**Preface to "HVDC/FACTS for Grid Services in Electric Power Systems"**

Electric power systems are headed for a true changing of the guard, due to the urgent need for sustainable energy delivery. Fortunately, the development of new technologies is driving the transition of power systems toward a carbon-free paradigm while maintaining the current standards of quality, efficiency, and resilience. The introduction of HVDC and FACTS in the 20th century, taking advantage of dramatic improvements in power electronics and control, gave rise to unprecedented levels of flexibility and speed of response in comparison with traditional electromechanical devices. This flexibility is nowadays required more than ever in order to solve a puzzle with pieces that do not always fit perfectly. This Special Issue aims to address the role that FACTS and HVDC systems can play in helping electric power systems face the challenges of the near future.

The Special Issue is composed of 13 papers submitted from Asia and Europe that cover three major areas:

Review papers;

Transmission system applications;

Distribution system applications.

The first review paper by Dr. Kaushal et al. from KU Leuven gives an overview of the role that HVDC lines are playing on the provision of ancillary services in the current European Network of Transmission System Operators for Electricity (ENTSO-E). Transmission system interconnections play a significant role in decarbonized power systems to reduce the energy from fossil-based sources. From this point of view, controllable DC interconnections are key to foster this energy transition. In addition, the use of DC assets is also of interest on the distribution side. In this regard, the second review paper by Dr. Maza-Ortega et al. from the University of Sevilla analyzes the new possibilities that hybrid AC/DC networks may have in future last-mile distribution systems. The proliferation of DC devices (loads, renewable energy sources, and storage systems) in low voltage networks allows the envisioning of the coexistence of AC and DC networks, which may considerably reduce the use of auxiliary AC/DC power interfaces and open new possibilities from the ancillary service provision point of view. The last paper in the review section of this Special Issue highlights the importance of the controller and power hardware in the loop (CHIL and PHIL, respectively) testing practices. Dr. Kotsampopoulos et al. from the National Technical University of Athens underline the benefits and limitations of using these cutting-edge industry practices.

The second block of papers in this Special Issue is devoted to transmission system applications. The flexibility provided by voltage source converter (VSC) HVDC (VSC-HVDC) systems by means of the simultaneous controllability of active and reactive power flows is unquestionable. The provision of high-quality ancillary services, however, requires precise and advanced control algorithms that are capable of satisfying the most stringent requirements even with adverse network conditions, such as voltage unbalance. In this regard, the paper by Dr. Miao et al. proposes a novel control technique intended to reduce the active and reactive double frequency ripple caused by voltage unbalance. In addition, VSC-HVDC may increase the stability of the system in case of contingencies in complex power systems composed of AC parts as well as conventional line-commutated converter (LCC) HVDC (LCC-HVDC). This topic is explored in the paper by Dr. Yoon et al. which exemplifies the proposed control algorithm in a modified version of the IEEE 39 bus test system. The paper by Dr. Xiao et al., also related with LCC-HVDC, proposes a novel technique to improve the endurance capability of an AC system facing continuous commutation failures and to reduce the blocking risk of HVDC converters. Dr. Kim et al. also explore new controller applications for FACTS by presenting a neural network controller for reducing the oscillations between interconnected power systems using a unified power flow controller (UPFC). The provision of ancillary services by FACTS and/or HVDC devices may require the intervention of advanced communication infrastructures to provide remote measurements required by the control algorithms. The paper by Dr. Renedo et al. analyzes the impact that the communication delay may have on the transient stability of the system and how the critical clearing time is affected. All of these academic contributions focus on the application of new controllers to FACTS or HVDC devices in order to enhance the operation of the system by means of advanced ancillary services. However, it is also important to highlight that new contributions on alternative power electronics-based devices or components continuously emerge. This is the case of the DC circuit breaker with fault current limiting capability presented by Dr. Wang. Finally, this section closes with a detailed analysis of the installation of the first thyristor-controlled series capacitor (TCSC) installed in Korea for the mitigation of subsynchronous resonance. Dr. Park et al. compare in their paper the performance of a conventional fixed series capacitor compensation with the TCSC, which reveals the TCSC's superior performance.

The third block of papers of this Special Issue refers to distribution system applications where the focus is probably to enhance, as much as possible, the power quality perceived by the final user. In this regard, the paper by Dr. Dai proposes a three-phase, star-connected, Buck-type dynamic capacitor (D-CAP) for reactive power compensation and unbalance reduction. This Special Issue also includes the paper by Dr. Rodr´ıguez et al. that proposes advanced synchronous power controllers (SPC) capable of controlling unbalance currents during faults with the aim of balancing the voltages as much as possible and attenuating the harmonic distortion in steady-state conditions. Finally, this section closes with a paper devoted to the increasingly important electrified railway sector where the application of FACTS and DC technology may bring several benefits. The paper by Mr. Chen et al. proposes a novel hybrid power quality compensator to compensate the voltage unbalance and harmonic distortion.

In closing, we would like to thank the authors for their work and contributions to the development of new technologies in this area and the reviewers for their valuable comments, which have helped to finalize the papers. We also appreciate the continuous support of the Energies Editorial Office and especially Dr. Billy Bay, MDPI Managing Editor, who greatly facilitated our work in inviting the editors of this Special Issue.

> **José M. Maza-Ortega, Antonio Gómez-Expósito** *Special Issue Editors*

## *Review* **An Overview of Ancillary Services and HVDC Systems in European Context**

#### **Abhimanyu Kaushal 1,2,\* and Dirk Van Hertem 1,2**


Received: 17 July 2019; Accepted: 26 August 2019; Published: 9 September 2019

**Abstract:** Liberalization of electricity markets has brought focus on the optimal use of generation and transmission infrastructure. In such a scenario, where the power transmission systems are being operated closer to their critical limits, Ancillary Services (AS) play an important role in ensuring secure and cost-effective operation of power systems. Emerging converter-based HVDC technologies and integration of renewable energy sources (RES) have changed the power system dynamics which are based on classical power plant operation and synchronous generator dynamics. Transmission system interconnections between different countries and integrated energy markets in Europe have led to a reduction in the use of energy from non-renewable fossil-based sources. This review paper gives an insight into ancillary services definitions and market practices for procurement and activation of these ancillary services in different control areas within the European Network of Transmission System Operators for Electricity (ENTSO-E). The focus lies particularly on ancillary services from HVDC systems. It is foreseen that DC elements will play an important role in the control and management of the future power system and in particular through ancillary services provision. Keeping this in view, the capability of HVDC systems to provide ancillary services is presented.

**Keywords:** ancillary services; HVDC systems; loss management; frequency control; voltage and reactive power control; black start; congestion management

#### **1. Introduction**

In a vertically integrated power system, the main task of the system operator is to operate the power system in a reliable and secure manner. With unbundling, the vertically integrated power systems have been unbundled into generating units, transmission system operators (TSO), and distribution system operators (DSO). Electricity markets have been established for a transparent and cost-effective trade of energy. However, with energy trade, other so-called ancillary services are also exchanged between different market players. Ancillary services are the resources required by TSO for reliable and secure power system operation [1]. Important power system characteristics such as frequency, voltage, load and system restart process are maintained by these services [2]. The nomenclature for ancillary services varies in different parts of the world. For example in the USA, for PJM operator area ancillary services for frequency control are known as regulations and reserves [3] (operating, primary, synchronized and quick start reserves [4]), for CAISO (The California Independent System Operator) area the services are termed as regulation up, regulation down, spinning reserve and non-spinning reserve [5]. The ancillary services for frequency control has been categorized as regulation and contingency by Australian Energy Market operator [6] whereas in Europe as per guidelines on electricity balancing the frequency control services are known as frequency containment reserves, frequency restoration reserves and replacement reserves [7]. These all services enable respective system operator with same functionality i.e., frequency control in this example however the names

are different in different energy markets. As in this review paper the emphasis has been placed on ancillary services in European context so the terminology as used in Europe is considered.

Real-time power system operation involves several uncertainties and these uncertainties have been further increasing as a consequence of augmented integration of distributed power generation from RES. For secure power system operation in such a scenario, ancillary services market has gained critical importance. Ancillary services can be market-based or non-market-based [8]. Market-based ancillary services are procured by the TSOs from different stakeholders from electricity market [9]. In some control areas it is mandatory for power system entities to provide ancillary services with or without payment, these ancillary services are termed as non-market-based ancillary services. While the ancillary services have been defined by ENTSO-E for the interconnected European power system, their implementation, the method of procurement, and activation for these services varies in different member states [10,11].

In this paper, a review of ancillary services definitions, procurement, and implementation methods in different ENTSO-E areas is presented. Various methods through which the HVDC system elements can participate in providing ancillary services are also reviewed. The paper is organized in 6 sections. Section 2, provides insight into definitions and technical aspects of ancillary services for ENTSO-E control areas. The overview of activation and market practices for procurement of ancillary services followed in different ENTSO-E member states has been addressed in Section 3. Section 4, is dedicated to an overview of HVDC system types, connection topologies, control structures, and time constants associated with HVDC systems. A review of various literature work about use of HVDC systems for participation in ancillary services has been undertaken and a comprehensive summary is shown in Section 5. Finally, the conclusion of this survey paper is presented in Section 6.

#### **2. Ancillary Services Overview**

The functions or services needed by a TSO to guarantee power system security (reliable and secure power system operation) are termed as *"Ancillary Services"* [10]. These services are either provided by TSO itself or procured from other stakeholders, for carrying out the power transmission from generating units to the load centers while meeting power quality standards [12,13]. The authors in [14] have mentioned that as per the definition, the number and types of the services is very broad. The ancillary services are used to provide the stakeholders with the following capabilities:


The details of the ancillary services shown in Figure 1 are presented in the following subsections.

#### *2.1. Loss Compensation*

The TSO must compensate for all the losses incurred in the process of power transmission from generation units to load centers. These losses correspond to transmission line losses and losses in various other equipments. The TSO must procure energy to make up for these losses. If the generation plant for this energy is not located in the TSO control area, the TSO must take into account the losses for the power transmission in other zones also [14].

**Figure 1.** Ancillary services classification.

#### *2.2. Frequency Control*

In conventional AC power system, the system frequency is a universal characteristic for the synchronous system i.e., it remains same at every measurement point in the system. For reliable and secure power system operation, it is desired that the frequency of the system shall remain constant at nominal system frequency value (50 Hz for ENTSO-E area). Any deviation in frequency can be attributed to a mismatch in power generation and power consumption (load). A set of parameters have been defined for the assessment of reliability and quality of frequency for ENSTO-E area by European Union commission regulations vide guideline on electricity transmission system operation [15]. These parameters are defined as follows:


The frequency ranges (recovery, standard, steady state, and frequency deviation) vary from system to system depending upon the size of the system, typical generation mix, and the time required for activation of reserves. For a smaller islandic system such as GB or IE/NI, these frequency ranges are larger as compared to the larger CE power system. This is due to the fact that deviation in frequency has direct relation with deviation in active power and same power imbalance will result in large frequency deviation for the smaller systems as compared to the same for larger CE system [16] i.e., (Δ*P*/ ∑ *Plarge*) < (Δ*P*/ ∑ *Psmall*). The range for these parameters as defined in the grid code for CE, GB, IE/NI and Nordic power system [15] is shown in Table 1.


**Table 1.** Frequency quality parameters [15].

*Frequency control* is a set of control actions aimed at maintaining the system frequency at its nominal value. Frequency control is implemented in different stages, the commonly defined services for frequency control in ENTSO-E area are categorized as follows:


system imbalances, including generation reserves [15]. RR are activated manually as a result of system optimization by the system operator [21].

The sequence of activation of above-mentioned frequency control services followed by Belgian TSO (Elia) after an imbalance is shown in Figure 2 [22]. Inertia support acts immediately and FCR reacts within a few seconds (full activation within 30 s to any discrepancy between power generation and load with the objective of restricting the frequency deviation. FRR are activated starting from 30 s to bring the system frequency back to its nominal value after the imbalance. RR are activated within 15 min to make FRR available for any other system imbalance. New re-dispatch set points are updated by Elia for economical system operation within 1 h. The sequence of activation for reserves is same for other ENTSO-E control areas also; however, the implementation varies (activation time, threshold value, participating entities etc.).

