Applied Energy 347 (2023) 121410

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Design of sustainable offshore hybrid energy systems for improved wave

energy dispatchability
Mariasole Cipolletta , Anna Crivellari , Valeria Casson Moreno , Valerio Cozzani *
Laboratory of Industrial Safety and Environmental Sustainability - DICAM, University of Bologna, Via Terracini 28, Bologna 40131, Italy

H I G H L I G H T S

• A methodology for sustainable design of offshore hybrid energy systems is developed.

• Wave energy dispatchability issues are considered.
• Stranded gas in end-of-life reservoirs may be used to foster energy transition.
• Wave energy exploitation enhanced by offshore stranded gas-based back-up systems.
• Test cases in the North Sea and Adriatic Sea were carried out.

A R T I C L E I N F O A B S T R A C T

Keywords: Wave energy is a renewable energy source having a highly exploitable potential in several locations worldwide.
Offshore hybrid energy systems In the framework of energy transition, the exploitation of wave energy combined with end-of-life offshore
Sustainable design stranded gas reservoirs may lead to two positive impacts: the stabilization of the energy supplied to the grid and a
Offshore renewable energy sources
better penetration of renewable energy in areas where the grid is not able to compensate the fluctuations
Wave energy
Grid integration
associated to renewable energy production. Moreover, in order to guarantee the dispatched energy schedule,
wave energy needs to be coupled with back-up systems aimed at valley filling. In the present study, an innovative
approach to the conceptual design of hybrid energy systems based on wave energy is developed, entailing an
operation strategy that complies with the dispatching needs of grid-connected generation systems. The proba­
bility of correct dispatching that the producer assures to the Transmission System Operator is used as a parameter
to optimize the design of a Gas to Power back-up system used for valley filling. The approach supports the
preliminary design of offshore hybrid energy systems based on wave energy, starting from historical wave data
up to the definition of an optimal back-up system valorizing residual reservoir fuels and its operation strategy.
The proposed design is evaluated through a Multi-Criteria Decision Analysis, including the technological, eco­
nomic, environmental and safety aspects, which allows the assessment of the overall sustainability performance
of the hybrid system, considering the fluctuations associated to wave power generation during a typical oper­
ation period. The methodology was applied to two test-cases in different offshore operating theaters (North and
Adriatic seas), in order to test its potentiality. The results highlighted that, in both sites, similar design choices
are suggested for the hybrid system. However, the annual energy production resulted 6.5 times higher in the
North Sea test-case. The low energy generation in the Adriatic Sea test site caused a levelized cost of energy of
3960 EUR/MWh, much higher than the value obtained for the North Sea case (610 EUR/MWh). In both cases, the
gas turbine park impacts negatively on the cost of energy production, but is critical in meeting the design value of
the probability of correct dispatching.

profitably exploit Renewable Energy Sources (RESs). In 2019, 11.6% of

primary energy and 26.7% of worldwide generated electricity were
1. Introduction obtained from these unlimited sources [1]. The sharp increase in
renewable energy generation after 2001 [2], supported by public
In recent years, uncountable efforts were made to efficiently and

* Corresponding author.
E-mail address: valerio.cozzani@unibo.it (V. Cozzani).

https://doi.org/10.1016/j.apenergy.2023.121410
Received 30 December 2022; Received in revised form 1 March 2023; Accepted 6 June 2023
Available online 25 June 2023
0306-2619/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Nomenclature LCOE Levelized Cost Of Energy

LEE Levelized Energy Efficiency
List of Acronyms LGHG Levelized Greenhouse Gases emissions
AEP Annual Energy Production LHI Levelized inherent Hazard Index
AHP Analytic Hierarchy Process LVOE Levelized Value Of Energy
ASI Aggregated Sustainability Index MCDA Multi-Criteria Decision Analysis
CAPEX CAPital Expenditures NG Natural Gas
CDF Cumulative Distribution Function OPEX OPerative Expenditures
EFLH Equivalent Full Load Hours PTO Power-take-off
G2P Gas To Power RE(s) Renewable energy(ies)
GHG Greenhouse Gas RES(s) Renewable energy source(s)
GT(s) Gas Turbine(s) TSO Transmission System Operator
GTP Gas Turbine Park WD Wave Dragon
HVAC High Voltage Alternating Current (HVAC) WEC(s) Wave Energy Converter(s)
KPI(s) Key Performance Indicator(s) WF(s) Wave Farm(s)

