solar production mainly depends on both the solar irradiance and the ambient temperature.Considering the location of the PV, and from average solar irradiance and temperature databases, the monthly electrical energy production can be predicted, as seen in Fig. 13.The PV production is 45.18 GWh/year, i.e., 12.35% of the annual energy consumed by the HRM.Therefore, over 35% of the electrical energy consumed by the HRM can be supplied from renewable sources.The simulation model of the PV comprises two key components: a detailed model of a reference subinverter of 390 kW, and an averaged model controlled by the power injected into the grid by the reference subinverter.This averaged model supplements the remaining 29.61 MW under standard measurement conditions, thus achieving the targeted 30-MW output.The detailed model integrates input variables such as irradiance and temperature forecasts over a day.Within the detailed model, the solar generator is delineated, incorporating characteristics of selected panels and string configurations, along with the maximum power point tracker, inverter, and interconnection filter with the Point of Common Coupling (PCC).
Fig. 14 shows the evolution of irradiance and temperature over the course of half a day.The evolution of the active and reactive power injected by the PV into the grid over the same time interval can be seen in Fig. 15.The reactive power setpoint is adjusted throughout the whole time interval for the power factor at the PCC to be as close to unity as possible.The setpoint can also be adjusted for the solar inverter to manage reactive power.The limits imposed by the inverter rated power must be respected and the active power injection must be prioritized under normal operating conditions according to the existing irradiance.The power injected into the grid changes as irradiance and temperature vary.
Considerable variations in the electricity production of both the WF and the PV may occur due to the inherent variability of wind and solar resources.For example, Fig. 16 shows the PV power generation over two separate days with markedly different solar irradiance.These power variations modify the load curve of the synchronous generators, which transiently affects their speed and the system frequency.

C. Battery Energy Storage System
The BESS is made up of 20 arrays of batteries (BAs), each with a capacity of 1 MW and 2.5 MWh.Each array includes 5 battery boards (BBs) and 15 battery clusters (BCs), each cluster having a capacity of 165.88 kWh, a voltage of 604.8/766.8V, and a rating of 240 Ah.Each cluster is composed of 18 battery modules, each with a capacity of 240 Ah and a voltage of 33.6/43.2V, arranged in series.Each battery module (BM) consists of 24 lithium single cells (SCs) with capacities of 120 Ah and voltages of 2.5/3.65 V, organized into two parallel strings of 12 SCs each (Fig. 17).The bidirectional converter AC/DC is a three-level active front-end rectifier connected to a 34.5-kV grid via a transformer [29], [30].This converter adapts to the grid voltage, sequence and frequency while controlling the demand/injection of current at the fundamental frequency, as well as its displacement factor and harmonic content, and the DC bus voltage.The DC/DC conversion at the converter output enables the control of the BESS charge and discharge (Fig. 18).For this purpose, several bidirectional converters control the BC voltage and current.The dual active bridge and the interleaved dc-dc are the most common DC/DC converters for managing the energy stored in the BA.
Fig. 19 shows the evolution of the active and reactive power of one of the 1-MW storage units when a setpoint modification occurs (from charge to discharge at t = 18 s).Fig. 20 shows the evolution of the phase voltage and line current under such circumstances.
A detailed model simulates the functionality of the 1-MW Neutral Point Clamped (NPC) three-level converter, serving as the interconnection component with the distribution grid.Building upon this detailed model, an averaged 19-MW model Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply. is constructed to replicate the operation of the remaining 19 storage units within the system.