**Figure 2.** Frequency control ancillary services activation time [23,24].

#### *2.3. Black Start Capability*

The ability of a power system to perform black start operation is known as '*Black Start Capability*' [25]. Black start operation is the process of reviving a power system or a part of power system back to the operational mode from a partial or full shutdown (independent of another power system). Blackouts (situation of total or partial power loss in power system due to unexpected transmission system or generation failure) are the least desired scenarios for power systems and result in social and economic loss [26]. Restoration of power system after a blackout comprises a set of coordinated actions of many power system components and is very complex given the numerous generators, loads and transmission system constraints [27]. In present power systems, it is necessary to recognize the generating units capable of starting without external support and provide power locally. As a consequence of electricity market de-regularization, black start service is treated as a separate ancillary service and is procured by the TSOs from the energy market [28]. As per the regulations, a TSO must identify the generators with black start capabilities in its control area and use these capabilities in a manner to minimize the system restoration time.

#### *2.4. Voltage or Reactive Power Control*

'*Voltage or reactive power control*' is a set of measures or control actions intended to maintain a constant voltage level or reactive power value at each node of the system [15]. These control actions are carried out at different nodes (generation nodes or transformers or AC transmission line ends or HVDC systems or other means) of the power systems. Contrary to frequency, which is a system wide variable, voltage is a local quantity varying for every node of the system. The voltage varies depending upon the system topology, generator, or load location and type of loads. Frequency in the power system is affected by active power balance, voltage is affected in the similar manner by the reactive power balance. Voltage control is implemented by controlling the injection of reactive power in the power system and for this purpose automatic voltage regulators, static VAR compensators, capacitor banks, and reactors are deployed. As it is difficult to transmit reactive power, it is important to control the voltage locally [29]. In view of this limitation, it is very crucial that voltage control equipment is located at critical locations.

Depending on the connection point voltage, the operational voltage limits for steady-state power system operation have been defined for ENTSO-E control area by the European Union commission regulation on electricity transmission system operation [15]. These limits are given in Table 2.


**Table 2.** Steady-state operational voltage range [15].

Ensuring adequate volume and time response of remedial actions to keep voltage within the limits in its control area is one of the tasks of TSO [15]. Thus, a TSO must ensure that sufficient reactive power regulating capacity is available, and this capacity can be activated when needed. The regulating actions to control voltage level can be tap change of power transformer or switching of capacitors/reactors or control of HVDC systems or change in reactive power output of generators etc. The voltage or reactive control service can be split into two hierarchical levels i.e., local and centralized control [29].


In some countries for example France, voltage control is implemented in three hierarchical levels i.e., primary, secondary, and tertiary control. Primary control is activated locally and is activated automatically. Secondary control is an automatic control and controls the voltage at main transmission buses. Tertiary control is activated manually at utility level after power flow analysis to free reactive power reserves.

#### *2.5. Oscillation Damping*

In power system operation, it is desired that the frequency and voltage values shall remain within the stable operation range during or after internal (excitation loss, generator instability etc.) or external disturbances (transmission line fault, loss of generation or load etc.) [32]. As a consequence of these disturbances, low frequency oscillations occur in the power system. These oscillations can be local (to a single plant or generator or a region) or inter-area (geographically spread and involving several remote generators) [33]. Local oscillations (0.7–2 Hz [34]) occur due to presence of fast exciters in the power system whereas inter-area oscillations (0.1–0.7 Hz [34]) are a result of over loading of weak transmission links [35]. If not damped properly, these oscillations may cause partial or total power system blackouts. Automatic voltage regulators equipped with a power system stabilizer (PSS) [36] and flexible AC transmission system (FACTS) devices [33] such as static VAR compensator (SVC) and static synchronous compensator (STATCOM) are employed in the power system for damping these oscillations.

#### *2.6. Congestion Management*

Congestion in power system is a situation in which the transmission system is not able to fulfill all the desired transactions due to power system's physical and operational limitations [37]. These physical and operational limitations can be thermal limits of transmission lines and transformer, voltage limitations, and transient or other stability limits [38].

In grid codes for capacity allocation and congestion management (CACM) [39], 3-types of congestion i.e., market, physical, and structural congestion has been defined. A situation when cross-zonal capacity or allocation constraints limits the economic surplus for single day-ahead or intraday coupling is termed as '*Market congestion*'. When the thermal limits of grid elements and voltage or angle stability limits of power system are breached during forecasted or realized power flows, it is defined as '*Physical congestion*'. '*Structural congestion*' has been defined as transmission system congestion that is predictable, geographically stable over time, and occurs frequently under normal power system conditions. In electricity markets power system congestion leads to price split between various regions. One such case was observed on 3rd October 2018 when the price difference for day-ahead wholesale price between Germany and Belgium was e 105–152 per MWh. This price difference was due to physical congestion between Belgium and Germany [40].

*Congestion management* is the process of making use of available power system infrastructure (economical and operational) while operating within system constraints [41]. Congestion management gives long-term investment signals to the TSO for strengthening local (to a single TSO) or cross-zonal (shared with other TSOs) transmission system infrastructure. A TSO responsible for a given control area or multiple TSOs responsible for the concerned control area must compensate the cost for remedial actions for congestion management [15]. A number of methods have been proposed for congestion management in [38–43], these can be broadly categorized into two methods i.e., technical and non-technical methods. Technical methods of congestion management can be cost free and not cost free [44]. Use of FACTS devices, phase-shifters, and transformer tap change for congestion management comes under cost free congestion management methods. These methods are readily available with the TSO, have limited economic impact and do not involve other stakeholders such as generation or distribution companies. Load shedding and rescheduling of generating units for the purpose of congestion management comes under not cost-free methods. Technical methods are ordered by the TSO. Non-technical congestion management methods can be market-based (auctioning, counter trading, nodal, or zonal pricing etc.) and non-market-based (pro rata or first come first serve). There is no involvement of TSO in non-technical congestion management methods and these are just observed by the TSO. Classification of various congestion management methods has been illustrated in Figure 3.

7

**Figure 3.** Congestion management classification [45].

#### **3. Ancillary Services in De-Regularized Electricity Market Context**

Although ancillary services have been defined by ENTSO-E, the methods for procurement and implementation of these ancillary services vary across different control areas of ENTSO-E. A survey carried out by ENTSO-E on '*ancillary services procurement, balancing market design 2018*' [11] shows that different EU member states have implemented the centrally defined ancillary services in very different manners. For instance, different balancing and ancillary services markets in EU member states have different market closing time for procurement of the ancillary services, different set of participants, different activation time (deadband before activation or instantaneous activation) and different procedure for recovering the cost of ancillary services. As an illustrative example, few features related to FRR (energy) ancillary services for balancing and ancillary services markets in Germany, Belgium, France and Norway are tabulated in Table 3. From this table it is clear that the ancillary services are different products in each of these countries, even when defined in a single grid code.

It has been set in [15] that each TSO must procure 30% FCR from within its load flow block. The volume for FRR and RR to be procured from within load flow block is 50%. A platform for primary and secondary control reserves has been established in Germany for sharing the reserves among the TSOs from Germany, France, Belgium, the Netherlands, Austria and Switzerland [46]. Such flexibility in procurement of services from other control areas gives the TSOs an economical opportunity and encourages the optimal use of inter-zonal transmission capacity. This flexibility of balancing and ancillary services markets has motivated the TSOs to enhance the cross-zonal or inter-country transmission capacities. While procuring the services from different control areas, care needs to be taken about various features of the offered service. From Table 3, it can be observed that the product resolution in time is 1 h for the FRR in balancing and ancillary services market in Germany whereas the same is 30 and 15 min for markets in France and Belgium, respectively. Therefore, to buy the same FRR from Belgian and French markets, a TSO must procure 4 and 2 products respectively as compared to 1 in balancing and ancillary services market in Germany.



#### **4. HVDC Transmission Systems, Control and Dynamics**

HVDC transmission systems are seen as a key enabler for bulk power transmission over long distances with better controllablity as compared to conventional AC systems. HVDC transmission systems can be asynchronous interconnection, embedded transmission line in synchronous system and offshore to onshore grid interconnections. Asynchronous HVDC interconnection is the connection between AC systems operating at different frequencies or systems operating at same frequency but different phase angles [47]. An example of such interconnection is NORNED HVDC link between Norway and the Netherlands. HVDC synchronous interconnection or embedded HVDC is the connection between two nodes in a synchronous area such as ALEGrO link between Belgium and Germany. For evacuation of bulk power from remotely located offshore windfarms, HVDC interconnections are becoming a preferred option. HVDC transmission link BorWin1 in Germany is an example of such system. The representative line diagrams of HVDC transmission systems for asynchronous, synchronous and offshore HVDC connection are shown in Figure 4a–c, respectively.

(**c**)

**Figure 4.** Line diagram for (**a**) asynchronous HVDC transmission system interconnection (**b**) synchronous HVDC transmission system (embedded line) and (**c**) offshore HVDC transmission system.

For HVDC converter stations two types of HVDC technologies are used namely Line commutated converter (LCC) which uses thyristors in current source converters(CSC) topology and voltage source converters (VSC) that uses gate turn-off thyristors (GTOs) or insulated gate bipolar transistors (IGBTs) [48] as shown in Figure 5a,b respectively. Both technologies have some advantages and disadvantages. LCC technology is a very mature technology, cheaper, has less converter station losses, more short-term overload capability and has higher converter ratings in comparison to VSC-based HVDC systems [12]. However, this technology requires large AC and DC harmonic filters, significant dependency of active and reactive power, no blackstart capability and requirement of strong connecting AC systems. It is complicated to use LCC-based technology when power reversal is frequently required because the voltage polarity needs to be changed to change direction of power flow [14]. LCC-based HVDC systems are normally used for back-to-back or point-to-point interconnection of asynchronous

systems [49]. VSC-based HVDC technology is relatively new and can be used to control both active and reactive power separately. This technology can provide AC voltage to blacked-out grids and provide reactive power support similar to statcom [14]. VSC converters are smaller due to requirement of smaller filters [12] and more dynamic in particular with respect to power reversal. VSC-based technology is considered to be a better choice for multi-terminal HVDC grids.

**Figure 5.** (**a**) CSC and (**b**) VSC-based HVDC systems [48].

Interactions between HVDC systems and AC systems have increased due to the increasing number of HVDC systems. A control structure as per Figure 6 has been proposed in [50] to highlight the similarities for primary, secondary, and tertiary control for HVDC and AC system using well known power system interaction between lower controllers and power dispatch.


**Figure 6.** Combined AC and DC control scheme [50].

It is expected that HVDC systems shall also participate in ensuring secure power system control and operation. A converter control for VSC-based HVDC systems as illustrated in Figure 7 [51,52] is generally used for controlling the firing of the IGBTs to control the HVDC grid and AC system voltage. The active power balance in the HVDC system is reflected by the DC voltage in a manner similar to frequency in the AC system [12]. Active power-DC voltage control in the DC system is therefore

similar to active power-frequency control in AC systems. Furthermore, it is possible to implement active power and reactive power droop control from the converter stations by controlling the values *kp* and *ka* as shown in Figure 8a,b, respectively. However, as mentioned earlier for LCC-based HVDC systems it is not possible to control reactive power independent of active power and extra reactive power compensation is required to change reactive power independently [14]. *kp* and *ka* in Figure 8a,b are the active power-dc voltage and reactive power-ac voltage droop coefficients respectively.

**Figure 7.** Control scheme for VSC converter station.

(**b**)

**Figure 8.** Reference (**a**) DC voltage based on active power-DC voltage droop and (**b**) AC voltage based on reactive power-AC voltage droop.

It is also possible to incorporate AC system frequency-active power droop to make the converter react to the frequency deviations in the AC system. The control diagram is presented in Figure 9. *k*<sup>1</sup> in Figure 9 is the frequency-active power droop coefficients.

**Figure 9.** AC system frequency-based active power-DC voltage droop control [53].

In contrast to AC systems which have large inertia, the DC systems have negligible inertia (only small amount of energy is stored in cables and capacitors). A consequence of this small inertia is that the DC systems respond faster to system imbalances than AC systems [54]. An overview of time constants associated with various services for AC and DC systems is presented in Figure 10.

