Augmenting Offshore Wind-Farm Yield with Tethered Kites

Abstract

Offshore wind-farm performance remains constrained by persistent wake deficits and turbulence that compound across intra-farm, intra-cluster, and inter-cluster scales, particularly under atmospheric neutral–stable stratification. A concept is advanced whereby offshore wind-farm yield may be augmented by pairing conventional horizontal-axis wind turbines (HAWTs) with lighter-than-air parafoil systems that entrain higher-momentum air and re-energise wakes, complementing yaw/induction-based wake control and enabling higher array energy density. A concise synthesis of wake physics and associated challenges motivates opportunities for active momentum re-injection, while a review of kite technologies frames design choices for lift generation and spatial keeping. Stability and control, spanning static and dynamic behaviours, tether dynamics, and response to extreme meteorological conditions, are identified as key challenges. System-integration pathways are outlined, including alignment and mounting options relative to turbine rows and prevailing shear. A staged validation programme is proposed, combining high-fidelity numerical simulation with wave-tank testing of coupled mooring–tether dynamics and wind-tunnel experiments on scaled arrays. Evaluation metrics emphasise net energy gain, fatigue loading, availability, and Levelized Cost of Energy (LCOE). The paper concludes with research directions and recommendations to guide standards and investment, and with a quantitative assessment of the techno-economic significance of kite–HAWT integration at scale.

1. Introduction

Although copious research into farm layout optimisation and turbine control strategies has led to a better understanding of and a successful reduction in wake effects, the industry faces an inflexion point where the assumption of infinite wind resources has become tenuous. The physical limits imposed by atmospheric dynamics may outpace current solutions as project densities increase, marking a transition from widely spaced developments to dense utilisation of offshore wind resources.

Offshore wind energy has undergone unprecedented global expansion in recent years. By the end of 2024, cumulative installed offshore wind capacity reached 83.2 GW [1].

China maintained its dominant position by seeing the addition of the most offshore wind capacity of any country for the seventh consecutive year, by commissioning 50 per cent of all new installations globally, bringing its national total to 41.8 GW, the highest of any country [1]. Europe installed 2.7 GW of new offshore wind installations, led by the United Kingdom and France, bringing its total to 37.6 GW [1]. Emerging offshore markets are also advancing rapidly; countries including Brazil, Japan, the Philippines, the Republic of Korea, and Vietnam are positioning themselves for post-2027 offshore build-outs.

Over the next decade, more than 350 GW of new capacity is anticipated, driving total global capacity to 441 GW by 2034, with annual additions surpassing the 30 GW milestone in 2030 and 50 GW by 2033 [1]. The Asia-Pacific region is set to lead this expansion, contributing 60 per cent of new installations (215 GW), while Europe is forecast to add 126 GW and North America to add 13 GW [1].

This scaling is not merely a logistical challenge of supply chains and grid connections but also a thermodynamic challenge. Recent large-scale studies indicate that sub-optimal layouts in very large offshore developments can experience wake-induced system losses on the order of 30–35 per cent and capacity factor penalties approaching 20 per cent for strongly wake-affected projects [2,3]. The review presented herein posits that the atmospheric wind resource can no longer be viewed as effectively limitless; rather, the clustering of gigawatt-scale projects generates a regional momentum sink that lowers capacity factors, and hence, the yield of offshore farms. To address this, a new perspective on wake mitigation is advanced herein, predicated on a novel mechanism of momentum entrainment.

Section 2, Section 3 and Section 4 provide a concise synthesis of multi-scale offshore wake physics, wake-loss evidence, and the advances and emerging limits of established mitigation approaches in dense clusters; this synthesis motivates the need for novel mitigation methodologies. Section 5 contains the original conceptual contribution of this article: a turbine-integrated, lighter-than-air parafoil operated as a controllable momentum-entrainment device to accelerate wake recovery. Some preliminary CFD results are presented therein to illustrate the plausibility of the proposed mechanism, not to provide a bankable performance prediction. Section 5.3 onwards outlines a staged validation roadmap and the metrics required to determine the net AEP/LCOE impact under realistic offshore atmospheric stability and turbulence conditions.

2. Multi-Scale Wake Losses as a Binding Constraint on Offshore Build-Outs

2.1. Turbine- and Farm-Scale Wake Physics

The rapid densification of offshore wind is increasingly constrained not only by turbine-scale wake physics but by multi-farm interference that propagates across clusters and national lease areas.

The wake is the primary aerodynamic signature of energy extraction: a coherent-to-turbulent flow region characterised by a deficit in streamwise momentum and elevated turbulent kinetic energy (TKE). Canonical wake structure, illustrated in Figure 1, comprises the upstream induction region, a near wake dominated by rotor geometry and coherent tip/root vortex systems, and a far wake that evolves toward self-similar behaviour with approximately Gaussian velocity-deficit profiles. Porté-Agel et al. [4] offer a comprehensive review of the fluid dynamics governing wind-turbine and wind-farm wakes.

Figure 1. Schematic figure showing the flow features resulting from the interactions of a wind turbine and ABL.

