Grid-forming control strategies for blackstart by offshore wind farms

Large-scale integration of renewable energy sources with power-electronic converters is pushing the power system closer to its dynamic stability limit. This has increased the risk of wide-area blackouts. Thus, the changing generation profile in the power system necessitates the use of alternate sources of energy such as wind power plants, to provide blackstart services in the future. This however, requires grid-forming and not the traditionally prevalent grid-following wind turbines. In this paper, four different grid-forming control strategies have been implemented in an HVDC-connected wind farm. A simulation 5 study has been carried out to test the different control schemes for the different stages of energization of onshore load by the wind farm. Their transient behaviour during transformer inrush, converter pre-charge and de-blocking, and onshore block-load pickup, has been compared to demonstrate the blackstart capabilities of grid-forming wind power plants for early participation in power system restoration.

work energisation (mainly long cables & lines), withstand transformer inrush transients, and cater to block loading, provided sufficient wind is available and controllers are adapted for blackstart. This makes HVDC transmission preferable for future OWFs due to the large charging Var required for long distance HVAC cables (Erlich et al., 2013).
The current turbine & converter controls are designed assuming a strong grid connection point which means that the gridside converter (GSC) of the WT latches onto a pre-existing voltage signal provided by the onshore grid in case of an AC-70 connected OWF, or produced by the offshore HVDC converter operating in voltage-frequency control mode in case of HVDCconnected OWF (Bahrman and Bjorklund, 2014). However, to allow outward-energisation of the network of inter-array cables & transformers, create a power island that can supply local loads and energize the HVDC link converters & export cable with the ultimate aim to supply onshore block load, the WT should be able to produce its own voltage signal. This requires gridforming, traditionally referred to as voltage-injecting control, as opposed to the conventional grid-following or current-injecting 75 control. The two control philosophies are very well explained by (Rocabert et al., 2012). Voltage-injecting control concepts have been shown to deliver a superior performance compared to their current-injecting counterparts in system-split scenarios as demonstrated by Heising et al. (2019). Erlich et al. (2017) also shows that the temporary over-voltages (TOV) following islanding due to transformer-cable interaction can be avoided. Moreover, grid-forming WTs can also minimize the use of diesel generators that are currently employed to supply backup auxiliary power required for energization. Although, most modern 80 WTs have an on-board UPS to power communications, protection & control for few hours during emergency shutdown (Göksu et al., 2017), a larger internal backup supply would be required for self-starting the WT for blackstart, especially after extended shutdown periods.

Hybrid
Traditional current-controlled WTs can be used with an external power supply (eg. diesel generator/energy storage) and a 85 Synchronous Var Generator (SVG) or STATCOM, combining services into a joint hybrid blackstart unit (BSU) to facilitate WT participation in BS procedure as proposed in Aktarujjaman et al. (2006). The external supply provides startup power and sets the reference voltage & frequency for the isolated system, the SVG/STATCOM supports the Var requirement of the cables & transformers and stabilizes the voltage, after which the WTs connect to meet the load power demand. Zhu et al. (2018) shows that earlier participation of WT in the restoration procedure is feasible as grid-forming control allows blackstart & 90 stand-alone island operation with better inherent synchronous-machine like inertial response during a transient, that can help absorb the initial impact of energization and ensure smooth load pickup, thus mitigating large voltage/frequency excursions that might occur during restoration. However, only the transients during load pickup and re-synchronisation to the grid have been studied, while energization of collector lines, export cables & transformers, that cause more transient stability challenges during energization, are not shown. Additionally, the major energization transients are taken by the ESS & SVG, while the 95 WTs behave as passive power-sources to meet load-demand during the last stages of restoration.

HVAC
Recent studies by Martínez-Turégano et al. (2018) and Aten et al. (2019) demonstrate the potential capability of HVACconnected OWFs to blackstart onshore grid using grid-forming controls in less than 25% WTs and assuming adequate wind resource. The results show that it is possible to do sequential energization of the array-cables & WT transformers, starting 100 with one WT energizing its string followed by others synchronizing to it and then sharing the control of voltage & frequency.
Shorter cable sections are energized first until enough WTs are connected to absorb the Var generated by subsequent cable sections. However, according to Elia (2018) & National Grid (2019b), a large gap to bridge is the energization of the export link while meeting grid code requirements.