**Figure 10.** Time constants for activation of services for AC & DC system elements [54].

As stated earlier, DC voltage in HVDC systems plays the same role as frequency in the AC systems and can be considered to be an indicator of stable grid operation. Hence, active power imbalances can be controlled by controlling the DC voltage. It can be observed from Figure 10 that the activation time of DC equipments is two orders of magnitude lower than AC system equipments. Primary DC voltage control is activated much faster than FCR and it can play a similar role of inertia support (if frequency-based active power droop is implemented). Using HVDC for frequency control also results in improvement of frequency nadir [55]. It has been concluded in [56] that combining the frequency control reserves (among CE, Nordic and GB systems) using HVDC systems improves ROCOF, frequency nadir, and frequency quality. It can be deduced from these studies that HVDC systems can be an alternative to provide frequency support in the manner which is equivalent to inertia support. Various scenarios for providing frequency support (inertia, FCR, FRR and RR) by providing required active power from HVDC interconnections are shown in Figure 11. It is pertinent to mention here that inertia support and FCR can only be provided by asynchronous or offshore windfarm HVDC interconnections and cannot be provided by the HVDC connections embedded in synchronous zone (as HVDC systems do not store/generate power themselves and the additional power comes from other generation sources). It is possible to provide FRR and RR using synchronous HVDC systems by making remote generation units contribute by changing their set points. It is not possible to provide blackstart using HVDC systems independent of other AC system as some power source is required.

(**c**) **Figure 11.** *Cont.*

(**d**)

**Figure 11.** Change in active power flow through HVDC transmission line for providing (**a**) inertia support, (**b**) frequency containment reserves, (**c**) frequency restoration reserves and (**d**) replacement reserves.

#### **5. Ancillary Services and HVDC Systems**

As mentioned in Section 3, liberalized electricity markets and integration of remote renewable energy sources have highlighted the need for enhanced cross-zonal transmission capacity among various European countries. HVDC transmission systems seems to be the most viable solution for cross-zonal interconnection as these can carry more power and have better controllability. Also, for integrating offshore renewables generation HVDC systems are being preferred. It will not be an over statement that in near future the DC elements will play an important role in power system management. Guidelines have also been published for participation of HVDC systems in power system operation support in ENTSO-E area [57]. For instance, the guidelines allows the TSO to require the HVDC system to control the active (and reactive) power output to maintain stable AC system frequency, provide synthetic inertia in event of frequency deviation in the connected AC system and remain connected and in operation if the network frequency changes at a rate from −2.5 to +2.5 Hz/s .

HVDC systems can actively participate in providing ancillary services to AC systems [14] as detailed in Section 4 also. At the same time HVDC systems will also need ancillary services such as energy balance, loss compensation, black start and restoration for smooth operation [12]. In [58], the capability of windfarms connected through HVDC to provide ancillary services to DC systems has been presented. The aspects of ancillary services required by HVDC systems are not further covered in this paper.

Based on the requirements of ancillary services for AC system, research has been going on to develop the possible solution for providing the ancillary services from HVDC systems to AC systems and in the literature several possible solutions have been proposed. The literature review of possible methods for ancillary services provisions from LCC-based and VSC-based HVDC systems is presented below:

i **LCC-based HVDC system**: The authors in [59–62] have proposed some approaches to control the system voltage from LCC-based HVDC systems. Voltage stability analysis for multi-feed HVDC system using STATCOM has been presented by authors in [63]. Various methods for providing frequency control service from such HVDC systems are presented in [64,65]. In [66], the authors have proposed virtual synchronous generators (VSG) approach for providing frequency control from islanded windfarms. Methods for providing blackstart service with LCC-based HVDC systems has been detailed in [67,68]. Methods for power oscillation damping using LCC-based HVDC systems for connected AC systems are detailed in [62,69,70].

ii **VSC-based HVDC system**: In [71,72], the authors have proposed VSG approach for providing fast frequency control and virtual inertia from the VSC-based HVDC converter stations. Fast frequency and AC system voltage control has been proposed by the authors in [73]. For providing primary frequency support (FCR) from offshore windfarms, HVDC converter control techniques have been proposed in [74–78]. In [59,79–81], the authors have discussed the provisions for AC system voltage control service from HVDC systems. The method for providing frequency from the energy stored in HVDC link is highlighted in [82]. The authors in [83–85] have presented the various studies on blackstart capabilities of VSC-based HVDC systems. In [86–90], the authors have highlighted the control aspects for oscillation damping for VSC-based HVDC systems.

Based on the literature review, some of distinctive features of HVDC systems and their possible use for providing ancillary services have been summarized in Table 4. It is assumed that sufficient reserves are available in the systems to provide the considered ancillary services.


**Table 4.** Ancillary Services from HVDC Transmission systems.

Note. The symbol *-, +, ++ and +++* means that the HVDC systems *cannot provide the service, are able to provide the service, are able to provide the service similar to conventional AC systems* and *can provide the service better than AC systems* respectively. \* implies that the HVDC system requires appropriate controls at the offshore side to provide this service. NA implies that it is not possible to provide this service from respective HVDC system.

#### **6. Conclusions**

Ancillary services play a pivotal role in ensuring reliable power system operation. It is critical that various power system equipments which can provide these services could be used in an economical way while enabling smooth power system operation. This review paper details the definitions of ancillary services, procurement, and activation practices for these services as followed in different control areas of ENTSO-E. It can be concluded from this study that every control area in ENTSO-E follows its own practice for procurement and activation of ancillary services, despite these services being defined by one grid code.

A major outcome of this review paper is the significant potential of HVDC systems (specially VSC-based) in providing ancillary services. A comprehensive analysis of possible control methods and time frame of activation of HVDC equipments highlights the robustness and fast control aspects of HVDC systems. These characteristics are the major drivers for considering ancillary service support from HVDC systems especially within the context of an interconnected grid and offshore grids. From the comparative analysis for different HVDC systems it can be inferred that VSC-based HVDC systems can provide ancillary services in a manner better than or similar to that of the conventional AC systems. This analysis expands on the features of HVDC connections types in providing different ancillary services based on a literature review.

To facilitate the sharing of ancillary services from HVDC among different operators of an interconnected system, a coordinated evaluation of the most optimal use of ancillary services on pan-European level would be necessary. The categorization of different ancillary services from HVDC systems and their characteristics as presented in this paper, could be a starting point for such analysis.

**Funding:** This work is supported by the project HVDC Inertia Provision (HVDC Pro), financed by the ENERGIX program of the Research Council of Norway (project number 268053/E20) and the industry partners; Statnett, Equinor, RTE, and ELIA. https://www.sintef.no/en/projects/hvdc-inertia-provision/.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Ancillary Services in Hybrid AC/DC Low Voltage Distribution Networks**

**José M. Maza-Ortega 1,\*, Juan M. Mauricio 1, Manuel Barragán-Villarejo 1, Charis Demoulias <sup>2</sup> and Antonio Gómez-Expósito <sup>1</sup>**


Received: 31 August 2019; Accepted: 18 September 2019; Published: 20 September 2019

**Abstract:** In the last decade, distribution systems are experiencing a drastic transformation with the advent of new technologies. In fact, distribution networks are no longer passive systems, considering the current integration rates of new agents such as distributed generation, electrical vehicles and energy storage, which are greatly influencing the way these systems are operated. In addition, the intrinsic DC nature of these components, interfaced to the AC system through power electronics converters, is unlocking the possibility for new distribution topologies based on AC/DC networks. This paper analyzes the evolution of AC distribution systems, the advantages of AC/DC hybrid arrangements and the active role that the new distributed agents may play in the upcoming decarbonized paradigm by providing different ancillary services.

**Keywords:** distribution networks; hybrid AC/DC networks; ancillary services

#### **1. Introduction**

Distribution systems have been traditionally designed, built and operated to fulfill the requirements of the large number of customers who are connected to them. Most of these traditional final users are characterized by absorbing power in an almost inelastic manner from the grid, i.e., irrespective of the energy price, to supply mainly electromechanical loads. The usual network design approach has considered these inherent characteristics of customers to achieve a reasonable quality of supply with the minimal investment cost and always considering the network expansion to face the gradual demand increase [1]. Nowadays, however, this traditional customer is steadily evolving to take advantage of the technological evolution which is paving the way to the smart grid paradigm. On the one hand, customers now may play an active role with the advent of the information and communication technologies (ICT), which offer the possibility of changing their consumption according to economic signals. On the other hand, the electronic-based loads are shifting the traditional electromechanical ones. This trend is unstoppable considering the expected massive deployment of electromobility [2–4] and distributed generation [5–7]. Without any doubt, this is a priority of our society which is focused on decarbonizing the transportation sector and increasing the participation of renewable energies in pursuit of the energy self-sufficiency. The most conservative forecast of the International Energy Agency (IEA) envisions an increase of the worldwide electric vehicle (EV) fleet from the current 2 millions up to 56 millions in 2030 [8]. Similarly, the International Renewable Energy Agency (IRENA) predicts a worldwide increase of the photovoltaic (PV) technology in the generation mix used to cover the demand from the current 233 TWh up to 1.104 TWh in 2030 [9]. In this new context, it will be of utmost importance to incorporate into the system new devices and procedures to provide the required flexibility for a safe and secure network operation. Finally, a huge increase of

battery energy storage systems [10–12] (BESSs) is expected at the distribution level from three major drivers: EV deployment, utility-scale applications to tackle the network congestions, and domestic storage units used to modulate the final user demand. It is important to note that IRENA predicts an increase of BESS installed power from the current 1 GW to 250 GW in 2030 [13].

As a result, the distribution business has to evolve to cope with the new constraints imposed by these new agents distributed throughout the network, with stringent regulatory issues, scarce investment capability and final users more and more concerned with the power quality of supply. However, it is possible to take advantage of the flexibility that these new agents may provide to the distribution system, turning the problem into a solution. In fact, PVs, EVs, BESSs (collectively referred to as "PEB" in the sequel) and other power electronic based devices may bring several benefits to the distribution system operation, if properly managed so that they provide a number of ancillary services to the distribution system in addition to their corresponding primary functionality.

Moreover, note that in this new context a huge number of electrical devices connected to the low voltage (LV) distribution networks (PEBs, variable-speed drives associated to different appliances such as air conditioning systems, etc.) could be connected to a DC supply. Therefore, it is arguable whether the current AC networks are still the most efficient way of distributing electricity, owing to the need to include AC/DC converters for these loads. These front-end components, based on power electronic elements, increase the cost of these loads, reduce their efficiency and pollute the distribution system with harmonic currents affecting the system power quality. The use of hybrid AC/DC networks can be considered a logical step forward from the current AC distribution system, allowing loads of DC nature to be directly connected to the DC side of the network.

This paper elaborates on the role that the new agents (PEBs) and other power-electronic devices (DFACTS) may play in the current AC and future hybrid AC/DC LV distribution systems to provide flexibility by means of ancillary services. The paper is organized as follows: First, the main characteristics of traditional AC LV distribution systems are analyzed, followed by an update of the challenges that they are facing with the incorporation of the new distributed devices. Second, a comprehensive review of the most promising hybrid AC/DC LV network topologies is included. Third, different ancillary services for both conventional AC and future hybrid AC/DC networks are detailed. Finally, the paper closes with the main conclusions and future research work on this topic.

#### **2. AC LV Distribution Networks**

The aim of the distribution system is to supply the loads of a large number of customers considering safety, service continuity, power quality, flexibility and expandability issues at the minimum cost [14]. The design of traditional distribution systems comprising MV and LV networks can be broadly grouped into the so-called American or European layouts. The American distribution scheme is characterized by an unbalanced MV network comprising three-phase main feeders with several single-phase or two-phase laterals which directly supply a group of LV loads through pole-mounted MV/LV transformers of reduced rated power. Therefore, the LV network of the American layout is not large, as the main network extension is done in MV in order to reduce the system power losses as much as possible. On the contrary, the European layout is based on a balanced three-phase distribution using three-wire and four-wire configurations for MV and LV networks, respectively. In this case, the MV/LV transformers are always three-phase units and the phase-to-neutral connection of single-phase LV customers is balanced as much as possible among the three phases. Therefore, the LV network extension for the European layout is typically much larger than in the American case. For this reason, the following subsections focus only on the European LV distribution systems.

#### *2.1. One-Way LV Distribution Feeders*

Figure 1a shows the one-line diagram of a typical European LV distribution system, which is fed from a secondary substation equipped with a MV/LV three-phase power transformer. Usually, the connection group of the transformer is Dyn or Yzn to distribute the neutral wire for connecting the single-phase LV customers. The transformer secondary winding is connected to a LV switchboard where the protection of the different radial feeders is embedded. This protection customarily consists of simple fuses, the most cost-effective solution considering the short-circuit power levels of LV distribution systems, the large number of existing LV feeders and their radial nature. The number of LV radial feeders mainly depends on the rated power of the MV/LV transformer, but it typically ranges 4–10. Unlike in the MV case, the LV feeders are structurally radial in most cases, i.e., there is no way to modify the normal topology by acting on normally open switches, which would be helpful for instance to reduce the restoration time after a fault [1]. It is common, however, to find specific locations where the end nodes of those radial feeders are really close to each other, especially in urban areas.