The near wake typically spans roughly two to four rotor diameters; helical vortex filaments destabilise via mutual inductance and break down into turbulence [4]. In the far wake, recovery is controlled by turbulent entrainment of higher-momentum air from above and laterally, making ambient turbulence intensity and thermal stratification decisive [2,5,6]. This dependence is central when interpreting any wake-control concept: mean flow recovery inferred from steady simulations must be viewed as a partial indicator unless the concept is subsequently addressed in higher-fidelity, unsteady, turbulence-resolving frameworks, such as through Large Eddy Simulation (LES).

Wake meandering, driven by large Atmospheric Boundary Layer (ABL) eddies that can exceed the rotor diameter, introduces substantial unsteady loading and is a primary mechanism for downstream fatigue amplification [4]. These single-wake dynamics scale nonlinearly when turbines are assembled into arrays, forming an Internal Boundary Layer (IBL) that grows with fetch. At high installed capacity densities of around 6 MW/km², such deep-array conditions can be severe: for very large arrays, the most heavily waked interior turbines may generate only 32–35 per cent of the power of turbines exposed to freestream conditions [2].

2.2. Far-Field Cluster Wakes and Regional Wind Shadows

At farm and cluster scales, as defined in Figure 2, empirical [7,8,9,10] and numerical [11,12,13,14,15] evidence demonstrates that offshore wakes extend far beyond individual project footprints, undermining the assumption that spacing within a lease area alone is sufficient to protect yield. Airborne observations in the North Sea show that, under stable stratification, cluster wakes can exceed 50 km before the mean wind-speed recovers to 95 per cent of the freestream speed, consistent with Akhtar et al. [16] and quantified directly by Cañadillas et al. [7]. The same campaigns report substantially shorter recovery distances in neutral and unstable conditions, typically on the order of 15 km, with cases spanning from 1.5 to 24 km depending on event and definition [7]. More recent airborne datasets in the German Bight by zum Berge et al. [8] reveal stable-regime wake lengths exceeding 70 km, with specific flight evidence indicating 60 to 65 km deficits between major cluster groupings. LES studies by Maas and Raasch [12] imply that as North Sea-style build-outs approach “very large” farm dimensions, roughly 100 km long, wakes may extend beyond 100 km in shallow boundary layers, with wake recovery increasingly influenced by farm-scale pressure–gradient dynamics rather than turbulent momentum flux alone. Similar large-area shadow effects are predicted for emerging Atlantic-scale developments: WRF simulations for U.S. East Coast lease areas suggest a region with a mean velocity deficit exceeding 5 per cent extending to 2.82 times the footprint of the lease areas [11].

Figure 2. Schematic diagram of interactions at different length scales: intra-farm (blue), intra-cluster (green), and inter-cluster (red).

2.3. Production and Economic Impact of Long-Range Wakes

The production consequences of such long-range wakes are now sufficiently large to be treated as a first-order design and planning variable. In a 2030 North Sea scenario, inter-farm wakes are projected to reduce the Annual Energy Production (AEP) of 25 out of 69 operational wind farms by over 5 per cent, with 13 farms experiencing losses exceeding 10 per cent, and some German EEZ assets potentially surpassing 18 per cent loss [13]. These annual averages can mask operationally critical intermittency: the planned Princess Elisabeth zone is expected to reduce output of a neighbouring cluster by more than 15 per cent during roughly one out of every five high-production 24 h periods [17]. Historical-to-future comparisons by Warder and Piggot [14] further show density-driven escalation of long-range impacts: in the German Bight, inter-cluster wake effects rise from zero in 2010 to approximately 2.5 per cent by the 2027 build-out stage, while intra-cluster effects reach 4.3 per cent. At shorter inter-farm separations, SCADA-based evidence by Fischereit et al. [15] similarly confirms measurable yield suppression: the upstream Nysted farm reduced the median wind resource at Rødsand II by 3 to 4 per cent, translating to a 6 to 12 per cent power reduction, and even as far as 25 km downwind. These observations align with wind-farm-internal deep-array behaviour, in which interior turbines can generate between 15 and 20 per cent less than edge turbines [4,18].

A consistent message across these studies is that the dependence on wind speed is operationally consequential: inter-farm losses peak just below rated conditions, implying that long-range wakes disproportionately affect the wind-speed regimes that drive high-capacity factor and revenue.

2.4. Modelling Uncertainty and Parametrisation Sensitivity

Notably, the magnitude of these effects remains model-sensitive, challenging today’s planning and finance assumptions. Rapid engineering models can underpredict far-reaching farm-to-farm deficits relative to RANS and mesoscale frameworks such as the Weather Research and Forecasting model (WRF) [15], consistent with the observational inferences of Cañadillas et al. [7]. Pryor and Barthelmie [2] discuss that even within WRF, the choice of wind-farm parameterisation materially changes projected outcomes: the Fitch scheme yields substantially larger wake extents and higher wake-induced power losses (27–37%) than the Explicit Wake Parameterisation (EWP) (11–19%) for comparable U.S. East Coast layouts, with wind shadows predicted to cover areas 1.18–1.38 times larger than EWP. This divergence is no longer an academic nuance; it directly alters lease valuation, risk pricing, and engineering design margins.