105
HVDC with Voltage Source Converters (VSC) can also be used as a standby facility for blackstart and restoration of the onshore AC grid, as demonstrated by the excellent voltage and frequency control performance in real system tests done by Jiang-Hafner et al. (2008), proving for the first time that VSC-HVDC helps reduce restoration time while facilitating a safer & smoother restoration process with lower investment & maintenance cost. With HVDC transmission gaining momentum as the preferred choice for longer distance connections to larger OWFs, Sørensen et al. (2019) show that the Skagerrak-4 (SK4) 110 VSC-HVDC link between Norway & Denmark (DK) can be successfully used to ramp-up the voltage of an islanded 400 kV & 150 kV DK-network to energize overhead transmission lines, transformers and block load, followed by synchronisation to continental EU. Additionally, a top-down restoration test of the NEMO link between Belgium & the UK also demonstrates the capability of the VSC-HVDC interconnector to energize a dead Belgian grid from the live UK side (Schyvens, 2019). However, a diesel generator was used to provide auxiliary power for the dead-side converter. Simulation results by Becker et al. (2017) 115 5 https://doi.org/10.5194/wes-2020-34 Preprint. Discussion started: 17 March 2020 c Author(s) 2020. CC BY 4.0 License.
show, without any details of the transformer/cable energization transients, that a VSC-HVDC connected OWF can respond to onshore load changes and participate in load restoration. Cai et al. (2017) analyzes the inrush current of transformers and cables (HVAC/HVDC) using EMT simulations, but with a diesel generator to pre-charge the offshore converter that then energizes the offshore collector grid, and the onshore converter pre-charged from onshore AC-grid, contrary to what is expected from an OWF to provide blackstart service. Simulation results presented by Sakamuri et al. (2019) demonstrate, for the first time, 120 an HVDC-connected OWF with grid-forming control, sequentially energizing the offshore AC network including transformer, cables & converter through a pre-insertion resistor, followed by HVDC link energization and onshore converter pre-charging & de-blocking for picking up block load, successfully participating in restoration as a BSU. However, the energy imbalance in the HVDC link during the DC-side uncontrolled pre-charging of the onshore converter leads to a significant dip in HVDC voltage and large transients in the offshore & onshore converter cell voltages & valve currents. Grid-forming control, in addition to 125 enabling blackstart and islanding capabilities of WTs, can also allow the use of Hybrid-HVDC connection with a diode rectifier unit (DRU) instead of the offshore VSC. The application of controls proposed in Blasco-Gimenez et al. (2010) for an OWF to ramp up the offshore AC grid voltage & control frequency, considering it as an inverter-based microgrid, has shown improved steady-state regulation during islanding when the DR-HVDC is not conducting, and smooth transition to current-control during grid-connected operation. This significantly reduces the cost-vs-performance, due to lower losses, especially for higher power 130 levels, and lesser capital cost, along with increasing efficiency & reliability due to a lower probability of commutation failure than a VSC (Andersen and Xu, 2004).

2.3
This paper attempts to provide a generalized structure of different grid-forming control strategies that can be applied for controlling wind turbines as voltage sources to enable the blackstart and islanding capabilities of offshore wind farms. It then 135 focuses on testing four different methods during the various stages of energization of an onshore block-load by an HVDCconnected OWF. The aim is to characterize the different techniques and compare their capability to deal with the transients in a controlled manner while maintaining stable voltage and frequency at the offshore terminal.

Grid Forming
Grid forming control of power electronic converters has been well studied for microgrids, where the role of PECs is to act 140 as an interface between the small-scale distributed/renewable power generation units and the consumption points, leading to intertial decoupling of the rotating machines and making the microgrid system susceptible to oscillations caused by network disturbances. Grid forming allows a PEC to mimic Synchronous Generators (SG) for droop-based load-sharing, synthetic inertial-emulation, synchronized & stand-alone operation and blackstart behaviour, ensuring voltage and frequency stablility in low-inertia microgrids during varying loads, network disturbances and system configurational changes (islanding ⇐⇒ grid- An OWF is like a microgrid rich in power electronics although very different in that the voltage and power levels are much higher. Moreover, wind farm operators maintain a large amount (>100s) of WT-assets that are located very far from each other. Current sharing techniques for low rated inverters like the centralized controllers and the master-slave approach can be used only for paralleled systems that are close to each other and interconnected through high-bandwidth communication channels 150 (Rocabert et al., 2012). These communication-based solutions can not be used for microgrids spread across several kilometres, as ensuring globally available, bidirectional, reliable & robust, low-power secure communication architecture becomes increasingly costly. Moreover, larger communication links increase delays which is undesirable in cases where a fast high-bandwidth link is required. This gave way to droop control algorithms with a hierarchical structure being used in microgrids, especially for islanded operation of many micro-sources located far away from each other (Pogaku et al., 2007). Although rated at much 155 lower power, these grid-forming droop-based strategies can be extended to large power OWFs operating in islanded mode, taking into account their characteristcs, as demonstrated by Blasco-Gimenez et al. (2010).