**Figure 1.** One-way LV distribution feeders: (**a**) one-line diagram; and (**b**) voltage profile along the feeder.

As a consequence of the radial design and the passive character of the final users, the power always flows in the same direction, from the secondary substation to the customers, and the voltage magnitude decreases monotonically, as shown in Figure 1b. In any case, the receiving voltage must be within the limits imposed by the standards [15]. For this purpose, the MV/LV transformers are equipped with off-load tap changers, allowing the utility to adjust the LV voltage at the head of the feeder to fulfill the voltage requirements. However, this device can be exclusively operated by disconnecting the transformer and interrupting the service to the final user. For this reason, the tap position is seldom readjusted once the transformer is commissioned. Those secondary substations close to the primary one feature higher MV voltages and, therefore, upper tap positions are usually selected to reduce the LV voltage. On the contrary, lower tap positions are used in those secondary substations far from the primary ones.

#### *2.2. Two-Way Distribution Feeders*

Figure 2 is the counterpart of Figure 1, updated in accordance with the smart grid paradigm arising by the incorporation of PEBs, which are transforming consumers into prosumers. These can play now a more active role by controlling their consumption/production profiles, considerably complicating the operation of the LV radial distribution system. First, the simultaneity coefficients of rooftop PV production and EV charging can be much higher than those of traditional commercial, residential or industrial loads. Second, the deployment of these new technologies within the distribution network is not homogeneous, being it possible to find different traditional customer versus prosumer ratios depending on several external factors. Third, the power flows are no longer one way, from the secondary substations to the final users. This completely modifies the voltage profiles along the feeders, as shown in Figure 2b, being it more difficult to guarantee that the final user voltage lies within the regulatory limits. Moreover, it is worth noting that this load/generation scenario is dramatically changing throughout the day, as the PV generation and the EV demand should be peaking during noon and midnight hours, respectively. Therefore, a massive penetration of PEBs may create distribution network congestions, i.e., currents above the ampacity ratings or voltages beyond regulatory limits, which cannot be alleviated by merely acting on the off-load tap changer of the MV/LV transformer.

**Figure 2.** Two-way LV distribution feeders: (**a**) one-line diagram; and (**b**) voltage profile along the feeder.

Classical network reinforcement, i.e., increasing the cross-section area of cables or installing new LV feeders, may solve the voltage regulation problem. However, in cases where the feeder congestions arise a few days a year, it is questionable whether the investments in new assets are really justified. Moreover, those network reinforcements are not always straightforward from a practical point of view in densely populated urban areas where underground installations are used. The main issue, however, is that network reinforcement does not tackle the root of the problem, namely the radial topology of the LV distribution network, which constitutes the main technical barrier to the deployment of the new decarbonizing technologies.

A promising alternative solution to this problem comes from the application of new smart-grid technologies, particularly new power electronics-based devices and new control systems, as proposed in Figure 3 and discussed in the following subsections.

**Figure 3.** Advanced two-way LV distribution feeders: (**a**) one-line diagram; and (**b**) voltage profile along the feeder.

#### 2.2.1. Utility-Scale Power Electronic Devices

Four different utility-scale devices are incorporated in Figure 3 to host as much PEBs as possible by providing additional flexibility:


peak shaving [27] and energy shifting [28]. In addition, it is possible to contribute to the voltage regulation by means of active and/or reactive power injections.

• DC link. The aim of this device, also known as flexible link, is to create a controllable loop between the radial feeders to which it is connected. It is composed of two VSCs connected in back-to-back configuration sharing a common DC bus, as shown in Figure 4d [29]. This device may control the active power flow between the interconnected feeders and two independent reactive power injections. This provides an extraordinary flexibility to the distribution system operation [30,31]. In fact, this device can fully overcome the barrier related to the radial nature of the LV distribution system, because network congestions in one feeder can be alleviated using the neighbor feeder as a back-up supply point. The use of this DC link is advantageous when compared to a conventional meshed operation, where the active and reactive loop flows cannot be controlled. In addition, the meshed operation requires an expensive protection system to cope with short-circuit faults [32]. On the one hand, the short-circuit current increases as both interconnected feeders contribute to the fault current. On the other hand, note that the protections of both interconnected feeders should trip in the case of a short-circuit fault and, therefore, disconnecting a larger number of customers in comparison to the radial operation case. The use of DC links, based on back-to-back VSCSs, prevents this undesired effect, as the healthy feeder can be quickly isolated from the faulted one by just inhibiting the gating signals to the IGBTs. This way, it is possible to maintain the conventional, simple and reliable protection system used in LV radial distribution networks. It is worth noting that it is possible to interconnect *N* feeders by means of multi-terminal arrangements. This multi-terminal device, composed of *N* converters, has 2*N* − 1 degrees of freedom, *N* − 1 of them corresponding to active power flows (as one of the VSCs must control the DC bus voltage) and *N* to reactive power flows [33]. Finally, it is interesting to highlight that the DC bus of this device may incorporate a PV generator, a battery or even an EV fast charger station, providing even more flexibility of operation [31]. Alternative topologies have also been proposed to reduce as much as possible the rating of the power electronic components, aimed at reducing its cost [34].

**Figure 4.** Power electronic devices incorporated in the smart grid distribution system: (**a**) solid state tap changer (SSTC); (**b**) three-phase, three-wire STATCOM; (**c**) three-phase, three-wire BESS; and (**d**) DC link composed of two back-to-back VSCs.

#### 2.2.2. Control Systems

The new paradigm arising around the active distribution systems discussed above and depicted in Figure 3, involves the development of a Secondary Substation Control System (SSCS) [35]. After decades of incremental technological advances, this constitutes a leapfrog with respect to

the current status of LV AC distribution networks. Nowadays, those networks are unobservable in real-time, being equipped in the best cases with just a data concentrator to gather the data from the smart meters downstream [36]. This Advanced Metering Infrastructure (AMI), however, is thus far exclusively used for billing purposes, its extension for improving the network operation still being on the blackboard. The aim of the control system included in Figure 3 is twofold:


The adoption of this hierarchical control scheme, as shown in Figure 5, is advisable in case of distribution systems where the number of power assets and prosumers is huge [38,39]. The traditional centralized approach, where a unique control center is in charge of receiving and processing all the raw field measurements, as well as computing the required control actions for the set of control assets [40], is not scalable if the whole LV network is incorporated, as this may entail thousands of components for a single primary substation. Undoubtedly, ICTs will play a crucial role in this new context, as they are required for exchanging the information between the different agents [14]. The ICT requirements are quite demanding if real-time network operation is envisaged, in addition to back-office functionalities such as billing of customers. Therefore, special attention must be paid to ICT resiliency, time latencies and cybersecurity issues.

**Figure 5.** Hierarchical control of active distribution systems.

#### **3. Hybrid AC/DC Networks**

Hybrid networks, composed of AC and DC parts, are not new in power systems. Their existence dates back to the fifties of the last century in transmission applications when the first HVDC was built to supply the Götland Island [41]. The use of HVDC has been mostly restricted to applications characterized by the need of transmitting large power through large distances, where AC lines are constrained by stability issues [42]. The technological evolution in terms of rated power and DC voltage has since been impressive. Some installations of modern HVDC based on VSC technology reach 1.8 GW with a rated voltage of ±500 kV [43,44] while some traditional HVDC units based on thyristors go up to 8 GW and ±800 kV [45,46]. HVDC has also been used to evacuate the active power of large off-shore wind power plants to the transmission system due to the reduced capability of long AC submarine cables [47,48].

In the opposite extreme of the grid spectrum, hybrid AC/DC networks have also been explored in microgrids. The microgrid concept [49], originally proposed in 2002 as a building block of the future distribution system, has been of interest for the scientific community due to the new advances in power electronics and digital processors with extended computation capability [50,51]. During the last years, different microgrid proposals based on either DC or AC technologies have emerged without a clear consensus on the best topology aiming at optimizing DC [52] and AC [53] resources simultaneously. Hybrid AC/DC networks represent a natural evolution of microgrids, merging their advantages. Wang et al. [54] proposed these hybrid networks as an efficient solution for supplying DC loads. This achieves a reduction of intermediate conversion stages with the subsequent cost reduction of electronic devices. In addition, the power quality of the AC network, mainly imbalances and harmonic distortion, can improve, because of the use of advance control algorithms in the centralized power converters used for connecting both networks. Multiple configurations of hybrid AC/DC networks can be found in the specialized literature, which can be classified according to different criteria: network topologies, involved power electronics converters and applied control algorithms. The following subsections are devoted to analyzing each of these topics.

#### *3.1. Network Topologies*

The simplest hybrid AC/DC network is based on the use of the back-to-back VSCs for supplying the DC loads from the DC bus [55,56]. However, it is also possible to find more futuristic approaches where a hierarchically organized power system composed of several hybrid AC/DC subnetworks connect each other by a number of power electronic converters, somehow mimicking the architecture and behavior of the Internet [57,58]. Due to this diversity of solutions, [59] raised an interesting taxonomy of the different existing topologies including a comparison with regard to different technical criteria.

In any case, to assure an adequate deployment of hybrid AC/DC networks, it is required to rely on two basic design principles: (1) to leverage as much as possible the already existing infrastructure of the existing AC network in line with [60]; and (2) to minimize the number and complexity of power devices used for the active management of the network. Considering these underlying ideas, Figure 6 shows a straightforward derivation of a LV hybrid AC/DC distribution network by adding just the DC conductors to the already existing four-wire AC branches. The existing reserve conduits of underground installations can be sometimes used for placing the new conductors. In this topology, the DC network is supplied from a central VSC within the secondary substation but also from other AC/DC links which can be distributed along the network. AC loads are supplied as usual while DC ones are directly plugged into the DC network, eventually using DC/DC converters. It is important to point out that the bridges between the AC and DC networks allow a bidirectional flow of active power. This opens the possibility of optimizing the power flows between the AC and DC sides to obtain the best operation in terms of power losses while maintaining the power quality to the final user.

**Figure 6.** Hybrid AC/DC LV distribution network by adding new DC conductors.

It is worth noting that the hybrid AC/DC grid proposed in Figure 6, which is derived from an original AC grid, increases the network loadability because of the new DC lines. However, this loadability increase is limited by the rated power of the MV/LV transformer within the secondary substation. A possible alternative, which maintains the existing LV line infrastructure, is shown in Figure 7. In this case, two AC phases are transformed into DC lines. The single-phase AC and DC loads are directly fed from the AC and DC networks, respectively, while the three-phase ones require the use of three-phase VSCs supplied from the DC bus (in general, however, three-phase loads are a minority in LV systems). The DC network is supplied from the secondary substation with a three-phase and four-wire central VSC but additional single-phase VSCs can be connected to create bridges between the AC and DC sides.

**Figure 7.** Hybrid AC/DC LV distribution network by replacing two AC conductors by DC ones.

Finally, it is worth noting that those two network topologies are not mutually exclusive, depending on the characteristics of the customers. This way, LV networks with a predominant presence of domestic users, which are characterized by DC and single-phase AC loads, should be adapted to the topology proposed in Figure 7. On the contrary, Figure 6 is preferred for those networks with commercial and industrial users with DC and three-phase AC loads.

#### *3.2. Power Electronic Converters*

Undoubtedly, AC/DC and DC/DC power electronic converters are the cornerstone of the AC/DC hybrid networks. In the case of three-phase AC systems, traditional three-phase converters are usually applied [54,55]. However, for single-phase AC systems, it is possible to find different alternative solutions to the conventional single-phase VSC [61]. For example, the authors of [62,63] presented a single-phase high-power density converter with two stages composed of a DC/DC regulator and a classical AC/DC bridge.

On the other hand, DC/DC converters have been widely used for adapting the voltage levels of renewable power sources and energy storage systems, both in isolated [64,65] and non-isolated arrangements [66].

#### *3.3. Control Algorithms*

Hybrid AC/DC networks are inherently flexible distribution systems, as they are composed of several AC/DC converters bridging AC and DC sides at different nodes. As a result, it is possible to apply different real-time control strategies aimed at maximizing the network performance. For instance, [67] controlled the EV charges in an effective way so that technical constraints violations are eliminated. In [68], the distributed generation is scheduled to satisfy the demand at minimum supply cost. [69] solved a centralized dispatch of generation and storage units in a hybrid AC/DC network considering the variability of the renewable energy sources. [70] analyzed the application of hybrid AC/DC networks within buildings equipped with renewable energy sources, energy storage and controllable loads. As can be noticed, a wide range of control algorithms can be found in the specialized literature. The interested reader is referred to the hierarchical classification and technical characterization proposed in [71].