IV. VIRTUAL POWER PLANT
The VP includes two high-demand loads (the EAF and the HRM), two power generation systems (the PV and the WF), the BESS, and the 220-kV PCC.The VP seamlessly integrate and optimize renewable energy sources, storage units, and electrical demand management to bolster steel production while achieving sustainability, efficiency, and CO 2 emission reduction objectives.Fig. 21 shows the single-line diagram of the VP.
The hybridization of the WF and the PV is of particular interest in locations where wind and solar resources are largely complementary to each other either diurnally or seasonally.The concurrent evacuation of energy from both installations through a shared substation in the VP, and at the same PCC, yields technical and economic advantages.These include: a) reduced grid connection costs and times; b) improved utilization of land and infrastructures; c) optimized operation and maintenance tasks; and d) enhanced quality and stability of electricity supply.Additionally, the integrated control and monitoring system the of overall power to maximize grid access at all times.
The power demand at the 220-kV network is analyzed without the contribution of the BESS.The generation and demand profiles determine the evolution of that power demand over time.
The demand profile of the EAF exhibits a high degree of flexibility.When the melting and refining cycles coincide with diurnal patterns or periods of forecasted high average wind speed, local generation mitigates the need for power from the 220-kV grid.As a result, substantial technical benefits (such as forecasting transmission and distribution losses and ensuring node voltage stability), economic advantages (through local generation with low CO 2 emissions), and environmental benefits (from renewable sources generation) are realized.Generation of energy from local systems enables a certain flexibility in the management of reactive power.Active power injection mainly depends on weather conditions and on day/night and seasonal cycles.Adjusting the inclination angle of solar panels optimizes the capture of maximum solar irradiance under typical conditions.This angle can also be altered to decrease solar energy production during periods of excess generation.Similarly, the setpoint of the DC/DC converter between the solar generator and the inverter can be modified to achieve the same objective.On the other hand, adjusting the blade pitch angle enables wind turbines to reduce the power injected into the grid when necessary.
The BESS charging/discharging processes are controllable and bring clear advantages to the management of active and reactive power.

V. CASE STUDY
The VP operation without the participation of the BESS is analyzed in the first place.The active and reactive power demand profiles are established for the course of a day.Although the demand and weather conditions are specific for one day, the order of magnitude of the power variations to be expected in the VP can be detected.The evolution of the reactive power in the grid and the order of magnitude of the displacement factor can be analyzed in a similar manner.The role played by the BESS regarding load levelling and frequency stability is analyzed in the second place.

A. Operation of the VP Without the BESS
The operational parameters of both the EAF (Fig. 22) and the HRM (Fig. 23) heavily influence the installation power demand under normal conditions.Due to the need to reduce the data representation load, a low sampling frequency was employed to generate both figures, each spanning 10 hours.However, it is essential to consider power consumption from other loads within the steelworks, which fluctuates over time.On average, these loads require approximately 20 MW.
The WF and the PV are assumed to inject active power into the grid according to the generation profiles in Figs.11 and 15.The evolution of the active power at the 220/132 kV substation under these circumstances is shown in Fig. 24.A great variability can be observed: power demand is extremely low in certain time intervals, in which case the distribution and transmission grids work with low load indices, whereas the demand is very high in other time intervals, the grids thus working with high load indices.

B. Operation of the VP With BESS: Load Leveling
Load leveling by using a BESS enables a better utilization of the distribution grid, thus preventing the transmission network from overloading, and improves node voltage profiles in reducing power variations [31], [32], [33].
The BESS must be charged when the VP active power demand is low.In such scenarios, the WF and PV systems consistently maintain reasonable production levels, and heavy loads do not require excessively high power.However, during periods of elevated active power demand in the VP due to the absence of local generation, either from low wind or insufficient irradiance (Fig. 25), BESS must discharge (Fig. 26) to supply loads with stored energy, thereby reducing the power demand from the transmission grid.
In terms of energy, 27.5 MWh are stored and 5 MWh are discharged, which yields a positive storage balance of 22.5 MWh.The adjustment of the upper and lower limits that determine the BESS charge and discharge periods enables a daily compensation of these periods and, therefore, the extension of the useful life cycle of the batteries.