These aerodynamic realities now translate into measurable techno-economic constraints and a discernible diminishing return with capacity density. Increasing turbine density can raise absolute areal power output while sharply reducing the capacity factor of interior installations; an efficiency collapse which is consistent with deep-array behaviour [2,19]. The Levelized Cost of Energy (LCOE) consequence of such densification is evidenced by optimisation studies by Barnabei et al. [19], showing a clear knee-point: in deep-water Mediterranean layouts, a minimum LCOE of 105.72 €/MWh aligns with low wake-loss, while higher-loss cases yield a higher LCOE of 128.67 €/MWh, despite reduced turbine counts. More broadly, wake losses can reduce AEP by around 15 per cent relative to idealised no-wake baselines [19], and a 5 per cent reduction in freestream wind speed can translate to a 7 per cent AEP reduction [11]. High-density spacing in Thomas et al. [20] further illustrates the potential magnitude of the penalty, with the LCOE rising from 107 €/MWh in an idealised reference case to nearly 140 €/MWh under severe wake conditions. Taken together with global-scale value estimates suggesting that a 1 per cent capacity-factor increase across a 650 GW fleet corresponds to more than $1.1 B/year [21]. These results position inter-farm wakes as a critical lever in national decarbonisation economics.

2.5. Planning and Governance Gaps

Finally, these findings expose a planning and governance gap. Despite growing empirical and model-based evidence of long-range wind theft, regulatory frameworks in Europe do not yet systematically account for cross-border wake externalities or mandate buffering in high-density areas [3,9]. The Belgian concession zone exemplifies the prevailing approach: high-density co-location to share infrastructure, at the expense of increased intra- and inter-farm wake coupling [17]. This global absence of actionable planning tools, such as publicly available Geographic Information System (GIS) layers that quantify wind-farm wind shadows [11], limits the integration of wake constraints into forward lease design and cross-operator coordination [3,11].

3. Limits of Current Wake-Mitigation Approaches

Current evidence that multi-farm wakes can persist for tens of kilometres under stable atmospheric stratification and impose multi-per cent penalties on AEP at cluster scales implies that the key mitigation question is no longer simply whether established strategies work in idealised single-farm settings, but whether they remain effective when capacity density, atmospheric constraints, and multi-owner governance dominate the design space.

Recent advances include cooperative wind-farm control informed by LiDAR and data assimilation; dynamic wake-mixing strategies through the modulation of rotor positioning aimed at accelerating recovery at the expense of added unsteadiness; improved deep-array and farm-to-farm wake modelling; and alternative blade-based devices and interventions. However, emerging evidence suggests that largely local, farm-bounded, and quasi-static approaches encounter fundamental limits in high-density offshore wind clusters.

3.1. Layout Optimisation

Layout optimisation is the first constraint to saturate. Even when optimisation algorithms identify lower-loss solutions, the physical boundary of lease areas and the economic pressure to share offshore infrastructure leave limited freedom to increase spacing. This is clearly reflected in spacing-dependent loss trends: at an inter-turbine spacing of roughly 4D, wake losses can reach 30 per cent, whereas 12D can reduce losses toward 10 per cent, yet optimisation studies of separation distance typically identify a working range of 5–10D for in-line Horizontal-Axis Wind Turbines (HAWTs), beyond which further spacing would not result in lower costs of generation due to the increasing inter-array cabling requirements [4,20]. Optimisation is also sensitive to objective choice and policy context: Barnabei et al. [19] discuss how attempts to minimise visual impact or concentrate assets for shared infrastructure can drive highly compact row-like arrangements that increase rotor–wake interactions and reduce AEP by 15 per cent relative to idealised layouts. More generally, the multi-parameter layout and hub-height optimisation studies reviewed by Awan and Sarosh [22] find that micro-siting can recover a few per cent of AEP at the farm-scale, but do not remove the underlying constraint that, in capacity-saturated areas, turbines and farms must ultimately operate in each other’s wakes.

3.2. Static Yaw Misalignment

Static yaw misalignment offers meaningful gains in aligned flow but is intrinsically fragile when applied at cluster density and under the uncertainties of realistic offshore spatial planning [5,17,23]. The strategy depends on accurate, stable wind-direction estimates; small directional errors can negate benefits or even worsen downstream impacts, and performance degrades relative to active-mixing approaches when wind-direction uncertainty is only a few degrees [5]. The upstream penalty is non-trivial: yawing incurs a local cost in power that approximately follows a cosine-cubed relationship, with Frederik et al. [23] reporting upstream losses on the order of 15.9 per cent in representative cases. This trade is not guaranteed to close in low-veer or medium-to-high turbulence conditions, where natural recovery already limits marginal steering benefits and where wake deflection can displace turbulent transport into the wake core, reducing re-energisation efficiency [5,23]. Operationally, persistent high yaw angles in dense zones may amplify blade, tower, and yaw-bearing fatigue, raising lifetime and availability concerns [6,17,21]. At the cluster scale, steering also exposes a coordination problem: greedy farm-level optimisation underperforms cooperative, cluster-level control, yet multi-owner data-sharing and aligned incentives remain limited [3,17].