It is important to highlight that the control architecture usually poses a hierarchical layout, similar to the one shown in Figure 3 and discussed in Section 2.2.2, where the centralized controller acts as a tertiary control level in charge of computing the optimal set-points for the AC/DC and DC/DC converters. In turn, those devices are equipped with primary controllers which are in charge of tracking the references sent by the central secondary controller [72]. In this sense, it is worth stressing that the centralized controller solves an optimization problem to minimize a given objective function, including a set of inequality an equality constraints, of the form:

$$\begin{aligned} \min & \quad & f(\mathbf{x}, \mathbf{u}) \\ \text{s.t.} & \quad & \mathbf{g}(\mathbf{x}, \mathbf{u}) = \mathbf{0} \\ & \quad & \mathbf{h}(\mathbf{x}, \mathbf{u}) \le \mathbf{0} \end{aligned} \tag{1}$$

where vector **x** comprises the dependent or state variables and vector **u** the control variables. The inequality constraints take into account the technical limits of the network (ampacity limits of lines and maximum nodal voltage variations) and the AC/DC and DC/DC converters (maximum converter current and DC voltage range). The equality constraints stand for the network equations comprising both AC and DC parts. This requires specific formulations, such as the one proposed in [73–75] for solving the load flow problem in the presence of VSC-based HVDC networks and/or microgrid systems. Finally, note that these optimization problems should be fed with real-time measurements from the field, if they are to be used for control purposes. For this reason, it is essential to rely on state estimators for computing the maximum likelihood state of the network from the raw field measurements [76]. However, suitable models for the specific case of HVDC systems have recently been developed [77–79].

#### **4. Ancillary Services in AC and Hybrid AC/DC LV Distribution Systems**

The advanced AC and hybrid AC/DC LV distribution grids presented in the previous sections, comprising several power electronics based components with accurate and fast control capabilities, allow an unprecedented and sophisticated active network management. In this upcoming paradigm, the new controllable assets may improve the distribution grid operation by providing different ancillary services, mimicking to a large extent the well-known operation of transmission systems.

The definition of ancillary service is somewhat fuzzy, as it encompasses any service required by the transmission or distribution system operator to maintain the integrity, stability and power quality of the system [80]. In fact, there is no unique or harmonized definition even for countries within the European Union, where different provision schemes are articulated [81]. In this work, the following ancillary services for AC and hybrid AC/DC LV distribution grids can be defined: voltage control, congestion release, imbalance reduction, harmonic distortion mitigation, power smoothing, inertial response and power frequency response. A classification of these ancillary services and the side of the network where they can be applied is detailed in Table 1. The following subsections outline each of these ancillary services, some of them related to the frequency and voltage control, others dealing with power quality issues.


**Table 1.** Classification of ancillary services for AC and hybrid AC/DC distribution networks.

#### *4.1. Voltage Control*

The objective of this ancillary service is to maintain the voltage levels at the different nodes of the distribution system within the technical limits [15]. The presence of distributed generation within the AC distribution system may create voltage rises due to inverse power flows that were not common in distribution systems [82]. On the contrary, the massive EV integration may create deep undervoltages. Active power curtailment has been proposed to mitigate this problem [83,84]. However, the practical difficulties of its implementation, related to a fair curtailed power allocation between the involved generators, and the high cost of the curtailed power, prevent its use. As an alternative, reactive power injections by distributed generators [85–87] or EV charging stations [88] can be used for controlling the voltage. The reactive power injections can be done depending on local voltage measurements [89] or according to control signals computed by a centralized algorithm [90]. The utility-scale power electronic devices previously outlined (BESSs, STATCOMs and DC links) can also be used for providing this ancillary service. Note that BESSs [91] and DC links [30] may also contribute with adequate active power management. In addition, SSTCs play an important role in the voltage control of LV distribution systems [92]. This is even more relevant in those cases where the high R/X ratio leads to a low sensitivity of voltage with respect to the reactive power injections [93]. In those situations, controlling the voltage by means of reactive power injection is not a cost-effective solution, which prevents the use of large utility-scale devices or oversized distributed generators. Undoubtedly, a coordinated control of all these resources is the best option to provide the voltage regulation within the AC distribution system [94].

Regarding the voltage control on the DC side of the hybrid AC/DC distribution systems, it is important to highlight the lack of standards on DC networks, beyond those related to data centers and ICT infrastructures [95,96]. In any case, this ancillary service has to be done by adequate active power management of the DC generators, BESSs and AC/DC converters supplying the DC network from the AC side [97–99]. In this sense, the use of AC/DC bidirectional converters spread along the network is of special interest to regulate the active power flow on the DC and/or AC sides, depending on the network loading.

#### *4.2. Congestion Release*

This ancillary service aims to maintain the power flows through the different branches of the distribution system below the ampacity limits. In traditional passive systems, this problem can only be solved by network reinforcement. However, this is not the best solution considering the cost, connection time and the spare capacity of the new assets over a large number of hours per year [100,101]. The active management of the future LV distribution systems offers multiple solutions for maximizing the loadability of the existing assets below their ratings. Considering the utility-scale devices, in addition to BESS [102], the DC link is probably the most promising solution for congestions release in radial distribution systems [30,31,34]. Moreover, PV generation or EV charging can be connected to its DC bus, facilitating the integration of these new components [33]. As far as distributed generation is concerned, it is possible to resort to generation curtailment [103], but also to the exchange of reactive power aimed at maintaining the voltages within acceptable bounds [104,105]. Similarly, the EV charging can be done in a controlled fashion to minimize its impact on the distribution system [106,107]. However, the most interesting results are obtained by exploiting the synergies between the distributed generation and EV charging [108,109]. In fact, the use of hybrid AC/DC networks maximizes this synergy, as the active power flow through the AC/DC bridges corresponds to the net active power of EV and distributed generation, minimizing their impact on the AC side.

#### *4.3. Unbalance Reduction*

Load unbalance, mainly due to the presence of unevenly distributed single-phase customers but also to domestic single-phase PV installations, is one of the main power quality problems of LV distribution networks [20,110]. The phase unbalances lead to several problems such as neutral and ground currents, neutral point-shifting and higher losses than in balanced operation [111]. For this reason, the voltage unbalance is usually limited in the current standards [15,112]. In this situation, those technologies based on VSCs (STATCOMs, BESSs, PV generation and DC links) may provide the unbalance reduction as ancillary service to the distribution grid. Some of these devices may have the unbalance reduction as their main target (e.g., STATCOM) but it is also possible to rely on alternative devices equipped with advanced control algorithms with other primary functionalities (e.g., BESSs, PV generation or DC links). For this purpose, those devices have to inject different current in each phase trying to balance the current upstream of their point of connection [113–115]. This implies injecting not only positive sequence current but also negative and zero sequence components. Particularly, the VSC plays an import role in this regard [21]. As three-phase, three-wire VSCs may only inject positive and negative sequence components, four-wire topologies are needed capable of injecting zero sequence currents. Different topologies have been proposed for the three-phase, four-wire VSC [22,23], which can be also composed of a set of three single-phase units [24]. Finally, it is also worth noting that the use of SSTCs with individual phase operation capability may also contribute to reduce the system voltage unbalance [116,117].

#### *4.4. Harmonic Mitigation*

Harmonic distortion is a relevant power quality issue in AC distribution systems because of the proliferation of non-linear loads. The use of power electronic devices, thyristors and/or diodes, for rectification purposes creates *m* · *k* ± 1 low order harmonics where *m* represents the number of pulses of the device. The problems created by these harmonics are well documented, including increase of power losses, overload of capacitor banks, electromagnetic interferences, overvoltages, malfunction of protection devices, and overload of neutral conduct and resonances among others [118]. Harmonic distortion is limited by several standards [15,112,119] and traditionally have been reduced by applying passive filtering techniques [120]. However, the development of power electronics, particularly the self-commutated technology [121], has enabled the application of active filtering [122]. Despite the shunt active filter being probably the most common topology, series and shunt-series active filters have also been proposed for reducing the rating of the power electronic components and increasing the filtering performance [123]. However, any device based on VSCs (STATCOMs, BESSs, PV generation and DC links) may incorporate this advanced functionality and provide this ancillary service to the distribution system [124]. Basically, the operation principle is

similar to the one used for the unbalance mitigation: injection of harmonic currents to achieve an almost sinusoidal current upstream of the point of connection. However, note that the local compensation of harmonic currents usually does not lead to a global reduction of the harmonic distortion [125]. For this reason, coordinated control strategies have to be applied for achieving this global goal [126–128]. Regarding the harmonic reduction on the DC side of hybrid AC/DC networks, the harmonics caused by the previously mentioned line commutated converters are of orders *m* · *k*. Despite methods for computing the harmonics of DC networks being available for special applications, such as HVDC installations [129], there is not a clear standard limiting them. However, the harmonic reduction on the DC side of the hybrid AC/DC systems can be provided either by the DC/DC or AC/DC converters connected to them.

#### *4.5. Power Smoothing*

The introduction of renewable generation relying on uncontrollable primary energy sources may produce large variations of the active power injections, which may lead to several problems in the power system. These variations are perceived by the conventional power plants based on synchronous generators as power ramps, which have to be compensated to maintain the power balance in the system [130]. These power ramps are especially problematic in weak AC networks where system operators try to limit their magnitude to assure the system stability [131–133]. Conversely, the power variations also have to be limited on the DC side of hybrid AC/DC networks to control as much as possible the DC voltage fluctuations [99]. This ancillary service can be provided by distributed generators, BESSs and other energy storage systems [134] depending on the rate of power change. Fast power variations, with time scales lower than 60 s, can be compensated by distributed generation equipped with advanced controllers, such as the rotor inertia control [135] and pitch angle control [136] in wind energy systems, without even involving complementary energy storage device. In addition, it is possible to use fast energy storage systems, such as supercapacitors [137], which are able to manage large amounts of active power in short-time periods without any risk of accelerated aging. Conversely, BESSs are most suitable for slow power variations [138] with different control strategies based on moving average [139] or ramp-rate control [140] algorithms.

#### *4.6. Inertial Response*

The introduction of renewable generation is displacing the conventional units based on synchronous machines, which are the main source of inertia of the power system. The inertia of a synchronous generator counteracts the frequency changes by injecting or absorbing an additional active power, instantly taken from the kinetic energy of the rotor [37]. Thus, as the penetration level of renewable generation increases, the system inertia decreases and its dynamic control becomes challenging [141]. A possible solution to overcome this shortcoming is to incorporate BESSs, either in standalone installations or combined with generation assets, to emulate the behavior of the synchronous generators, namely to provide virtual inertia. This ancillary service can be provided following different strategies [142]:


#### *4.7. Power Frequency Response*

The objective of this ancillary service is to contribute to the frequency regulation of the power system when a frequency disturbance, due to an imbalance between generation and load or any frequency stability issue, appears in the power system. Primary frequency response is implemented in the governor control of the synchronous generators by defining a given power versus frequency droop curve. This local implementation allows an immediate reaction against frequency variations without the intervention of any higher supervisory control [141]. This primary frequency response is slower than the inertial one outlined in the previous subsection, with an actuation time scale from a few milliseconds up to 30–40 s. The progressive integration of renewable energy sources replacing traditional generation requires the provision of this ancillary services by BESSs. On the one hand, it has to be considered that BESS, as a power electronics based device, has a faster response in comparison to traditional synchronous generators. On the other hand, it has to be considered the limitations of this technology. First, BESSs are constrained in energy which prevents its use as conventional generation units. Second, BESSs are constrained in power with limited ramping rates. Third, the number of frequency events and their duration are related to the number of cycles and the depth of discharge of the battery which are strongly related to its aging [150,151]. A straightforward way of providing this ancillary service is by implementing a power versus frequency droop curve in the BESS controller in such a way that active power is injected or absorbed when the frequency decreases or decreases, respectively [152,153]. To reduce the cycling of the battery, it is usual to incorporate a dead band around the fundamental frequency [154]. In addition, this ancillary service can be at least partly provided by renewable generation units, which usually operate maximizing the power conversion from the primary energy source [155]. As an alternative, the plant can be operated below the maximum power point, allowing reactions to frequency excursions of both signs [156,157]. Otherwise, the non-dispatchable generation units may act exclusively on over frequency events.