investments in the sector [3], is due to the growing awareness that the neglected, thus leading to the overestimation of the potentialities of
worldwide exploitation of RESs is a key requirement considering several locations with a high average wave resource power [17]. Also, wave
issues: the unequal distribution of fossil resources, their depletion, and, energy is challenged by its aleatory variability, since marine and meteo-
mostly, the effects of climate changes caused by Greenhouse Gas (GHG) climatic conditions determine the real-time output of WECs [18].
emissions. These factors, however, strongly claim the need for an even In order to face the variability of power production from RESs, the
higher penetration of cleaner energy sources and for the development of Transmission System Operator (TSO) requires the company operating
more efficient systems for RESs exploitation in production and trans­ the RESs converters to define in advance a daily dispatching power plan
formation processes [4]. Thus, the widespread and improved use of [19]. To reduce the risk related to the randomness of RESs and to foster
Renewable Energy (RE) can make great strides towards the sustain­ the improvement of the dispatching forecast, a common principle is to
ability of anthropic activities from the social, economic and environ­ impose to the energy suppliers a probability of correct dispatching (i.e.
mental perspectives. to correctly produce the scheduled power), thus penalizing them when
To this purpose, seas and oceans are unlimited sources of wave en­ the produced power diverges for an error higher than the allowable one
ergy, tidal energy, currents, salinity gradients and ocean thermal energy, (calculated with respect to the forecast). In addition, when the installed
scoring a total theoretical potential up to 114,000 TWh/y of RE, whose capacity of renewable systems is higher than 5 MW, simulation data of
exploitation would exceed the global electricity demand by 400% [5]. the dynamic transients have to be provided, since the grid core re­
Moreover, 95% of countries worldwide are bathed by seawater [6]. Yet, quirements have to be strictly respected for the stability and operation of
the exploitation of such energy potential is still limited: the International the grid itself [20].
Energy Agency highlighted that ocean energy generation is not on track Short-term forecasting is frequently used to balance generation and
with the targets foreseen for 2025 and 2030 (4 and 15 TWh, respec­ load in the case of RESs exploitation, and shows better predictability for
tively) set by the European Commission [7]. Several studies addressing wave than for solar and wind energy [21]. Even so, uncertainty affects
the strategies for producing and delivering offshore renewable energy the actual possibility of complying with the dispatching plan. This issue
are available in the literature, assessing the relevance of different energy can be overcome by fossil fuel-based back-up systems able to respond to
vectors [8]. In recent years there was a fast development of Wave Energy renewable power shortages, applying the so-called valley filling tech­
Converters (WECs) [9], exploring a variety of hydrodynamic and gen­ nique [22]. Gas turbines (GTs) are the generation systems most widely
eration principles [10] as well as looking at several possible applica­ applied to this purpose, due to their compactness, the short time
tions, such as built-in ocean observing platforms [11] and autonomous required for their start-up and the wide declination of sizes [23].
underwater vehicles [12]. Nevertheless, the very specific design, the Moreover, Du Toit et al. (2020) evidenced the flexibility of Natural Gas
demanding installation and operation procedures and the uncertainty in (NG) microturbines towards co-combustion of fuel mixtures containing
the actual RE delivered to the shore still result in a high Levelized Cost of CO2 and H2, in the perspective of off-grid and emergency back-up of
Energy (LCOE) (0.30 – 0.55 USD/kWh) with respect to other RESs power applications [24]. In offshore, specific benefits may derive from
conversion technologies [13]. As a consequence, hybrid systems are synergies with fossil fuel exploitation and/or the decommissioning of
often considered a viable and convenient option for the valorization of depleted gas fields [25]. Actually, the decommissioning of hydrocarbon
wave energy since the combination of different offshore energy sources fields is usually not started after the complete depletion of the reservoir,
may avoid the presented bottlenecks: Hu et al. (2020) approach the but rather at the break-even point for economic benefits considering the
design of a floating wind farm supporting WECs through a numerical progressive reduction in the well pressure and/or in the quality of the
study that enables the optimization of the number and configuration of extracted fossil resources, that causes increasing costs in the resource
the converters [14]; Kluger et al. (2023) evidenced the power balancing delivery to the shore [26]. Thus, depleted oil&gas fields facing decom­
effects deriving from wind-wave hybrid systems [15]. missioning are often not completely exploited, while the associated
As in the case of other RESs, the profitable exploitation of wave infrastructure is still operable and the stranded gas is available on site,
energy depends on the capability to accurately characterize the wave representing a possible fuel for a back-up system and for Gas To Power
features, in order to perform the optimized design of the energy har­ (G2P) applications [25].
vesting devices. For instance, despite several research efforts in this Several studies addressing the coupling of wave farms (WFs) with
field, Jiang et al. (2022) highlighted that seldom wave characteristics G2P systems are available in the literature. Ou et al. (2017) demon­
are derived by accounting for the different wave systems composing the strated the application of a novel damping controller for the static
wave field [16]. Additionally, Coe et al. (2021) highlighted that the synchronous compensator in a hybrid power system made of an offshore
wave resource information is often wrongly accounted in WECs design, wind farm, a seashore wave farm, a battery storage system and a
being the RES intermittency and the device capacity factors too often microturbine park [27]. Oliveira-Pinto et al. (2019) explored the niche