Control Structure
According to the definition in Rocabert et al. (2012), grid-forming converters are controlled in closed loop to work as ideal AC voltage sources (low-output impedance), while grid-feeding/following converters are controlled as current sources with high 160 parallel output impedance and can't operate in islanding/stand-alone mode as they require a grid-forming converter or local SG to set the bus voltage and frequency. 7 https://doi.org/10.5194/wes-2020-34 Preprint. Discussion started: 17 March 2020 c Author(s) 2020. CC BY 4.0 License.
The control structure of grid-forming control consists of different functional blocks, as shown in figure 1. Since the main objective of grid forming control is to operate the PEC as an ideal AC voltage source of given amplitude V * o and frequency ω * , it consists most importantly of a voltage control loop C V . The short-comings of the single-loop approach, explained in 165 Zeni et al. (2015), are already known from switch-mode power supplies and electrical machine drives as over-currents during transients & faults can not be limited due to the lack of an explicit closed-loop current controller. In addition, sensitivity to disturbances and plant-parameter fluctuations eliminates open-loop control as a good choice. The most commonly used alternative thus, is the nested/cascaded voltage-current controller (Zeni et al., 2015), in which a faster inner current control loop C I is added. C I is designed to have a relatively smaller time constant than C V , for decoupling i.e. C I behaves as an almost 170 perfect current controller for the slower C V . The controllers are in the synchronous reference frame that uses an angle θ * (for abc dq transformation) obtained from the synchronization block .
While grid-feeding converters require perfect synchronism with the AC voltage at the point of connection to accurately regulate the power exchange with the grid, in the case of grid-forming converters the synchronization system must provide precise signals for both islanded and grid-connected modes of operation. It works as a fixed frequency ω * oscillator in the former 175 case while slowly varies the phase-angle & frequency of the island voltage during the reconnection transient to resynchronize with the grid voltage, in the latter. The most extended method used is a Phase Locked Loop (PLL), also called voltage-based synchronization as the frequency and phase-angle of the grid voltage vector is used for control. However, enhancements are needed to ensure stability under unbalanced and distorted voltage conditions as voltage sag, weak grids or off-grid operation can lead to instabilities. Alterately, power-based synchronization can also be used for synchronization as the structure of the 180 swing equation that governs synchronous machine (SM) dynamics, can be equated to that of a PLL, in the sense that the PLL structure can be modified to extract the derivative term of the frequency (inertia in SMs) and the speed variation (damping in SMs), as shown in van Wesenbeeck et al. (2009). This presents a more stable solution and allows the power controller to also act as the synchronization block.
The outer power control loops C P , C Q are required to regulate the real P and reactive Q powers exchanged with the grid 185 (in grid-connected mode) or meet the demand set by the load (in islanded mode), while ensuring communication-less real & reactive power sharing between the multiple paralleled inverters. The simplest method for this, by only relying on local measurements, is the droop control scheme, which was initially introduced for SGs in utility scale grids, and now is well incorporated into microgrids (Arbab-Zavar et al., 2019). The primary level of the 3-level hierarchical control, explained in Guerrero et al. (2011), employs droop control equations 1 & 2, based on grid X/R ratio, to mimic the self-regulation capability 190 of a grid-connected SG and allow power sharing in microgrids without using critical communication links (Rocabert et al., 2012).
between load-sharing & voltage-regulation, load dependent frequency-deviation, slow dynamic response due to filters for P, Q and non-linear load-sharing issues due to harmonics. A variable virtual impedance Z V can be used to add harmonic (f n ) droop characteristics and improve tradeoff between current harmonic sharing & voltage total harmonic distortion, by adjusting output impedance seen by harmonics and the fundamental. Additionally, this allows intelligent mode-switching with soft-start to take advantage of the fast converter response while avoiding large transients (Guerrero et al., 2011 Automatic Voltage Regulator (AVR) for reactive power sharing between the paralleled PEC-interfaced WTs.