#### **5. Conclusions**

This paper analyzes the evolution of LV distribution systems from the traditional AC passive networks to future hybrid AC/DC systems, and the role that the new power electronics based technologies may play in this new distribution paradigm by providing ancillary services. This change is being driven by the technological substitution of the traditional electromechanical devices by the new power electronics ones and the development of control systems based on advanced communication infrastructure. The new distributed agents, including PV generation, EV chargers and other utility-scaled devices such as SSTCs, STATCOMs, BESS and DC links, allow an active management of the distribution system to be implemented. The paper outlines the basic operational principles of each of those technologies and their use for an improved operation of the distribution grid. Actively operating the LV distribution grid requires the intervention of a control system in charge of supervising and determining the adequate set-points for each control asset. For this purpose, a hierarchical scheme is described where each LV distribution network departing from a secondary MV/LV substation is locally controlled. Once the LV AC distribution system has been transformed from a passive to an active network, by incorporating the power and control elements previously mentioned, it is questionable if the AC topology is the most adequate one, considering the DC nature of many of those power devices. The authors' opinion is that hybrid AC/DC LV networks constitute the logical way of incorporating new DC components to the existing AC LV systems. This is envisioned as a smooth transition, trading off the need to maximize the use of the AC legacy assets while allowing a massive roll-out of the DC technology. In any case, irrespective of whether the LV network is purely AC or hybrid AC/DC, it is of utmost importance to duly consider the contribution that the new controllable assets may bring to the whole system, by the provision of ancillary service. The paper has ended by outlining different ancillary services related to frequency regulation, voltage control and power quality issues, which will be required in the near future at the distribution level. The evolution of the power systems throughout the 20th century declared Tesla as the undisputed winner of the war of the

currents after Edison's fleeting initial success. The expected evolution over the 21st century, however, will very likely reveal that the last battle has not yet finished. The DC system revival for integrating the new technologies described in this paper is technically and economically feasible nowadays for the last mile distribution systems. In the authors' opinion, however, the harmonized coexistence of both systems through the deployment of hybrid AC/DC networks is the most cost-effective solution for integrating the new technologies into the current AC networks. Undoubtedly, hybrid AC/DC grids will provide extended flexibility allowing the new technologies to provide the required ancillary services for an optimal, secure and sustainable network operation.

**Author Contributions:** All the authors have equally contributed to the final version of this paper.

**Funding:** This research was funded by the Spanish Ministry of Economy and Competitiveness with grant number ENE2017-84813-R and the European Union Horizon 2020 Program with grant number 764090.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **FACTS Providing Grid Services: Applications and Testing**

#### **Panos Kotsampopoulos \*, Pavlos Georgilakis, Dimitris T. Lagos, Vasilis Kleftakis and Nikos Hatziargyriou**

School of Electrical and Computer Engineering, National Technical University of Athens, 15780 Zografou, Greece **\*** Correspondence: kotsa@power.ece.ntua.gr; Tel.: +30-2107721499

Received: 17 May 2019; Accepted: 28 June 2019; Published: 3 July 2019

**Abstract:** The role of flexible alternating current transmission systems (FACTSs) in the provision of grid services is becoming increasingly important, due to the massive integration of intermittent renewable energy sources, energy storage systems, and the decommissioning of thermal plants. A comprehensive literature review of grid services offered by FACTS is performed, focusing on the different grid services that they can provide, such as power flow control, reactive power control, voltage control, power quality improvement, harmonic mitigation, improvement of transient stability, and damping of inter-area and intra-area oscillations. These grid services need to be realistically and economically validated in suitable testing environments. A review of relevant standards, guides, and the literature is performed, which covers the entire range from functional specification and factory testing up to the field testing of FACTS. Advanced industry practices, such as controller hardware in the loop (CHIL) testing of FACTS controllers by the manufacturer, and recent trends, such as CHIL testing of replica controllers by the owner, are underlined. Limitations of conventional testing and CHIL testing are explained and the use of power hardware in the loop (PHIL) simulation for FACTS testing is discussed. CHIL and scaled-down PHIL tests on a transmission static synchronous compensator (STATCOM) are performed and a comparison of the results is presented.

**Keywords:** FACTS; grid services; CHIL; PHIL; lab testing; field testing; standards; STATCOM; replica; review

#### **1. Introduction**

Power systems are subject to an unprecedented transformation, characterized by the wide integration of intermittent distributed generation, energy storage, consumer engagement, and the decommissioning of thermal plants, in order to meet environmental goals, while maintaining the quality of supply [1–3]. In this transformation, flexible alternating current transmission systems (FACTSs) can play an important role by facing several challenges of the transmission system [4]. They can offer a variety of grid services, such as power flow control, reactive power control, voltage control, power quality improvement, harmonic mitigation, improvement of transient stability, damping of inter-area and intra-area oscillations, and black-start capability, among others.

Many researchers have investigated various models and methods for the analysis of FACTS devices and their optimal operation and planning within power systems. In addition to research articles, there are also important review papers that summarize the research findings on FACTS devices and their grid services in a comprehensive manner. There are review articles devoted to the review of one particular FACTS device [5–7], and articles that review more FACTS devices [8–10]. The study of [5] reviews the models and methods of unified power flow controllers (UPFCs) in smart grids, sets the future research goals, and provides future research directions in this field. Another [6] provides a comprehensive review on static synchronous compensators (STATCOMs) and their future research

potentials. Furthermore, [7] provides a systematic review of static phase sifters (SPSs), compares SPS configurations, and highlights their advantages and limitations. The work of [8] presents a bibliography review of FACTS applications for enhancing power quality and ensuring efficient utilization of energy in power systems with increased penetration of renewable energy sources. The work of [9] presents a review of methodologies for optimum allocation and coordination of FACTS devices and distributed generation units. Finally, paper [10] reviews electromagnetic transient models of FACTS devices that do not use voltage source converters and summarizes key characteristics of each model.

There is no review paper focused on the different grid services provided by various FACTS devices. This paper aims to cover this gap by providing a comprehensive review of the grid services offered by FACTS. These services include power flow control, reactive power control, voltage control, power quality improvement, harmonic mitigation, improvement of transient stability, and the damping of inter-area and intra-area oscillations. Due to space limitations, only a small set of representative research works is reviewed in Section 2 for each one of the different grid services offered by FACTS.

Moreover, suitable testing procedures and setups for testing grid services by FACTS are necessary. In the last decade, several standards and guides related to FACTS testing have been published. These include factory testing and field testing for specific types of FACTS, such as static var compensator (SVC) and STATCOM; however limited attention has been placed on grid services. A method for efficiently testing the provision of grid services is hardware in the loop simulation [11], where a hardware controller (controller hardware in the loop simulation, CHIL) or hardware power device (power hardware in the loop simulation, PHIL) is connected to a real-time simulated power system. Testing of the control systems of FACTS using CHIL simulation has been performed by manufacturers for several years and more recently by transmission system operators (TSOs).

Although there is a plethora of papers that have performed simulation studies of FACTS, limited experiences on hardware testing have been reported, concerning both laboratory testing and field testing. Most of these papers describe specific test cases for a specific FACTS device at a specific location (as described in Sections 3.1 and 3.2). A review on the testing of FACTS is still missing. This paper addresses this gap by reviewing papers, standards, and guides on the testing of FACTS. It efficiently presents all the different testing stages, including conventional testing, i.e., factory and field testing, and emerging industry practices, such as the CHIL testing of FACTS controllers by manufacturers and CHIL testing of replicas by utilities. Particular attention is placed on the testing of grid services.

The contributions of this review paper are manifold:


The structure of the paper is as follows: Section 2 presents a literature review on the grid services offered by FACTS. Section 3 reviews the conventional testing of FACTS, advanced industry practices, and recent trends. The CHIL and PHIL results of a STATCOM performing voltage control are presented and discussed. Section 4 concludes the paper and summarizes the main findings.

#### **2. FACTS Providing Grid Services: A Review**

FACTS devices are based on power electronics and are used in order to improve the control of electric power transmission systems in both steady state and transient state conditions [12]. FACTS devices also help increase transmission lines' power transfer capacity [12]. The ability of an alternating current (AC) transmission line to transfer AC electric power is constrained by various factors, including the thermal limit, voltage limit, transient stability limit, and short circuit current limit. These limits define the maximum power, called the power transfer capability, which can be transferred through the AC transmission line without causing damage to the transmission line and the electrical equipment.

A FACTS device provides control of one or more parameters of an AC transmission system. These parameters include the voltage magnitude, voltage angle, and the impedance of the transmission line. Through the control of these parameters, the FACTS device can control the real and reactive power flow, the voltage magnitude, and the shunt reactive power compensation. There are four different types of FACTS [12]:


In this section, representative research works are reviewed for each one of the different grid services offered by FACTS.

#### *2.1. Power Flow Control*

The power flow control services include:


different FACTS, namely TCSC, SVC, and thyristor controlled phase angle regulator (TCPAR), shows that a SVC is slightly better in the improvement of the voltage profile, while a TCPAR is better in reducing the total active power loss [24]. In comparison with the case without a UPFC, the optimal allocation of one UPFC provides a greater reduction of transmission lines' total active power loss and buses' voltage deviation for both the IEEE 57-bus and the IEEE 118-bus test systems [25]. In an AC-DC power system with 96 AC buses and two DC terminals, the optimal allocation of one UPFC, under contingency conditions, has a significant impact on the minimization of power loss and voltage deviation [26].


#### *2.2. Voltage Control*

The voltage control capability of the SVC was investigated and a reactive power dispatch model was developed that restores the SVC operating point and regulates the bus voltage [40]. A photovoltaic inverter is controlled as a STATCOM and provides voltage control in power distribution systems [41]. The limits of the UPFC were included into a steady state power flow model, which was validated by simulations that highlight the capabilities of a UPFC for coordinated voltage control and power flow control [42]. An optimization methodology was developed that identifies the optimal parameter settings of one UPFC and manages to relieve voltage violations and overloads that are caused by line outages [43]. A probabilistic methodology improved the steady state bus voltage profile by optimally sizing the TCSC, STATCOM, and UPFC [44].

#### *2.3. Improvement of Power Quality*

FACTS devices, such as SVC, STATCOM, and UPFC, offer significant power quality services to the grid, including enhancement of the power system reliability and mitigation of voltage sags, harmonics, and unbalance [8,45–49]. The capability of the SVC, STATCOM, and dynamic voltage restorer (DVR) to mitigate voltage sags, harmonics, and unbalance was shown in [45]. Distribution STATCOM and SVC minimize voltage sags and the economic losses in power distribution systems [46]. SVC, STATCOM, and DVR minimize economic losses due to voltage sags [47]. An appropriate control strategy allows the UPFC to provide harmonic isolation in case of nonlinear loads [48]. In order to

improve the reliability problem that is due to the loading of a transmission line, a UPFC was installed on that line and, as a result, power system reliability as improved [49].

#### *2.4. Improvement of Power System Stability*

FACTS devices offer significant stability services, including:


#### *2.5. Multiple Grid Services*

The coordinated use of multi-type FACTS devices offers multiple grid services [65–69]. The multiple grid services are mathematically formulated as multi-objective optimization problems. An optimally allocated UPFC simultaneously minimizes total active power loss and maximizes power system predictability in systems with a high penetration of wind power [65]. In these systems, predicting the system state is very difficult due to the uncertainties in wind power generation.

An optimally allocated UPFC simultaneously minimizes transmission lines' total active power loss and maximizes the voltage stability limit [66]. The best results are obtained when the transformer taps are optimized in combination with the optimization of the UPFC location and parameter settings.

Optimally allocated TCSCs and SVCs simultaneously minimize the total active power loss, minimize load voltage deviation, and maximize static voltage stability margin, considering single outage contingency criterion, line thermal limits, and bus voltage limits [67].

An optimally allocated hybrid flow controller (HFC), phase shifting transformer (PST), and UPFC simultaneously minimize total active power loss, total fuel cost, and cost of FACTS installation, and maximize power system loadability [68]. The HFC provides better results in comparison with PST and UPFC.

Optimally allocated TCSC, SVC, and UPFC simultaneously minimize total active power loss, and minimize system operation cost that includes the cost of FACTS, the energy loss cost, and the congestion cost [69]. The FACTS' location, size, and parameter settings are optimized in combination with existing reactive power sources.

#### **3. Current Practices and Future Trends in the Testing of FACTS**

#### *3.1. Conventional Testing of FACTS*

The testing of distributed energy resource inverters according to ancillary service requirements has been well established recently. Several standards and guidelines are in place stating requirements and respective testing procedures [70,71]. Factory testing takes place at manufacturer's facilities, compliance testing at independent accredited institutes, and, finally, commissioning testing occurs in the field. On the other hand, the testing of FACTS is more challenging mainly due to their high rating and large size. For example, the typical capacities of recent transmission STATCOMs applied in the USA are +/−100 to +/−200 MVAr with some large units reaching +/−250 MVAr, while SVCs can have capacities up to 500 to 600 MVAr [4]. For this reason, most of the studies on FACTS preform digital simulations, while fewer involve laboratory or field tests.