2
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. 1. Flow chart of the methodology developed for the conceptual design of offshore hybrid energy systems based on wave energy and G2P.

market of wave energy power generation for the power supply to an oil an important role on the overall sustainability of non-programmable
and gas production platform in the Norwegian Continental Shelf, by RESs exploitation, in particular with respect to technical and eco­
halving the capacity of the existent gas turbine park [28]. G2P was also nomic constraints. Therefore, in the present study, the dispatchability of
selected as the suitable back-up strategy for a desalination facility the energy produced was assumed as a main factor in the design and
mainly powered by wave and solar energy on the island of Tenerife [29]. operation of a hybrid system. Moreover, the method developed provides
Recently, Dincer et al. (2021) proposed the design of offshore hybrid a set of Key Performance Indicators (KPIs) addressing the performance
energy systems based on Multi-Criteria Decision Analysis (MCDA), in the different sustainability domains (economic, environmental and
specifically addressing G2P applications to support wind energy har­ social) of the hybrid systems under design, in addition to those
vesting [30]. addressing the technological performance. The methodology is espe­
In the work of Nasrollahi et al. (2023), a comprehensive review of cially valuable in the comparison of alternative sites considered for wave
WEC assessment methods and MCDA tools is reported. The authors also energy exploitation. Actually, the MCDA approach developed enables
report the application of different assessment criteria to a set of WECs the conceptual design optimization and the assessment of the broad
proposed for industrial scale-up [31]. The proposed prioritization sustainability of technological alternatives considered for RESs exploi­
approach, entailing both the Fuzzy Delphi and the PROMETHEE tation in each site.
methods, is demonstrated through an application to the Caspian Sea, Two test cases were analyzed, one in the North Sea and the other in
whose wave energy potential was confirmed by the results of extensive the Adriatic Sea, in order to explore the potentiality of the methodology
numerical simulations carried out by Jahangir et al. [32]. and to assess the influence of site-specific parameters on the hybrid
In this framework, the present study aims at the development of an system design and on its sustainability performance.
innovative methodology for the conceptual design of sustainable hybrid In the following, section 2 presents the proposed methodology. In
energy systems based on wave energy, using offshore G2P systems as a section 3, the test cases are described. Section 4 presents and discusses
back-up for valley filling. The main novelty of the methodology consists the main results of the test cases, while conclusions are reported in
in the inclusion of the constraints deriving from the preparation of an section 5.
optimal dispatching plan in the conceptual design of these novel sys­
tems, in order to comply with a stabilized power supply to the grid. 2. Methodology
Indeed, the influence of the dispatching plan on the design of a gener­
ation system was seldom considered in previous studies, even if it plays A methodology supporting the conceptual design of offshore hybrid

3
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table 1 WEC data also need to be collected in step 1 of the methodology.

Site-specific input data collected in step 1 of the methodology. Data about the available WECs technologies are gathered from dedi­
Wave motion cated literature and from the available technical sheets. The selection of
the proper WEC starts by considering the relevant site specific data
a. Wave periods (Tm , Tp )
b. Significant wave height (Hs )
(water depth and distance from shore). The technological readiness level
c. Wave direction (θ) and the operative ranges of the device are also relevant, requiring the
WEC selection of technologies demonstrated on the full scale in conditions
a. Technical features: sizes, installation requirements similar to those of the site of interest. The set of suitable devices selected
b. Operative data: ranges and performance curves by this procedure is the input to Step 2, where the specific device used in
c. CAPEX and OPEX the design will be selected considering the power generation
Reservoir
performance.
a. Reservoir fluid properties: composition, heating value Information concerning the reservoir fluid and its availability, as
b. Reservoir fluid availability: residual pressure, flow rate well as the features of the offshore platform, also need to be collected in
Offshore platform
step 1 in order to carry out the conceptual design of the back-up system
a. General information: type, remaining lifetime in the following steps of the procedure.
b. Structure-related data: size, elevation of decks
c. Layout of actual equipment and process flow diagrams (if possible)