3.3. Active Induction Control (AIC)

Active Induction Control (AIC) and related derating strategies face a different limitation: they often yield only marginal net gains in high-fidelity studies and can be structurally or aerodynamically counterproductive under dense alignment [21]. Reducing upstream thrust may leave recoverable kinetic energy concentrated along wake edges rather than replenishing the wake centreline, limiting the benefit for directly aligned downstream rotors [21,24]. Reduced induction can also weaken rotor-added turbulence, slowing recovery and offsetting the benefit of reduced deficit [21]. Comparative offshore studies suggest AIC is often inferior to yaw control across most wind directions in realistic layouts [25]. More aggressive dynamic mixing concepts discussed by Brown et al. [5] can improve recovery but introduce grid-power fluctuations, actuator wear, and potentially prohibitive increases in fatigue loads, with sensitivity to wake skewing under realistic veer conditions [5,21,23].

3.4. Rotor- and Device-Based Wake Manipulation

Beyond layout and turbine-level control, a broad set of rotor- and device-based strategies has been proposed to mitigate wake deficits by modifying the near-wake and blade aerodynamics. Awan and Sarosh [22] classify these into active methods, such as adding surfaces or actuators, such as microplates near the leading edge, cross-axis rotors combining HAWT and VAWT elements, trailing-edge flaps, dual-rotor configurations, vortex generators, and tip injection; and passive methods, including blade sweep, tailored rotor design, hub-height staggering, and optimised separation distance. Microplates, trailing-edge flaps, and leading-edge protuberances can delay stall and locally re-energise boundary layers, yielding modest increases in the lift and power coefficient and, in some cases, 1–7 per cent gains in AEP at the single-turbine or small-array scale [26,27,28,29]. Dual-rotor and cross-axis concepts extract additional energy from the near-wake swirl and from vertical components of the flow, achieving substantial power increases in laboratory tests but at the cost of sharply increased static and dynamic loads, higher manufacturing complexity, and uncertain maintainability offshore [22,30]. Tip injection and blade sweep can reshape tip-vortex structures and alter thrust and power characteristics, but their beneficial regimes are narrow [31,32]. Hub-height optimisation can recover a few per cent of farm-level AEP by staggering rotor planes in shear [33] but does not remove long-range deficits once the entire boundary layer is populated. Overall, these techniques confirm that local wake characteristics can be modified in many ways. Yet, they remain fundamentally near-wake interventions, largely validated in wind-tunnel or small-farm contexts, and have not been shown to address the persistence of cluster-scale wakes in stable offshore boundary layers.

Taken together, these limitations indicate that the dominant mitigation toolbox is reaching diminishing returns under high-density offshore clustering. Local optimisation and quasi-static control can reduce intra-farm penalties, but do not fundamentally address the physics of suppressed vertical replenishment in stable offshore ABLs and the resulting persistence of long-range deficits [7,10,12]. Nor do they resolve emerging policy and planning gaps where cross-border or inter-operator wake externalities remain weakly regulated or unaccounted for in lease design [1,3]. The proliferation of rotor- and device-based wake-reduction concepts further confirms that incremental improvements at the turbine-scale are possible but do not unlock new momentum pathways; they operate largely within the existing kinetic-energy budget of the farm-layer flow. This combination of physical persistence, economic densification, and governance constraints provides a clear rationale to broaden the mitigation paradigm beyond local steering and spacing toward novel wake-management concepts capable of actively accelerating wake re-energisation at scale and under stable stratification, including approaches that leverage higher-altitude momentum pathways.

4. Kites for High-Altitude Momentum Utilisation

4.1. Kite System Characteristics and Operation

The limitations of the existing wake-mitigation strategies in dense offshore clusters, discussed above, point to a fundamental constraint: the lack of controllable access to higher-altitude momentum within the stable, low-turbulence boundary layers that dominate offshore sites. The momentum availability of conventional HAWTs is effectively fixed to a narrow vertical band constrained by the rotor diameter and hub height. In contrast, kites and parafoils are intrinsically three-dimensional airborne devices that can reach high altitudes, enabling the exploitation of more favourable wind conditions that are inaccessible to fixed towers.

In practical implementations, these systems use soft-fabric parafoils made from low-permeability textiles and high-strength steel or Ultra-High-Molecular-Weight Polyethylene (UHMWPE) tethers flown either statically or in dynamic crosswind patterns (figure-of-eight or sine manoeuvres) that amplify the apparent wind speed and tether force by more than an order of magnitude relative to the freestream [34,35,36].