Control Strategies
In the last decade, many different control solutions have been proposed in literature to replicate the system-level functionali-

Virtual Synchronous Generator (VSG)
The VSG control structure implemented here is shown in figure 2 voltage amplitude reference. Virtual impedance can also be added to reduce sensitivity to small grid disturbances by providing additional damping. Since a sufficiently high enough damping leads to a low inherent steady-state ∆P ∆f | ∞ droop, an external secondary virtual governor f -P droop is needed to mimic the traditional 3-5% SG speed droop characteristic.

Power Synchronous Control (PSC)
The PSC control structure, explained in Zhang et al. (2010), is shown in figure 2(c). Here also power synchronization is used 235 just like in VSG, however unlike the swing equation where the power difference drives the rotor speed dynamics which is then changed to electrical angle i.e. double integration for P -θ transfer function (equation 3), the the PSC Loop (PSL) directly gets the phase angle by a single integration of the power difference, as given by the equation 4. Due to 1 less integrator, PSC has higher stability margin, however due to 0 inherent steady-state droop, outer droops are needed for paralleling multiple grid forming units. Moreover, no virtual inertia or damping is present due to absence of rotor dynamics.

Distributed PLL-based (dPLL)
The dPLL control structure is based on Yu et al. (2018) and shown in figure 2(d). Originally developed for DRU-connected OWFs, the real power controller is used to generate the d-axis voltage reference as power flow is determined by offshore voltage, governed by equation 9, and a reactive power droop controller regulates frequency to share the DRU demanded Var.

260
Instead of the conventional approach of setting the q-axis voltage reference to 0, since PLL output can be used as an indication of frequency deviation, a Frequency Control Loop characterized by equation 10, is embedded in the q-axis to use the output of the Q-f droop-controller.
265 Yu et al. (2018) demonstrates frequency controllability with plug-and-play capability providing successful sequential start-up of the grid-forming WTs and automatic synchronization of the offline WTs during connection with minimal impact, to supply the Var required to energize transformers, filters and finally ramp up the offshore voltage and start delivering active power to onshore grid. However, only the start-up and synchronisation of an islanded OWF to an energized onshore synchronous power system via a DR-HVDC link is studied while the energisation of export cable and onshore converter, expected from a blackstart 270 service provider, was not looked into.

Direct Power Control (DPC)
DPC was introduced by Noguchi et al. (1998)   The OWF consists of 50 Type-4 (fully-rated PEC interface) 8 MW WTs, as a partially-aggregated model shown in figure   3(b), based on Muljadi et al. (2008). It consists of 9 individual WT 1−9 models on the first string, the second string with

305
The model described above has certain limitations. Firstly, the WT Rotor-Side Converter (RSC) & changes to the turbine controller that are required for grid-forming operation, have not been modelled. In conventional grid-following operation of the WT, the RSC is controlled to extract maximum power from the generator while the GSC maintains power balance to control the DC link voltage of the back-to-back PEC interface of the WT and the reactive power output at the AC terminal. However, in grid-forming mode, the GSC can not control the DC link & reactive power anymore and the required generator and their related issues like voltage surges or TOVs are not modelled. Moreover, the WT transformer is modelled as a pure electrical impedance r + jx without any magnetic characteristics as it can be soft-started along with the WT voltage ramp-up, to avoid magnetic inrush & saturation effects. Secondly, for this study, although power sharing between the WTs inside the WPP has been controlled by including the outer power control loops, the WTs are started-up simultaneously as opposed to a more realistic sequential energization (eg. Yu et al. (2018)), as the study mainly focuses on the capabilities of the grid-forming 320 OWF to provide blackstart services to the onshore grid while dealing with offshore network transients due to energization of the large converter transformer, HVDC converters and export cable, in a controlled manner. This puts any synchronization dynamics of multiple grid-forming PEC interfaced WTs out of the scope of this study. Additionally, substation load has not been modelled here because it is negligible with respect to onshore load.