2.2. Conceptual design of the hybrid power system

systems exploiting wave energy was developed, entailing the prepara­
tion of an operation strategy that complies with the dispatching needs of As shown in Fig. 1, the preliminary design of the hybrid system (step
grid-connected generation systems. 2) is divided in two parts: step 2a aims at the conceptual design of the
Fig. 1 shows the main steps of the methodology. The procedure WF, while step 2b addresses the back-up system design.
initially requires the definition of the parameters needed for the hybrid
system specification: the nominal capacity of the WF (PWF ) to be 2.2.1. Conceptual design of the wave farm (step 2a)
installed and the probability of correct dispatching (Probd ) that the First, the WEC technology to be used has to be selected. The selection
producer commits to deliver to the TSO (step 0 of Fig. 1: System is carried out through the comparison of the generation performances of
specifications). the devices which were identified in step 1 as suitable to exploit the Pw,av
Step 1 consists in the collection and analysis of data related to the previously calculated. In detail, the hourly power generation curve
selected site(s) in a reference period. Step 2 foresees the design of the PWEC (t) [MW] in the t-th hour needs to be calculated for each device. The
hybrid system, starting from the WF (step 2a). On the basis of historical modelling approach proposed by [36] was used in the present study. It
real and forecast data, a dispatching plan is then obtained, driving the consists in the calculation of the WEC absorbed power by the use of the
optimal selection and sizing of the back-up system (step 2b). WEC specific power matrix and in its correction considering power-take-
In case alternative sites are considered for the installation of the off (PTO) and generator efficiencies:
hybrid generation system, a further step based on MCDA is proposed for PWEC (t) = Pw,abs (t) • ηPTO • ηgen (2)
the assessment and comparison of the sustainability performance of
each alternative. Pw,abs (t)[MW] is the power absorbed by the device. The parameters ηPTO
[-] and ηgen [-] are respectively the PTO and the generator efficiencies,
whose values depend on the specific type of PTO and electric generator
2.1. Data collection
used [36].
The power absorbed by the device, Pw,abs (t), is calculated applying
In step 1 of the methodology, the site-specific data needed for the
the specific WEC power matrix to the meteo-climatic parameters (bins
conceptual design of the hybrid system are gathered. Table 1 summa­
(Hs , Tm ) or (Hs , Tp )) occurring in the t-th hour [37]. The power matrix of
rizes the data that need to be collected.
a WEC is specified by the manufacturer and provides the WEC power
Meteo-climatic parameters describing the wave motion are collected
output for all operative combinations of wave height and wave period
over a reference period of at least 1 year with a minimum time resolution
((Hs , Tm ) or (Hs , Tp )) in the wave direction operative range (θmin - θmax ) of
of 1 h, as to study seasonality and to comply with the energy market
pricing methods [33]. The required data are: the significant wave height the WEC.
Hs [m], the mean wave period Tm [s], the peak wave period Tp [s] and The good practice suggests to consider the possibility of adaptation
of the device size to make the most of the specific meteo-climatic con­
the wave direction θ [-] [29]. In alternative to rough data directly ob­
ditions, taking into account that WECs are site-specific and scalable
tained from real-time measurements, aggregated data from long-term
devices [36]. The theoretical principles on which WECs are scaled are
trend analysis may also be used [34]. Starting from these parameters,
reported in the literature [38]. Details on the theory of wave generators
the quantification of the wave potential is carried out to allow the WEC
and a summary of the scale parameters used in the present study are
selection (step 2a). The wave potential is quantified calculating Pw,av
reported in section A2 of Appendix A.
[W/m], the average of the hourly wave energy flow Pw,t [W/m]
Finally, it should be mentioned that the developed approach is
throughout the reference period. Pw,t represents the power per unit of
limited to the conceptual design of the WF. Thus, the proposed design
surface of the crest length at the t-th hour and may be estimated as
does not address the lay-out of the WECs in the WF, neither its effects on
follows [19]:
the expected power output of the WF.
1 Once the power output of each WEC is obtained, the Annual Energy
Pw,t = •ρ 2
• ag 2 • Hs,t • Tm,t (1)
64π water Production (AEPWEC ) [MWh] is evaluated with Eq. (3):

where ρwater is the seawater density (1025 kg/m3) and ag [m/s2] is the ∑
m
AEPWEC = AV WEC • PWEC (t) (3)
gravitational acceleration constant. Forecast data of the same wave t=1
parameters with a time horizon of 6 h [35] and referring to the same
reference period also need to be retrieved. An in-depth discussion of real where m [h] is the number of hours in the reference period (8760 in a
and forecast data types and analysis is reported in section A1 of Ap­ year), AV WEC [-] is the WEC availability (i.e., the time fraction during
pendix A. which the device is available to produce power excluding any operation