4.2. Vertical Wind Resource and Operational Adaptability

Resource and optimisation studies discussed by Bechtle et al. [37] and Pereira et al. [38] provide clear evidence that kites can systematically exploit this vertical structure. Analyses of ERA5 data for Western and Central Europe show that allowing a variable harvesting height up to 500 m can more than double the wind-power density available 95 per cent of the time relative to a fixed 100 m height, with particularly strong gains for coastal and offshore regions in the UK, the Netherlands, and northern Germany [37,38]. A common use of kites, Airborne Wind Energy Systems (AWES) operate at heights below 400 m, with a modal operation around 200 m [39]. In this implementation, kites exploit the smooth, favourable shear in the 200–400 m layer.

Operationally, this vertical adaptability is realised through coupled control of tether length and elevation angle [40]. Optimisation frameworks driven by realistic WRF wind fields show that AWES can adjust the deployed tether length and reduce the elevation angle to reach higher-energy layers, systematically shifting the operating height to coincide with locally optimal wind speeds [39,40]. Fagiano et al. [41] notes that safety-related minimum heights are set by wingspan and airspace constraints, and ceiling heights are limited by tether drag and mass rather than by aerodynamics.

While resource analyses often assume vertical adjustment on hourly timescales [37], winch dynamics and control authority govern the practical response: full pumping cycles that traverse a large fraction of the vertical range occur on the order of 1–3 min, and radial motion is constrained by maximum reel-out speeds and generator torque [39,42]. Rigid-wing systems with onboard control surfaces exhibit minimal mechanical lag, whereas soft, ground-steered parafoils are more affected by tether elasticity [38].

4.3. Boundary-Layer Interaction and Momentum-Exchange Potential

Beyond access to stronger winds, kites also act as active momentum-exchange devices within the boundary layer. The LES of AWES operation in a neutral, pressure-driven boundary layer by Haas et al. [42] shows that crosswind flight and cyclic operation can amplify local TKE levels by up to an order of magnitude relative to ambient levels, indicating a strong capacity to disturb and mix the flow. Zgraggen et al. [43] demonstrate that kites can be stabilised in quasi-static angular positions, suggesting a continuum between power-maximising crosswind flight and deliberately parked or slowly manoeuvring states that primarily inject turbulence and redistribute momentum rather than maximise electrical output. Bechtle et al. [37] explicitly note that the effect of large-scale airborne energy extraction on vertical energy exchange and mixing remains uncertain and a subject of scientific debate.

Taken together, the AWES literature demonstrates three properties that are highly relevant to next-generation wake-management: (i) kites can inhabit and exploit a vertical band where offshore kinetic energy flux is significantly higher and more persistent than at turbine hub-height [37,39]; (ii) they can be repositioned within this band on operational timescales via tether and elevation control, subject to realistic winch and structural limits [40,42]; and (iii) their operation can substantially modify local turbulence and momentum transport [42], even though the detailed interaction with stable boundary layers is not yet resolved [37]. These characteristics suggest that kite and parafoil systems are not only alternative generators, but potential high-altitude momentum managers that could, if appropriately integrated with offshore HAWTs, be repurposed from purely harvesting devices into tools for targeted wake re-energisation in precisely those regimes where conventional passive strategies are least effective.

5. Integrating Parafoils with Turbines for Wake Re-Energisation

A novel application is thus proposed herein: using inflatable parafoils to inject momentum into the wake behind a wind turbine, thereby accelerating recovery and mitigating wake losses. The literature on this concept is sparse. The LES of AWES farms indicates that kites generate tip vortices and wake structures distinct from turbine wakes [42], but there is no quantitative assessment of how their induced downwash entrains high-momentum air into a turbine wake.

5.1. The Existing Literature on External Devices for Wake Manipulation

Early numerical work on external devices for wake manipulation has focused on deliberately introducing additional lifting surfaces in the rotor vicinity. Ploumakis and Bierbooms [44] investigate a small array of four aligned turbines embedded in an LES of a turbulent ABL and then intersperse tethered kites in steady flight between rotors. The kites are represented as distributed forces equal and opposite to the vector sum of their aerodynamic lift and drag, allowing the quantification of how crosswind-flying parafoils modify the Reynolds stresses, vertical momentum flux, and mean velocity on the farm. It is shown that appropriately positioned kites enhance the vertical transport of high-momentum air and partially replenish downstream turbine wakes beyond what is achievable by turbine spacing alone [44].

Bader et al. [45] pursue a complementary idea using stationary lifting surfaces mounted in the vicinity of conventional HAWTs. Large wake modifiers with a NACA 2412 cross-section are positioned close to the rotors in Reynolds-Averaged-Navier–Stokes (RANS)-based simulations and demonstrate that the forces resulting from flow redirection and the associated trailing vortices can substantially accelerate wake recovery in downstream rows [45]. Their analysis shows that these devices re-route wake recovery from being dominated by slow turbulent diffusion to a more efficient vertical advection process, with gains in downstream power of as much as 80 per cent, albeit in an idealised configuration where the lifting devices are suspended in the flow without an explicit structural-integration concept.