In practice, digital simulations (e.g., transient stability, dynamic performance) are used for the functional specification of a FACTS device, which refers to the definition of equipment requirements, typically performed by the buyer (typically the TSO) or a consultant [72]. IEEE 1031: 2011 [73] proposes an approach to prepare a specification for a transmission SVC using conventional thyristor technology, which can be partly used for STATCOM and other devices. The guide describes newer developments in SVC component equipment and, particularly, control systems and also the latest practices for SVC applications among other topics. Similarly, IEEE P1052/08 [74] aims to assist users in specifying the functional requirements for transmission STATCOMs, using forced commutated technology based on voltage source converter topologies. The guide covers specifications, applications, engineering studies, main component characteristics, system functions and features, factory testing, commissioning, and operations of the STATCOM systems. It addresses the following functions: Reactive power compensation, voltage regulation and control, transient and dynamic stability, and control and protection.

The manufacturing of the device, based on the specification, is followed by factory testing at the facilities of the manufacturer. This includes the testing of the valves as the most critical component. Standards for the testing of valves are IEC 61954:2011 [75] for the thyrisors of SVCs and IEC 62927:2017 [76] for voltage source converter valves of STATCOMs. More specifically, IEC 61954:2011 defines the type, production, and optional tests on thyristor valves used in thyristor controlled reactors, thyristor switched reactors, and thyristor switched capacitors, forming part of SVCs. Type tests aim to verify that the valve design meets the requirements specified (e.g., dielectric and operational tests), production tests aim to verify proper manufacturing (e.g., voltage withstand check), and optional tests are additional to the type and production tests (e.g., voltage transient test). IEC 62927:2017 applies to self-commutated valves for use in voltage source converters for STATCOMs. It includes type tests (e.g., dielectric, operational and electromagnetic interference tests) and production tests (e.g., voltage withstand check) for air insulated valves. Moreover, the manufacturer also tests all other components of the FACTS device (e.g., transformer, circuit breakers, etc.) according to their respective standards. Particular attention needs to be paid to the testing of the control system, which includes dynamic performance tests and protection system tests, among others. More information on the testing of the control system is provided Section 3.2.

When factory testing by the manufacturer is completed, field testing follows in order to verify the specified performance in real operating conditions. It should be noted that this is an important step as in many cases, FACTS are built to satisfy the needs of specific projects. According to IEEE 1303:2011 [77], a field testing program for SVCs should include the following: Equipment tests within the SVC system, tests of the various subsystems that comprise the SVC system, commissioning tests for the complete SVC system, and acceptance testing of the complete SVC system. Equipment, subsystem, and commissioning tests are usually performed by the supplier, while the acceptance tests are performed by the buyer or user. The standard provides general guidelines and criteria for the field testing of SVCs, before they are placed in-service. It identifies the main elements of a field testing program

so that the user can formulate a specific plan that is most suited for his/her own SVC. Parts of the standard are useful for compensator systems using gate turn-off thyristor technology (STATCOM) or other semiconductor devices, such as insulated gate commutated transistors.

Apart from the aforementioned standards and guides, which focus on SVC and STATCOM, other relevant standards and guides have been published. IEEE 1676:2010 [78] defines control architectures for high-power electronics. It covers FACTS, high voltage direct current (HVDC) systems, distributed generation, and energy storage, among others, with a power range from hundreds of kW to thousands of MW, but with emphasis on 1 MW to hundreds of MW. IEEE 1534:2009 [79] provides recommended practices for specifying TCSC installations used in series with transmission lines. Ratings for TCSC thyristor valve assemblies, capacitors, and reactors as well as TCSC control characteristics, protective features, testing, commissioning, training, operation, etc. are addressed in this standard. IEEE P2745.1 [80] provides functional requirements for UPFCs, using modular multilevel converters (MMCs), including application conditions, system architecture, function requirements, performance requirements, primary equipment requirements, control and protection requirements, testing, etc. Concerning the testing of valves, IEEE 857:1996 [81] and IEC 60700-1:2015 [82] provide recommendations for the type testing of thyristor valves for HVDC power transmission systems. IEC 62501:2009 [83] deals with the testing of self-commutated converter valves, for use in voltage source converters for HVDC power transmission.

In the literature, a plethora of papers performing simulation studies of FACTS is available, as shown in Section 2. On the other hand, limited experiences on hardware testing are reported. Indicative papers are presented next. Concerning lab testing, [84] explains the difficulty of testing HV thyristor valves, due to the increasing capacity of FACTS devices. A synthetic test setup was applied, where the large current and the high voltage are generated by different power supplies. Moreover, test facilities that can perform operational tests on thyristor valves are reported (e.g., [85]). In [86], real scale validation tests in the laboratory of a large 40 MVAr voltage source converter for a SSSC were carried out. The device is formed by four 3-level 10 MVA inverters, and a series of tests was carried out, including dynamic performance and harmonics. In [87], the performance of prototype tests of a 26 MVA STATCOM and a 13 MVA back-to-back, comprising smaller units connected in parallel, is described. A scaled-down setup for testing series connected FACTS (e.g., SSSC, UPFC, TCSC) was proposed in [88].

Experiences of field testing of FACTS can be found in the literature, starting from more than 20 years ago [89]. Aspects of the planning and execution of the commissioning and testing of +300 MVAr (capacitive) to −100 MVAr (inductive) SVC are described in [90]. Challenges concerning the commissioning and testing of the SVC connected to a relatively weak grid are reported, while considerations on the impact of the tests on transmission and distribution system equipment, generating facilities, and customers are provided. Dynamic performance tests were performed, by varying the reference voltage of the SVC, due to the difficulty of controlling the grid voltage in the field. In [91], a comparative analysis between field tests during commissioning of a −50/+70 MVAr SVC and computer simulations was carried out. The tests included the application of steps in the reference voltage of the SVC and the switching off a capacitor bank among others. In [92], a 1 MVA FACTS device, entitled "power router", designed for power flow control was tested in the field, including dynamic tests. In [93], a large load disturbance test on an HV line with a TCSC was performed (by opening-reclosing of a circuit breaker) combined with a wide area measurement system (WAMS). The field tests and simulations showed that the current issue of a transmission bottleneck of hydroelectric power could be resolved. Additional experiences on the field testing of FACTS have been reported e.g. [94–96].

From the above review, it is clear that important progress in the development of appropriate setups and procedures for laboratory/factory and field testing of FACTS has been performed. However, the standards and guides cover only some types of FACTS devices, while grid services are not thoroughly addressed. Due to the large rating of FACTS devices, it remains a challenge to fully validate all the device functionalities, and especially grid services provision, in the laboratory. For example, inverters that are interfacing distributed generation are typically tested in the lab by a grid simulator that can simulate several network conditions (e.g., voltage dip, frequency rise, etc.). However, such devices are hard and expensive to build for high ratings. Moreover, conventional lab testing treats the FACTS device as an independent component, while neglecting interactions with other devices (e.g., distributed generation, other FACTS, etc.) and the overall power system. This is becoming increasingly important due to the need of FACTS to provide advanced grid services in the complex contemporary environment. On the other hand, field tests can validate the performance of FACTS in actual conditions; however, limitations are present, such as limited flexibility and repeatability, as well as the possibility of adversely influencing network equipment. Digital real-time simulation is a promising tool that can tackle some of the aforementioned limitations. Its use for the lab testing of FACTS will be discussed in the following section.

#### *3.2. Hardware in the Loop Testing of FACTS*

Real-time hardware in the loop (HIL) simulation is gaining significant attention as an advanced power system testing method [11,97–100]. It allows the connection of a physical device to a power system that is simulated in real-time in a digital real-time simulator (DRTS). Exhaustive testing can be achieved in realistic, flexible, controllable, and repeatable conditions that allow de-risking equipment. The two main categories of HIL simulation are shown in Figure 1. In CHIL simulation, a hardware controller is connected to a simulated system executed in the DRTS. In PHIL simulation, a physical power device (e.g., motor, inverter) is connected to a real-time simulated system. In PHIL simulation, a power interface consisting of a power amplifier and sensor is required in order to connect the hardware under test with the DRTS. The voltage of the common node between the simulation and hardware is amplified and applied on the hardware under test. The measured current of the hardware under test is fed back to the DRTS and is inserted in a controllable current source in the real-time simulation. Stability and accuracy issues arise during PHIL simulation due to the non-ideal power interface, which needs to be considered.

**Figure 1.** Basic topologies of controller hardware in the loop (CHIL) and power hardware in the loop (PHIL) testing [101].

Decades ago, analogue simulators or transient network analyzers (TNAs), composed of scaled-down physical models, were used for FACTS and HVDC controller testing, which were simplified and presented limited flexibility [102,103]. The emergence of digital real-time simulation allowed a realistic representation of the network and FACTS converter, allowing efficient testing of FACTS controllers and thus leading to fewer problems at the commissioning phase and field operation. This allowed the optimal tuning of the controller parameters, debugging, and repairing. Since then,

several HVDC and FACTS manufacturers have been using digital-real-time simulation, and specifically CHIL simulation, for dynamic performance and factory acceptance tests [102,103]. It should be noted that the possibility of using digital real-time simulation for FACTS controllers testing is mentioned in guides/standards [73]. The FACTS actual control system is connected to a DRTS that simulates the power network and also the FACTS power circuit (valves, filters, circuit breakers, etc.) [104]. The actual control system receives measurements from the simulated VTs/CTs, sends firing pulses to the simulated converter, and trip commands to simulated breakers. More recently, voltage source converters, which are part of certain FACTS devices, have made use of MMCs, which can include hundreds of sub-modules, therefore requiring hundreds of input/output channels and posing additional challenges for testing and real-time simulation. Indicative examples of CHIL testing of FACTS are briefly presented below. In [105], an SVC controller intended to damp sub-synchronous oscillation was tested using CHIL simulation. In [106], comparisons between CHIL tests and field tests of an SVC were performed. The CHIL tests allowed the efficient tuning of the SVC controller, as well as effective fault tracing. A wide-area control of a STATCOM using a hybrid CHIL and software-in-the loop configuration was validated in [107]. The transient stability enhancement using a wide-area controlled SVC was investigated in a CHIL configuration [54]. Additional applications of CHIL simulation for the testing of FACTS and HVDC controllers are reported in [108].

A more recent trend is the CHIL testing of replicas of the control system of HVDC and FACTS devices by utilities, mainly TSOs. The complexity and increasing importance of HVDC and FACTS in contemporary power systems, the difficulty of maintaining simulation models (e.g., black box models provided by the manufacturer)—leading to differences between the modeled controllers and the actual controllers in the field—and the need to test control changes in a safe environment before field implementation are key drivers for this development [102,103,109]. Certain utilities operate a DRTS, and obtain from the HVDC or FACTS manufacturer a replica of the control/protection system, which is an exact copy of the actual control/protection system installed on site [110]. Moreover, the human–machine interface of the replica offers the same functionalities with the one installed in the field [102]. CHIL platforms with hardware replicas of the control system allow the investigation of system events and the optimum tuning of the control and protection systems [111]. Moreover, interactions between several HVDC systems and FACTS devices can be studied (e.g., provided by multiple vendors), maintenance can be supported, future software updates can be validated, and the training of engineers can be facilitated [112]. As a major source of modeling uncertainty is removed, such platforms are considered to be a reliable tool for longer term studies [109,111].