4
M. Cipolletta et al. Applied Energy 347 (2023) 121410

and maintenance intervention). A value of 0.90 was assumed for AV WEC , As a matter of fact, ξ(t) has a negative value if the energy generated
as suggested in the literature [39]. and supplied to the grid is lower than the planned value. If the energy
The most suitable WEC for the site is identified according to the generated is higher than the forecasted value, a positive value is
following performance indicators, which were selected due to their obtained.
relevance in determining the technological performance of the devices: Once estimated, the errors, ξ(t), are statistically analyzed over the
time intervals Ii considered for the reference period (i.e., one-month
• AEPWEC , as defined in Eq. (3): the yearly energy generation of the intervals in a reference period of one year, therefore i = 12). Then,
device in cumulative terms, independently of the time trend of the the error distribution is fitted applying a best-fitting model based on
power output; accuracy criteria. Herein, the Anderson-Darling fitting model was
• PWEC [MW]: the average power produced by the device in the refer­ considered due to both its suitability in comparing different series and
ence period, indicating if the renewable plant is able to provide a its adaptability to data tails. Subsequently, the probability density
defined design load; function and the cumulative distribution function (CDF) of the samples
• CF, the Capacity Factor [-] [40]: this parameter quantifies the ratio are obtained.
between PWEC and the nominal power of the device (PWEC,nom ), As stated earlier, in the definition of the dispatching plan, a given
providing the average operation level of the device within the value of Probd (lower than 100%) was assumed: i.e. a Probd of 80%
boundaries of its own capacity. Thus, CF is a critical indicator of implies assuming that in the 20% of the time the RES exploited is not
technological performance, also allowing a comparison with the ef­ sufficient to produce the forecast power [43]. Based on this probability
ficiencies of other renewable energy converters; value, the absolute dispatch error ξd (i) per each time interval is calcu­
• CW, the Capture Width [m] [40], i.e. the ratio between PWEC and the lated. The prediction error corresponding to the CDF of probability of
wave potential Pw,av : this parameter allows the comparison of the incorrect dispatching equal to the complement of Probd in the i-th time
overall power generation of the WEC device with respect to the interval Nξ (i) is indicated as ξd (i) and may be calculated as follows:
average potential of the wave resource in the site considered for ξd (i) = ξ : Nξ (i) = 1 − Probd (5)
installation. Actually, the indicator expresses the equivalent wave
length that the WEC is able to absorb and to convert at a given site. As an example, if Probd is 80% and 12 months are analyzed, 12
allowable ξd (i) are calculated which correspond to a cumulative prob­
Clearly enough, a higher value of each parameter corresponds to a ability of incorrect dispatching of 20% (i.e. Nξ (i) = 20%).
higher performance of the WEC device. Thus, in order to select the most Finally, the hourly dispatched power Pd (t) can be obtained from Pf (t)
performing WEC, the device scoring the overall highest performance considering the absolute dispatch error ξd (i) [kW] for each time interval:
based on the above-listed parameters should be selected. To this pur­ In each Ii : Pd,i (t) = Pf ,i (t) − ξd (i) (6)
pose, a straightforward multi-objective approach is applied. The
approach is divided into two steps: the internal normalization and the Therefore, the Pd (t) declared for grid injection needs to be lower than
calculation of the average of the normalized indicators. A linear internal the hourly forecast power Pf (t) to avoid or at least to reduce prediction
normalization is adopted in the range [Imin , Imax ] according to the errors as much as possible. Thus, once defined the dispatching plan, the
equations presented in section A4 of Appendix A. Imin and Imax are conceptual design of the GTP may be carried out (step 2b). The neces­
defined as the lowest and the highest figures obtained for each criterion sary back-up power from the GTP (PGTP ) is defined considering the
considered in the MCDA within the compared WECs. In order to rank the maximum power that should be provided by the turbo machines, PGT (t)
overall performance of each device, the values of the normalized in­ assuming Probd equal to 100% (i.e., Pf equal to Pd at each hour):
dicators are summed and divided by four. The WEC device scoring the
PGTP = maxt (PGT (t)) (7)
highest performance index is selected for application.
The required number of WECs (NWECs ) is then calculated, considering The equipment model selection is then carried out, taking into ac­
the design WF capacity PWF . count the nominal power and the footprint of the single machines.
The real output power of the WF, Pr , is obtained by multiplying PWEC Among the available GT models, compact and light-weight aero-deriv­
by NWECs . In this approach, the power supplied to the grid is assumed to ative GTs are suitable for power generation at offshore platforms in the
be the gross power produced by the WF, neglecting inter-array losses low-medium range (4–66 MW), while micro-GTs are usually the best
related to wake effects and electrical losses deriving from inter-array option for smaller capacities (<1 MW). After the selection, the nominal
cables, export cables and the High Voltage Alternating Current power at full load (PGT,nom ), the nominal efficiency at full load (ηGT,nom )
(HVAC) substation [41]. This assumption is suggested in the literature and the size of the machine are noted. The total number of GTs (NGT )
for HVAC cables operating at a maximum rating of 200 MW and needed to obtain the PGTP is calculated considering their nominal
150–170 kV. Such systems are used for small distances between the capacity.
offshore site and the onshore grid delivery point (i.e. 20–50 km), as in The total footprint of the GTs is also assessed. The footprint needed
the case of G2P offshore hybrid energy solutions [42]. for GT installation needs to be compared with the available free space of
Similarly, Pf , the forecast output power of the WF, is obtained by the decks of the offshore structure. In case the footprint of the back-up
applying Eq. (2) to the forecast meteo-climatic data, multiplying the system exceeds the free space available on the decks of the offshore
result by the required number of WECs, NWECs . platform, the system specifications have to be revised.
The GTP is assumed to be operated according to the approach sug­
2.2.2. Conceptual design of the gas turbine park (step 2b) gested by Guandalini et al. (2015) [43]. Thus, on the basis of the
This step is aimed at defining the type, the nominal power and the declared Pd (t), the GTP guarantees a flexible and efficient power pro­
total number of the machines to be installed in the gas turbine park vision through the intelligent operation of the turbines in parallel and at
(GTP) of the hybrid generation system. An innovative procedure, the same part-load. Further details on the calculation of the part-load
applying the method of dispatching errors proposed by Dincer et al. [30] efficiency of low-medium and micro-GTs are reported in section A3 of
was used. Appendix A. The hourly power provided by the back-up fuel is obtained
In the definition of the dispatching plan, power prediction errors are as the ratio of PGT to ηGT . The related fuel consumption and emissions are
calculated as in Eq. (4), where ξ(t) [kW] is the hourly absolute error then quantified and used in the sustainability assessment.
between the real and forecast output powers of the WF.
ξ(t) = Pr (t) − Pf (t) (4)