C. Operation of the VP With the BESS: Frequency Regulation
Present energy and environmental scenarios promote the incorporation of power generation systems based on renewable sources into the distribution grids of electro-intensive industries.Despite the many technical, economic and environmental advantages of this solution, there are still various drawbacks to it.Wind and PV converters normally operate according to the setpoints provided by maximum power tracking algorithms without any contribution to inertia.A significant loss of inertia can bring the grid frequency and/or the Rate-of-Change-of-Frequency (RoCoF) to unacceptable values, producing uncontrolled disconnections of both loads and generation stages, which can lead to system collapse [34], [35], [36], [37].In January 2019, Spain boasted an installed photovoltaic capacity of 4800 MW, which steadily increased to 25126 MW by January 2024, constituting 20% of the total installed capacity of 123776 MW.Concurrently, wind capacity reached 30748 MW by the same date, with wind energy contributing 25% to the energy demand over the preceding year.Together, solar and wind generation now account for over 40% of the global installed capacity [38].At the same time, the inertia in the rotating masses of synchronous generators and turbines has decreased due to the decommission of coal-fired power plants.
In the case study, the BESS is charged or discharged in response to changes in grid frequency.It responds to an increase in grid frequency by charging and a decrease by discharging.The first scenario assumes a sudden loss of renewable energy generation at the onset of the melting process.Consequently, the grid frequency transiently decreases until it returns to its nominal value (Fig. 27).
The VP is considered to be connected (in isolated grid mode) to a 400-MVA, 60-Hz equivalent synchronous generator with 2 pairs of poles.The order of magnitude of power variations is up to 50% of the generator rated power (200-MW), including the loss of local generation (40-MW) and the increase in power demand.The 20-MW contribution (5% of the rated power) from the BESS reduces power variations to 45% and frequency variations by more than 12%.In the second scenario (Fig. 28), a transient increase in frequency occurs due to a substantial loss of load, specifically attributed to the conclusion of cycles in the EAF and the HRM, amounting to 160 MW.Simultaneously, there is a rise in power generated by local systems, with the WF and PV systems injecting a combined 40 MW into the grid.To mitigate the load loss and constrain the Rate of Change of Frequency (RoCoF) and zenith values, the 20-MW BESS demand serves to compensate.
The observed frequency variations depend on several factors: fluctuations in power demand, the inertia of synchronous generators, and the speed and power control systems regulating their frequency.Additionally, the penetration rate of solar and wind technologies in the distribution grid, along with their synthetic or virtual inertia, play crucial roles.Battery storage systems emerge as a viable solution capable of ensuring frequency stability in steel environments under significant stress due to abrupt power variations resulting from new load regimes and the loss of grid inertia.

VI. CONCLUSION
Conventional steelworks must undergo a deep transformation in the coming years for decarbonization purposes.The incorporation of electric arc furnaces and the electrification of a large part of the production processes are to increase electricity consumption in an already electro-intensive industry.The inclusion of wind and photovoltaic generation systems in the supply networks of steelworks is currently being considered to support the aforementioned electrification without increasing CO 2 emissions.The hybridization of both systems offers significant technical, economic, and environmental advantages.
It is crucial to investigate solutions that reconcile the low inertia inherent in modern power systems, which increasingly rely on non-synchronous generation sources such as wind and photovoltaic systems, with the significant variability in power demand resulting from the integration of fluctuating loads like electric arc furnaces and rolling mills.Steelworks must collaborate closely with grid operators to seamlessly integrate these new loads, ensuring compatibility and enhancing frequency control amidst evolving operational conditions.This paper proposes the incorporation of an energy storage system capable of mitigating load-induced power fluctuations as one potential solution.Additionally, research directions include exploring the integration of synthetic inertia into renewable energy systems, analyzing operational dynamics across varying penetration levels of wind and solar technologies, and investigating frequency control mechanisms, particularly during periods of low demand at night or high solar PV penetration during the day.These research directions aim to address the challenges posed by the changing energy landscape.

Fig. 4 .
Fig. 4. Reactive power demand resulting from the incorporation of the STAT-COM.

Fig. 5 .
Fig. 5. EAF power demand during the melting and refining of three scrap baskets.

Fig. 6 .
Fig. 6.Detail of the evolution of electrode phase-neutral and line voltages.

Fig. 7 .
Fig. 7. Detail of the evolution of the EAF electrodes line current.

Fig. 8 .
Fig. 8. Detail of the evolution of the EAF active power demand.

Fig. 10 .
Fig. 10.Evolution of wind speed, v wind , over half a day.

Fig. 11 .
Fig. 11.Evolution of active and reactive power injected by the WF over half a day.

Fig. 12 .
Fig. 12. Configuration of one of the twenty 1.5 MW solar inverters that are part of the solar plant.

Fig. 14 .
Fig. 14.Evolution of irradiance and temperature over half a day.

Fig. 15 .
Fig. 15.Evolution of active and reactive power injected by the PV over half a day.

Fig. 16 .Fig. 17 .
Fig. 16.Evolution of active power injected by the PV over two days with different solar irradiance.

Fig. 18 .
Fig. 18.Topology of the main components of the BESS.

Fig. 19 .
Fig. 19.Evolution of active and reactive power of one 1-MW storage unit from charge to discharge.

Fig. 20 .
Fig. 20.Evolution of voltage and current of one 1-MW storage unit from charge to discharge.

Fig. 22 .
Fig. 22. Evolution of the active and reactive power demanded by the EAF.

Fig. 23 .
Fig. 23.Evolution of the active power demanded by the HRM.

Fig. 27 .
Fig. 27.Left: frequency variation under a sudden increase in power demand.Right: power injection from the BESS.

Fig. 28 .
Fig. 28.Left: frequency variation under a sudden decrease in power demand.Right: power injection from the BESS.