325
In this section, the results of the dynamic simulations performed in PSCAD are presented. The energization sequence, events of which are described in detail in table 1, is based on Sakamuri et al. (2019), but includes an extra stage of DC-side controlled pre-charging of the onshore MMC cells and the outer C P,Q control loops for real & reactive power sharing amongst the WTs  inside the WPP. The entire sequence is simulated, however the main focus is on testing the characteristics of the different control strategies in enabling the OWF to deal with the energization transients and so we focus on the real & reactive power outputs of 330 the WPP, and with the voltage & frequency at the offshore PCC-2. Hard-switching is used here despite the advantages of softstart energization, as the former is more demanding on the grid-forming OWF in terms of the transients (eg. TOV, oscillations) linked to energization of transformer, cable and HVDC link. Figure 4 shows the waveforms for the real & reactive power outputs of the WPP during the different stages of the energization sequence. Since the grid-forming offshore WPP is operating in islanded mode, the real & reactive power demand is set by the 335 load, which depends on the particular stage of energization. For stage 2, it is the Var required for magnetic energization of the offshore transformer and AC-side pre-charging of the offshore MMC cells. A PIR is inserted for PIT to limit the inrush peak. In stage 3, power is required to energize the HVDC cable when the offshore MMC is de-blocked to control the HVDC link voltage, while in stage 4, the DC-side pre-charging of the onshore MMC cells draws power from the OWF to maintain the energy balance on the HVDC link. Finally, the OWF supplies power to match the onshore block load in stage 6. Since the scope 340 of this study is to focus on the offshore wind farm behaviour as a voltage source during the different stages of energization, the waveforms for the voltage & frequency at the offshore PCC-2 are presented in figures 5 & 6, respectively.   Figure 5. Voltage at offshore PCC-2 with zoomed insets to show transients in selected stages of the energization sequence.
The grid-forming WPP, controlled as a voltage source, has different characteristics based on the control method used. DPC is the most straightforward control technique with direct voltage and frequency control, without any inner loops, and doesn't have any electro-mechanical characteristics like inertia or damping. Thus it has the highest frequency swing, as can be seen can be seen especially, from the shifted frequency nadir after the 2.8s event in stage 4 of figure 6 and the shifted power peak after the 1.6s event in stage 2 of figure 4. Additionally, the non-minimum phase system type behaviour of the PSC due to its RHP zero is also evident from the frequency rise as opposed to dip for other 3 methods, at the 1.6s event in stage 2 of figure   6. Contrary to the power-synchronization based VSG & PSC methods, the frequency swing in dPLL is lowest in all stages 2-6 with no overshoot in stage 1, as is clear from the figure 6, due to the frequency controllability of the PLL-based frequency 355 control loop, which also provides damping and reduces the transient power peak, seen especially after the 1.6s event in stage 2 and stage 3 in fgure 4. However, after the 1.3s event in stage 2 in figure 4 the transient power peak for dPLL is higest during energization of the transformer with a PIR. This is because the dPLL is a voltage source controlled with Q-f , P -V coupling which not only leads to a drop in frequency when Var is demanded by the magnetic reactance of the transformer, but also causes a significantly higher distortion in voltage, compared to the other 3 methods, as shown in stage 2 of figure 5, due to the 360 PIR reducing the decoupling, to which the dPLL controlled WPP reacts with a surge in power output.
It is clear from the P Q waveforms shown in figure 4    during stages 2-4 that are recovered fast by the grid-forming controls. However, significant difference difference can be seen in the frequency transients in figure 6 which also demonstrates the characteristics of the different control methods. This is summarized in table 2.

Conclusions
Recent tests on HVDC interconnectors like SK-4 and NEMO link have shown that VSC-HVDC can be used for blackstart 370 services in a top-down restoration strategy. This makes VSC-HVDC connected offshore wind farms promising candidates for providing blackstart and islanding operation capabilities, as the conventional large thermal power plants are being phased out and wind farms grow bigger, to meet the decarbonization aims. This paper presents an overview of the different strategies for the participation of offshore wind in a traditional bottom-up power system restoration procedure and focuses on grid-forming as the main control change required to enable blackstart and islanding services from wind turbines, facilitate their earlier participation 375 and minimize the dependence on auxiliary diesel generators. The overall structure of grid forming control has been explained with the constituent functional blocks, along with conceptual explanation of 4 different techniques viz. VSG, PSC, dPLL and DPC. These methods were then tested in a study of the blackstart of onshore load by an HVDC-connected offshore wind