5
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table 2
Reference values assumed for the normalization of the five sustainability indicators.
Key Performance Definition Reference Sustainability Reference Description Source
Indicator Equation domain boundary

LEE[%] Energy efficiency of the hybrid system in Eq. (8) Technological Imin 0 Worst case efficiency (no —
the analyzed period energy production)
Imax 78 Highest performance from
[50]
overtopping WECs
LCOE[EUR/MWh] Distributed production costs of produced Eq. (11) Economic Imin 210 Minimum LCOE for
[36]
energy from the hybrid system over its overtopping WECs
lifetime Imax 664 Maximum LCOE for
[36]
overtopping WECs
LVOE[EUR/MWh] Distributed revenues from the sale of Eq. (12) Economic Imin 0 Minimum recorded incentive
[51]
produced energy from the hybrid system on marine energy
over the analyzed period Imax 582 Maximum recorded incentive
[51]
on marine energy (Portugal,
2006)
LGHG[tonCO2eq/ Levelized GHG emissions from the hybrid Eq. (13) Environmental Imin 0 Less polluting EU country in
[52]
MWh] system over the analyzed period 2018 (Norway)
Imax 0.9 More polluting EU country in
[52]
2018 (Estonia)
LHI[m2/y] Levelized Inherent hazard risk from the Eq. (14) Societal Imin 0 No GT operating —
operation of the hybrid system over the Imax 0.18 GTP at regime coupled with NG
[46]
analyzed period extraction offshore platform
operation