The application of CHIL simulation for FACTS testing is an industry practice, either with factory testing of the control system by the manufacturer, or in cases with the testing of a replica by the utility. On the other hand, the use of PHIL simulation for testing of FACTSs has been much less explored. PHIL simulation has been successfully used in the testing of PV inverters, wind generation systems, motors, drives, etc., and is also under consideration for standardized testing [97,99,113]. Concerning FACTSs, PHIL simulations were performed in [114]; however, a hardware PV inverter was used, while a D-STATCOM was simulated in the DRTS. PHIL simulation for studying the synchronization issues of voltage source converters, including STATCOMs, was performed in [115]. Contrary to CHIL simulation, where only the control system is tested, in PHIL simulation, the hardware power device is tested, which includes both the control system and power circuit. PHIL simulation can test the actual end-product that is going to be installed in the field, which is not possible with CHIL simulation. Physical valves, filters, circuit breakers, etc. are used, instead of a real-time simulation model. In this way, complex interactions of the actual device with other simulated power system components (e.g., distributed generation, FACTS from different vendors, etc.) within the power system can be examined. This is increasingly important for testing the grid services provided by FACTSs. Therefore, it is suggested that efficient PHIL tests could reduce the need for some of the field tests, or reveal hidden issues prior to field testing. On the other hand, the performance of PHIL testing of full-scale transmission FACTSs is difficult, as a power amplification unit is required with at least the same rating

as the hardware under test. PHIL simulation capabilities in the 5 MVA [97] and 7 MVA [116] range have been reported, while the rating of AC grid simulators (that can be used as power amplifiers) is constantly increasing (e.g., 12 MVA or higher [117]). However, reaching the rating of large transmission FACTSs is not feasibly nowadays. One solution would be the PHIL testing of each converter separately, if the transmission FACTS is composed of different converters. Another approach would be the PHIL testing of a scaled-down model of the FACTS device. Initial results of this approach are presented in Section 3.3.2. Special investigations are necessary to ensure that the scaled-down model adequately represents the full-scale device. It should also be noted that the non-ideal power interface used in PHIL simulation can, in certain cases, render the experiment unstable or compromise the accuracy, therefore, attention should be paid to this issue [118,119].

Figure 2 summarizes the development and testing stages of FACTSs that were described in Section 3.1 (i.e., functional specification, factory testing, and field testing), including advanced approaches, such as the CHIL testing of the control system by the manufacturer and the CHIL testing of a replica controller by the owner, as presented in Section 3.2. Moreover, the option of performing the proposed PHIL tests (full scale or scaled-down) is included as a new step after the CHIL testing by the manufacturer.

**Figure 2.** Process from the functional specification to field testing of flexible alternating current transmission systems (FACTS), including the proposed PHIL testing stage.

#### *3.3. Testing Results*

Laboratory testing of a transmission STATCOM providing voltage control was carried out. First, the controller of a 100 MVA STATCOM was tested in a CHIL configuration. Then, a physical scaled-down model of the STATCOM (1 kVA) was tested in a PHIL configuration.

#### 3.3.1. Controller Hardware in the Loop Tests

The Electric Energy Systems laboratory of the National Technical University of Athens (NTUA) operates a DRTS by RTDS Technologies Inc. (including 2 GPC processors), which allows the simulation of switching phenomena in small time-steps in the range of 1 to 4 μs. A transmission system was designed in the DRTS, as shown in Figure 3. A 3-level 3-phase DC/AC voltage source converter of a 100 MVA STATCOM, including DC bus, filters, and transformer, was simulated in the DRTS, while its control algorithm was executed on a hardware controller. Voltage and current measurements from the real-time simulation were fed to the hardware controller, which returned the modulation signal to the

simulated DC/AC converter. It should be noted that the conversion of the modulation signal to PWM pulses was performed in the real-time simulation, in order to match the operation of the actual voltage source converter used for PHIL tests in Section 3.3.2. The control algorithm of the STATCOM aims to provide dynamic and steady state voltage control by comparing the locally measured grid voltage (bus 2) with a reference voltage and providing or absorbing reactive power. The model of the STATCOM was based on [120] using the following values for the proportional integral (PI) controller of the outer AC voltage control loop: *Ki* = 1000, *Kp* = 10.

**Figure 3.** CHIL test setup for the STATCOM controller.

A 3-phase symmetrical fault occurred at bus 1, which was cleared after 13 cycles (fault impedance equals to 100 Ohm resistive). CHIL tests were performed and the results are described below. The 3-phase current at the faulted bus is shown in Figure 4. Figure 5 shows the increase of the current of the STATCOM during the fault in order to provide voltage control, where only one phase is shown in order to facilitate comparison with the PHIL tests in Section 3.3.2 (in the PHIL test, a hardware single-phase voltage source converter was used).

**Figure 5.** Current contribution of the STATCOM during the fault (CHIL test).

Figures 6 and 7 show the resulting voltages of bus 2 and bus 3 during the short circuit, with and without the STATCOM. The CHIL results show that the mitigation of voltage deviations in dynamic and steady state conditions by the STATCOM is clear. The voltage improvement is greater at bus 2, as this is the point of connection of the STATCOM. The dynamic behavior of the STATCOM influences the grid voltages shortly after fault clearance, however, without adverse effects. Figure 8 shows the reactive power provision by the STATCOM in steady state and dynamic conditions in order to mitigate the voltage sag. Naturally, the reactive power provision is reduced after fault clearance. Active power is absorbed by the STACOM in order to maintain a constant DC bus voltage and it is shown that the active power is increased when the reactive power provision rises.

**Figure 6.** Voltage of bus 2 during the short circuit, with and without the STATCOM (CHIL test).

**Figure 7.** Voltage of bus 3 during the short circuit with and without the STATCOM (CHIL test).

**Figure 8.** Active and reactive power of the STATCOM during the short circuit (CHIL test).

#### 3.3.2. Power Hardware in the Loop Tests

After the successful completion of the CHIL tests, the hardware controller was connected to a physical voltage source converter in order to verify the performance of the device in a PHIL setup (Figure 9). As the rating of the transmission STATCOM is high (i.e., 100 MVA, 3-phase), it was not possible to perform full-scale tests. A small-scale 2-level voltage source converter (i.e., 1 kVA, 1 phase) was used in order to verify the performance of the full-scale device. The converter was connected to the real-time simulated system using a suitable power interface consisting of a fast and accurate linear power amplifier and a current sensor. The current sensor used was a current probe (Tektronix A622, scaling 100 mV/A), with a small time-delay and compatible output voltage range with the analogue input card of the RTDS. The high voltage of the transmission system was scaled down in the DRTS, in order to reach the low voltage level of the physical converter, while the actual current provided by the converter was scaled up in order to virtually increase the rating of the device, following a similar approach to [119]. Specifically, the voltage from the simulation was multiplied by the ratio of the nominal voltage of the hardware under test (*VHWn* ) and the full scale device (*VSWn* ) according to Equation (1). Similarly, the measured current of the hardware was multiplied by the ratio of the nominal current of the full scale device (*ISWn* ) and the nominal current of the hardware under test

(*IHWn* ) according to Equation (2). It should be noted that the single phase hardware under test (1 kVA) was seen as a 3-phase in the simulation (i.e., 3 kVA) by applying a similar approach to [114]:

$$a = \frac{V\_{HW\_n}}{V\_{SW\_n}} = \frac{400\text{ V}}{500\text{ kV}} = 0.0008,\tag{1}$$

$$b = \frac{I\_{SWn}}{I\_{H\&n}} = \frac{100 \text{ MVA} / \left(\sqrt{3} \cdot 500 \text{ kV}\right)}{3 \text{ kVA} / \left(\sqrt{3} \cdot 400 \text{ V}\right)} = 26.67. \tag{2}$$

**Figure 9.** PHIL test setup of the scaled-down STATCOM.

The CHIL tests of Section 3.3.1 were repeated in a PHIL configuration. Figure 10 shows the current at the faulted bus during the short circuit in the PHIL test, which is very similar to the CHIL test. Figure 11 shows the current of the actual voltage source converter before, during, and after the fault is cleared, measured by the aforementioned current probe and obtained from the software of the DRTS. This current was scaled up in the DRTS and was injected in the simulated power system via a controllable current source.

**Figure 11.** Current of the hardware voltage source converter during the short circuit (PHIL test).

The voltages of bus 2 and bus 3 during the short circuit, with and without the STATCOM, are shown in Figures 12 and 13. It is clear that the STATCOM manages to mitigate the voltage sag at both buses, similarly to the CHIL tests. The occurring voltages in the steady state and dynamic conditions are similar to the CHIL test (Figures 6 and 7), however, there is a smaller voltage improvement under dynamic conditions (around 2 p.u less at bus 2). The smaller voltage improvement is due to less reactive power provided by the converter in the PHIL test (Figure 14), compared to the CHIL test (Figure 8). The hardware voltage source converter exhibits additional constraints on its maximum admissible current, based on its actual switching devices and power circuit. In the CHIL test, where the voltage source converter was purely simulated, this was not considered accurately. Therefore, in the PHIL test, the less reactive current allowed to mitigate the voltage sag resulted in a smaller improvement of the voltage compared to the CHIL test.

**Figure 12.** Voltage of bus 2 during the short circuit with and without the STATCOM (PHIL test).

**Figure 13.** Voltage of bus 3 during the short circuit with and without the STATCOM (PHIL test).

**Figure 14.** Active and reactive power of the STATCOM during the short circuit (PHIL test).

Finally, the aforementioned CHIL and PHIL test results were compared with pure digital simulations. In the pure digital simulations, the power system, the control algorithm of the STATCOM as well as its power circuit were simulated in real-time using the software of the RTDS (RSCAD). Figure 15 shows that the pure simulation and CHIL test result in very similar voltages at bus 2 with the operation of the STATCOM. Small deviations are shown shortly after fault clearance. On the other hand, the PHIL test results in a smaller voltage improvement, which is due to the less reactive power provided, as already explained. Similarly, Figure 16 shows the similar behavior of the current contribution of the STATCOM during the pure simulation and CHIL test. It is clear that the reactive current of the hardware voltage source converter (scaled up in the simulation) during the fault was smaller in the PHIL test. Moreover, the non-ideal behavior of the hardware device at fault clearance (*t* = 0.48 − 0.52 s) was shown in the PHIL tests.

**Figure 15.** Voltage of bus 2: comparison of pure digital simulation, CHIL test, and PHIL test.

**Figure 16.** Current contribution of the STATCOM: comparison of pure digital simulation, CHIL test, and PHIL test.

#### **4. Discussion**

This paper addressed the provision of grid services by FACTS and their testing considering complex contemporary and future power systems. Several research articles, but also review papers summarize the research findings on FACTS devices and their grid services in a comprehensive manner. However, there is no review paper focusing on the different grid services provided by various FACTS devices. This paper aimed to cover this gap by providing a comprehensive literature review on the different grid services provided by various FACTS devices. Moreover, limited experiences on hardware testing of FACTS have been reported that mostly present specific test cases, for a specific FACTS device at a specific location, while a review on the testing of FACTS is missing. This paper addressed this gap by providing a literature review of standards, guides, and scientific papers on the testing of FACTS and presented an overview of all the different testing stages, including conventional testing and emerging industry practices. Moreover, PHIL simulation for the testing of FACTS was suggested as an efficient testing method prior to field testing.

Conventional laboratory testing of FACTS treats the device as an independent component, while neglecting interactions with other devices and the power system. On the other hand, field testing has limited flexibility and repeatability, as well as the possibility of adversely influencing network equipment. The advantages of CHIL testing of FACTS controllers to overcome the above limitations were explained, which is already a practice of manufacturers. It was explained that recently, owners of FACTS, like TSOs, obtain a replica of the control and protection system by the manufacturer and perform CHIL tests, in order to reduce modeling uncertainty in the long-term, provide optimal tuning of parameters, and facilitate maintenance and future updates. As CHIL simulation allows testing of only the control system and not the actual end-product, the use of PHIL simulation is suggested as a further step. PHIL simulation can test both the control system and power circuit (i.e., valves, filters, transformer, circuit breakers, etc.), reducing the uncertainty towards field testing. Full-scale PHIL testing of D-STATCOMs and relatively small transmission STATCOMs, if they consist of several converters that are tested separately, is feasible. However, the full-scale PHIL testing of large transmission STATCOMs is not feasible nowadays. Therefore, the execution of scaled-down PHIL tests was discussed. Indicative CHIL and scaled-down PHIL tests of a transmission STATCOM providing voltage control were performed. The CHIL and PHIL results presented a similar behavior at steady state and dynamic conditions. The differences were due to the current limitations of the hardware device that were not considered accurately in the CHIL tests. The CHIL and PHIL results were also compared with pure digital simulations.

In future work, PHIL testing of a full-scale D-STATCOM could be performed in a suitable laboratory in order to compare with conventional factory tests and CHIL tests. In this way, the specific benefits of performing PHIL simulation prior to field testing could be highlighted. Moreover, the effectiveness of scaled-down PHIL tests and their accurate representation of full-scale tests could be further investigated.

**Author Contributions:** P.K. and P.G. developed the main idea and structure of the paper. D.T.L. and V.K. performed the hardware in the loop tests. P.K. and P.G. wrote the paper. N.H. supervised the work and reviewed the paper.

**Funding:** This work is supported in part by the European Commission within the Horizon 2020 framework ERIGrid project under Grant agreement 654113.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