Building on this emergent line of wake mitigation using external flow control, Nash et al. [46] introduce an explicitly tertiary device: a large, passive external wake deflector in the form of a stationary vane. Scaled wind-tunnel experiments at 5D spacing show that the vane cuts the velocity deficit at the downstream rotor plane by about one-third and boosts the inferred downstream power by more than 15 per cent, without penalising the upstream turbine or increasing its structural loads, thereby illustrating the potential of non-turbine tertiary devices to achieve stronger wake diversion than conventional control alone, while also highlighting the substantial implementation challenges at full scale.

5.2. Proposed Turbine-Integrated Parafoil Momentum Injector

Existing concepts that use external devices for wake mitigation demonstrate that integrating lifting surfaces has the potential to manipulate wakes but fall short in implementation.

Figure 3 presents a novel concept where a lighter-than-air inflatable parafoil is tethered to a turbine, the wake effects of which it is intended to mitigate. What distinguishes the present concept from existing ones is that it is explicitly conceived as a high-altitude momentum injector, structurally and aerodynamically integrated with a specific turbine. In contrast, previous concepts are not envisaged as turbine-integrated devices. Ploumakis and Bierbooms [44] introduce kites between turbines but treat them primarily as additional energy extractors, represented via body-force terms, demonstrating enhanced farm-level kinetic-energy entrainment, without targeting the recovery of an individual rotor wake or exploiting lighter-than-air buoyancy. The stationary lifting surfaces [45] and passive vane concepts [46] are structurally and spatially idealised and are not necessarily designed to exploit the stronger, more persistent shear layers from aloft. In all three cases, the objective is predominantly to deflect the wake or to enhance generic TKE rather than to target the vertical re-energisation of the turbine wake core deliberately.

Figure 3. Illustration of the proposed setup.

By contrast, the parakite concept proposed here is a lighter-than-air, inflatable parafoil moored directly to the turbine substructure via a horizontal strut, designed to operate in the 200–400 m band and to generate controlled downwash and strong tip-vortex systems that entrain high-momentum air from aloft into the HAWT wake, while simultaneously providing a beneficial counter-moment at the tower base. This dual role, structural load alleviation plus wake re-energisation, has no analogue in the existing literature [44,45,46], where structural effects are largely ignored, and devices do not tap the vertically mobile high-altitude resource identified in the AWES studies [37,38,39].

The present concept is explicitly framed around the regimes where conventional wake steering and induction control are weakest, hence, the parafoil is not merely redirecting existing wake momentum, but is intended to promote entrainment and potentially inject additional momentum from higher altitudes into the farm-layer flow, using the body forces and tip vortices of a controllable, buoyant lifting surface that can remain deployed even in low-wind conditions. In that sense, it represents a qualitatively different class of wake-management tool: a turbine-centric, high-altitude momentum pump.

Preliminary investigations into this approach are being undertaken at the University of Malta under the ParaWIND project (Project Reference: REP-2024-003). Steady RANS simulations of an Actuator Disc (AD) with the integration of the rigid, lighter-than-air inflatable parafoil having an elliptical planform, as displayed in Figure 4, have shown promising results. Figure 5 presents illustrative steady RANS–AD mean-flow fields for a baseline case and a configuration with the parafoil placed 0.8D above hub height and 0.25D downstream. Figure 6 reports corresponding mean axial-velocity profiles downstream. In this limited modelling setting, the parafoil-induced downwash deflects and redistributes the mean wake, consistent with the proposed entrainment mechanism; however, the magnitude and net benefit under realistic offshore ABL turbulence, stability, and array operation remain to be established through higher-fidelity simulation and experiment.

Figure 4. The wing used in preliminary studies investigating the proposed concept.

Figure 5. Illustrative steady RANS actuator-disc results: contour of mean axial-velocity in the x–y plane for (a) single AD baseline; (b) single AD with an inflatable parafoil at 0.25D downstream.

Figure 6. Illustrative steady RANS actuator-disc results: mean axial-velocity profiles compared with the undisturbed inflow shear profile.

The supporting numerical evidence for the turbine-integrated parafoil mechanism presented herein is presently limited to steady Reynolds-Averaged-Navier–Stokes (RANS) simulations using an actuator-disc (AD) representation of the turbine and a rigid parafoil geometry. These simulations are used to probe first-order mean-flow responses, such as downwash-induced wake deflection and mean velocity recovery, and do not resolve unsteady wake meandering, coherent-structure breakdown, or stability-dependent turbulence production in the offshore Atmospheric Boundary Layer (ABL). Accordingly, the results are presented as mechanism-illustrative and are not asserted as predictive of full-scale AEP or fatigue outcomes.

5.3. Comparative Context and Complementarity

The mitigation strategies reviewed in Section 3.1, Section 3.2, Section 3.3 and Section 3.4 share a commonality: they operate predominantly within the farm-layer kinetic-energy budget accessible at or near rotor height, either by redistributing wake impacts or by seeking incremental recovery gains through local mixing, induction modulation, or near-wake shaping.