2.3. Sustainability Key performance indicators turbines activated in the t-th hour.
As suggested by Dincer et al. (2021) [30] and IEA (2020) [47], in
In step 3 of the methodology (see Fig. 1), KPIs related to the sus­ order to separately account for the effect of costs and revenues related to
tainability performance of the hybrid system are calculated. The KPIs energy generation, the economic performance of the hybrid energy
aim at scoring the sustainability performance of the hybrid generation system is assessed using two different indicators: the LCOE and the
system designed in step 2. Therefore, the KPIs may be used as drivers to Levelized Value of Energy (LVOE). The LCOE [EUR/MWh] provides the
select among alternative siting options, as well as to identify the most production cost of each MWh of electrical power according to the fea­
sustainable design options and/or to identify critical issues in the tures of the generation system considered. It refers to the expected
outcome of conceptual design obtained from step 2. project lifetime, and it is estimated assuming a constant performance
A total of five KPIs are defined, addressing the different pillars of over the lifetime of the system [29].
sustainability, as summarized in Table 2. As shown in the table, an The LCOE is calculated as follows:
innovative feature of the present study is the introduction of specific ( )
∑
indicators to assess the expected safety performance of the hybrid sys­ CAPEX WF + CAPEX GTP + Tt=1
OPEX WF,t + OPEX GTP,t
m
(1+m)r

tem [30]. The approach entails the use of inherent safety indicators that LCOE = ( ) (11)
∑T PWF,t + PGTP,t
proved effective in a large variety of applications, addressing both t=1 m
(1+mr )
conventional [44] and innovative processes for energy vectors produc­
tion, onshore [45] and offshore [46]. In the following, the procedure for
where T [h] is the total number of hours in the project lifetime. In order
the calculation of each KPI is described.
to apply Eq. (11), the costs of the hybrid system are needed: OPEXWF,t
A novel indicator, the Levelized Energy Efficiency (LEE), is defined to
and OPEXGTP,t [EUR/h] are respectively the operating costs of the WF
assess the WF performance in the selected site, given that the GTP al­
and of the GTP distributed over the m [h] hours of the reference period.
ways runs at the highest possible efficiency. The LEE [-] weighs the
The parameter r [-] is the discount rate referred to the reference period.
contributions of the WF and the GTP energy conversion efficiencies over
The LVOE [EUR/MWh] provides the average market price of each
the m [h] hours of the reference period:
MWh produced by a generation system according to the pricing system
ηWF,av hWF + ηGTP,av hGTP adopted by the local grid during the sale process. Consequently, the
LEE = (8)
m LVOE is a metric deriving from the market prices, the incentives and the
rules applied by the local TSO to energy delivery. The LVOE is calculated
where ηWF,av [-] and ηGTP,av [-] are the average energy conversion effi­ as follows:
ciencies of the offshore WF and of the GTP during their respective ( )
operation periods, hWF [h] and hGTP [h]. ∑m RBPS,t +RI,t +Runb+,t − Cunb− ,t − CeGHG
m
The values of ηWF,av and ηGTP,av are the averages of the hourly effi­
t=1 (1+mr )
LVOE = ( ) (12)
ciencies ηWF,t [-] and ηGTP,t [-], over their operative periods, defined as ∑m PWF,t +PGTP,t
t=1 r m
( m)
follows:
1+

PWF,t where RBPS,t [EUR] is the revenue from the base price power sale, RI,t
ηWF,t = (9)
Pw,t Lchar [EUR] is the contribution of incentives, Runb+ [EUR] is the revenue due
to the positive unbalances, Cunb− [EUR] is the cost paid due to the
ηGTP,t =
PGTP,t
(10) negative unbalances which are not covered by the GTP, and CeGHG [EUR]
PGT,nom NGT,t is the hourly cost associated to GHG emissions.
It is worth to remark that the economic indicators adopted highlight
where PWF,t [MW] is the real power produced by the WF in the t-th hour the cost of energy generation accounting for both capital and opera­
and Lchar [m] is the characteristic length of the WEC. Similarly, PGTP,t tional costs, rather than providing the overall revenue of the investment
[MW] is the GTP power output in the t-th hour, PGT,nom [MW] is the needed to install the hybrid systems. The latter may be addressed using
nominal capacity of a single turbine, and NGT,t [-] refers to the number of