In contrast, the proposed turbine-integrated parafoil concept is positioned as a distinct class of wake-management rather than merely rearranging momentum deficits within the farm layer; it is intended to promote vertical entrainment by creating a controllable downwash and a coherent vortex system capable of coupling the wake to higher-altitude momentum pathways. This distinction matters especially in the regimes that motivate this perspective-neutral-to-stable offshore boundary layers—where suppressed vertical exchange can make incremental, rotor-bounded interventions increasingly marginal. At the same time, the parafoil approach introduces a new set of considerations and therefore cannot be treated as a drop-in substitute for established measures; its value proposition must be assessed on the basis of net energy, loads, and availability under stability-resolving conditions.

Importantly, a high-altitude momentum-injection concept is not inherently a direct alternative. A plausible development pathway is that parafoils are combined with existing wake-mitigation measures complementarily: moderate wake steering could reduce direct alignment losses while the parafoil accelerates recovery of the wake core via enhanced vertical transport; and AIC could be scheduled during parafoil operation to avoid potentially adverse turbulence–fatigue trade-offs. Layout optimisation could explicitly incorporate tether corridors and exclusion zones to preserve controllability. At the cluster scale, where governance and inter-owner externalities constrain purely local optimisation, coordinated operation may also be envisaged in which parafoils are deployed opportunistically during stable, wake-dominant periods to reduce long-range penalties. These combinations frame the parafoil concept as a complementary augmentation method with potential to be co-optimised with established tools rather than as a replacement for mature mitigation strategies.

5.4. Roadmap for Concept Development and Deployment

Progressing the novel concept presented in Figure 3 into a deployable technology requires addressing challenges across several tightly coupled fronts: physics, control, structural integration, operations, and regulation/finance.

5.4.1. Aerodynamics of Wind Turbine Arrays Integrating Kites

The first requirement is a quantitative understanding of how inflatable, lighter-than-air parafoils, mounted above and between turbines, modify the flow. Ploumakis and Bierbooms [44] show that kites can enhance farm-level kinetic-energy entrainment, and Haas et al. [42] report up to order-of-magnitude increases in TKE for AWES compared to a conventional HAWT setup, but there is still a lack of CFD-based evidence and resolved statistics on (i) downwash strength and footprint from a buoyant parafoil at 200–400 m; (ii) how that downwash couples to a specific downstream rotor wake; and (iii) how the system performs under stable stratification, where cluster wakes are longest.

In multi-row deployment, turbine wakes superpose nonlinearly under stability-dependent entrainment limits, and parafoil-induced downwash fields may thus interact constructively or destructively depending on spanwise spacing, elevation, and control set-points. Additional considerations include kite-to-kite aerodynamic interference, tether-corridor management, and the possibility that enhanced mixing beneficial to recovery may also elevate turbulence-driven fatigue in downstream machines. Consequently, multi-unit assessment must quantify not only mean recovery but also changes in turbulence spectra, wake meandering statistics, and load-equivalent fatigue metrics under representative offshore stability classes.

Future work must therefore extend modelling RANS/LES frameworks to include both a realistic turbine and an explicit parafoil model across neutral and stable ABLs, and investigate array-level behaviour for multiple-turbine arrays equipped with parafoils, inter-device interference, and wake superposition across rows. The key outputs are not only reductions in the mean-velocity deficit but also changes in vertical momentum flux, wake meandering, added turbulence scales, and sensitivity to kite position, including spanwise offset, height, and elevation angle. The concept can only be credibly optimised after quantifying the additional momentum that can be injected and the resulting penalties in turbulence and drag.

5.4.2. Control, Stability, and Fail-Safe Behaviour

Unlike fixed tertiary vanes [45,46], an inflatable parafoil is a dynamically unstable, flexible lifting body attached by a long tether with its own associated considerations. Pereira et al. [38], Sommerfeld et al. [39], and Vermillion et al. [47] highlight that practical AWES altitude changes and gust rejection are limited by winch dynamics, tether mass/drag, and control authority. The proposed concept requires controllers that can (i) hold quasi-static states to provide momentum injection at a prescribed elevation/azimuth, (ii) transition to safe, de-powered, or parked configurations when the turbine cuts out or inclement weather is present, and (iii) coordinate with turbine yaw/pitch and farm-level control.

The presence of strong veer complicates this: Brown et al. [5] and Frederik et al. [23] show how veer can severely distort wake-control strategies, and similar distortions will affect parafoil effects and the wake footprint. Research priorities here include high-fidelity 6-DoF kite–tether dynamics models, robust control synthesis, including fault-tolerant control methods, and load-aware supervisory logic that enforces safe envelopes in extreme metocean events, loss-of-tension, or partial inflation scenarios.

5.4.3. Structural and System Integration

The explicitly turbine-integrated approach is both an opportunity and a challenge. The horizontal cross-arm or pylon supporting the tether anchor must transmit significant, time-varying forces and moments into the tower without compromising fatigue life. Tether mass and drag add overturning moments and dynamic excitation. Future work must therefore embed parafoil loads into aero–hydro–servo–elastic models of monopile or floating substructures, generate new load cases for aligned and oblique flows, extreme gusts, and partial kite-control loss, and run full design-loop iterations using contemporary standards. A realistic concept must also address maintenance access, lightning protection, icing, and potential interference with cranes, CTVs, and SOVs during routine turbine O&M.

5.4.4. Operations, Safety, and Regulatory Considerations

A kite-equipped farm has more complex operational requirements than current offshore arrays. Parafoils and tethers at several hundred metres might interact with radar and aviation and maintenance corridors; they encroach on airspace that is not necessarily considered in offshore wind permitting. Maritime navigation and fishing must be considered in failure modes such as a deflated wing or slack tether in the water. Existing AWES studies mostly assume stand-alone devices; integrating them into large farms implies new safety cases, emergency recovery procedures, and exclusion zones. On the regulatory side, Finserås et al. [3] already highlight gaps for cross-border wake effects; the proposal adds another layer of novelty that is not covered by current planning rules or certification schemes. Early engagement with regulators, aviation authorities, classification societies, and insurers will be essential to define acceptable operating envelopes and certification pathways.

5.4.5. Techno-Economic and Planning Integration

While the net benefit is not yet established, it is useful to anchor the concept against order-of-magnitude planning benchmarks. Turbine-power scales with $U^3$. Thus, even modest increases in the effective inflow to waked downstream rotors can translate into disproportionate power gains during below-rated operation. Conversely, any benefit must exceed penalties associated with added drag, turbulence-driven fatigue, and availability impacts from deployment constraints.

Even a technically successful concept must clear a techno-economic bar. The additional CAPEX of parafoils, winches, control systems, and support structures, plus O&M costs and potential availability losses, must be justified by robust AEP gains. This demands scenario-based LCOE analyses of the sort already used for density optimisation [11,19,20,48], but extended to include parafoil hardware, failure rates, and insurance premiums. It also suggests new planning tools: if the proposed concept or similar systems can materially reduce long-range wake penalties, they should be explicitly represented in lease valuation and cross-border planning models, rather than treated as an after-the-fact retrofit.

A conservative early benchmark for feasibility assessment is, therefore, (i) demonstrate repeatable reductions in wake deficit and/or accelerated recovery under stable-to-neutral ABL conditions at the single-unit level; (ii) show that the resulting downstream energy gain plausibly survives array superposition; and (iii) close an LCOE case in which added CAPEX/OPEX and risk premiums are offset by net AEP and/or reduced wake-externality losses at cluster scale. These benchmarks define what the proposed staged programme must quantify before bankable performance claims are made.

6. Conclusions

The rapid densification of offshore wind is exposing fundamental limits in current wake-mitigation strategies. Observations and high-fidelity modelling show that turbine-, farm-, and cluster-scale wakes can persist for tens of kilometres under stable conditions, causing multi-per cent AEP losses and higher LCOE. Existing approaches, mainly layout optimisation, yaw misalignment, and induction control, remain largely local and constrained by lease boundaries and suppressed vertical momentum-exchange. This paper proposes a complementary approach: turbine-integrated parafoils acting as high-altitude momentum injectors, entraining stronger, less wake-depleted winds from 200 to 400 m to actively re-energise wakes via controlled downwash and vortex-driven vertical transport.

The present paper has articulated the mechanism and integration logic at the level of a single turbine–single parafoil unit and has presented only steady, actuator-disc RANS computed illustrations as evidence of mean-flow response. Demonstrating farm-scale value requires a credible development roadmap for parafoil integration in offshore wind farms should proceed in stages, with progressively higher TRL and spatial scale:

• Conceptual and numerical phase (TRL 1–2):

  ○

  Parametric RANS/LES studies exploring single-turbine and parafoil configurations in neutral and stable ABLs, scanning kite geometry, height, and offset.

  ○

  Reduced-order models to capture key trends such as tether dynamics.

• Laboratory phase (TRL 2–3):

  ○

  Wind-tunnel tests with scaled rigid or semi-rigid wings integrated with actuator-disc or scaled turbine models, to validate predictions of downwash, tip-vortex structure, and wake-core re-energisation.

  ○

  Wave-tank tests for floating support structures with surrogate kite loads to study coupled platform motions and fatigue.

• Prototype phase (TRL 3–5):

  ○

  A single, instrumented offshore turbine equipped with a simplified parafoil system (limited envelope, conservative loads) and dense wake measurements (nacelle and scanning LiDAR, met masts, possibly drone-based sampling).

  ○

  Controlled campaigns in different stability regimes to quantify actual wake-recovery gains, additional turbulence, availability impact, and structural load effects.

• Farm-scale pilots (TRL 5–7):

  ○

  Deployment of parafoils on a row or sub-cluster of turbines, combined with farm-level control (coordinated yaw, curtailment, and kite positioning), to assess cluster-scale impacts and interactions with neighbouring assets.

Throughout, it will be crucial to maintain closure between models and data: The literature has provided starting points, but new studies in the specific parafoil–HAWT configuration are needed to constrain turbulence closures, stratification effects, and structural load models.