6
M. Cipolletta et al. Applied Energy 347 (2023) 121410

conventional methods based on Net Present Value or Internal Return Table 3

Rate indicators. However, in the framework of the present methodology, Data describing the sites selected for the test cases (reference year 2017).
the comparison of indicators directly related to the cost of energy gen­ Adriatic Sea North Sea
eration is considered more effective to assess the economic performance
Platform and coordinates Porto Corsini MW C L11b-PA (L11-B)
of the alternative hybrid generation systems. (44◦ 51′ N, 12◦ 37′ E) (53◦ 28′21″ N, 4◦ 29′23″ E)
In order to assess the environmental performance of the hybrid Platform’s distance from 8 km, 14 m 50 km, 250 m
system, an indicator considering the Levelized GHG emissions (LGHG) coast and water depth
[kgCO2eq/MWh] is used [30]. The indicator considers the emissions from Platform features 4-floors 8-legs, 48 m 5-floors 4-legs, 23 m × 20
(structure type, surface) × 22 m m
the GTP, eGHG,GTP,t [kgCO2eq] over the reference time period, divided by NG rate 4960 Sm3/h 6685 Sm3/h
the total energy production: NG composition CH4 ~ 100% CH4 (93%), ethane (3%),
∑m CO2 (2%) and N2 (2%)
eGHG,GTP,t Buoy and coordinates Nausicaa (44◦ 21′ N, L91 buoy (53◦ 61′ N, 4◦ 96′
LGHG = ∑m t=1 (13)
t=1 (PWF,t + PGTP,t ) 12◦ 48′ E) E)
Buoy’s distance from coast 8 km, 10 m 50 km, 250 m
The above defined indicator has thus a lower value for systems and water depth
relying mainly on RESs. Wave data source [56] [57]
The inherent safety indicator adopted to assess the safety perfor­ Data series type 30-minutes series 10-minutes series
Available data % 72 96
mance of the hybrid system is the Levelized inherent Hazard Index (LHI)
References [58,59] [60,61,62]
[m2/MWh], which evaluates the risk of fatalities for the human target
due to major accidents per each MWh added by the back-up system. The
LHI is the sum of the Inherent Hazard Indices HHIk [m2/y] associated to
Xecon = wLCOE • XLCOE + wLVOE • XLVOE (16)
each k-th process unit [48] and normalized with respect to the energy
produced by the sole GTP: The approach used for the normalization and to derive the weighing
NGT
coefficients is reported in section A5 of Appendix A.
∑ ∑
m
LHI = HHI k / PGTP,t (14) Finally, when the comparison of alternative sites or of alternative set-
k=1 k=1 ups is of interest, a sensitivity analysis is applied to the ranking of al­
ternatives based on the figures obtained for the aggregated sustain­
HHIk is referred to each GT, since such items are the only units pro­ ability indicator. Indeed, uncertainties might affect the procedure,
cessing hazardous substances (i.e. NG). mostly due to the variability of input data, to the normalization applied,
Further details on the calculation of the LCOE, LVOE, LGHG and LHI to the weighing parameters selected. Therefore, carrying out a sensi­
indicators are reported in section A4 of Appendix A. tivity analysis is advisable in order to identify possible aberrations and
In order to obtain an easy and clear metric of the overall sustain­ to verify the robustness of the final ranking. The Monte Carlo method is
ability performance, suitable for the application in a MCDA framework here proposed for application as it enables the random variations of
and useful to produce a ranking of the possible alternatives, a multiple key variables in a high number of iterations [53].
compensatory normalization and aggregation procedure is applied to The outcome of the methodology is thus the conceptual design of a
the above defined indicators [30]. A linear external normalization is sustainable hybrid generation system optimized for the exploitation of
adopted, defining two reference values as the upper and lower limits of RESs in the site of interest, based on a design value of the probability of
the range of each indicator (Imax and Imin ). The boundary limits are correct dispatching and on sustainability KPIs used for performance
selected as the extreme values which are unlikely to be exceeded by a assessment of alternative design options.
wave energy-based offshore hybrid generation system aiming at the More in detail, the application of the methodology provides the
valorization of the available RES (the wave resource). In an ideal sce­ identification of the most suitable WEC to be deployed, as well as the
nario, such system would not need any back-up power from the GTP. As number of these devices that need to be installed according to the
a consequence, the highest boundary limits correspond to the highest nominal potentiality of the wave farm and to design constraints. Based
achievable performances of the WF (with respect to the efficiency, the on the simulation of the real and of the forecast performance of the
lifetime energy generation, the incentives received and the capability in renewable system, the dispatching plan is obtained. Thus, the method
responding to the dispatching plan without power generation from the also enables the conceptual design of the back-up GTP, in terms of type
back-up system). On the contrary, the lowest boundary limits represent and number of machines and the definition of the GTP operating plan on
the worst-case scenarios, namely the poor energy conversion perfor­ the basis of the dispatched power curve. The set of KPIs provided allows
mance of the WF or its reduced operation, the low incentives, the high the quantitative assessment of the broad sustainability of the hybrid
power input from the GTP to compensate the underperformance of the system defined.
renewable power generation system. The values of the boundary limits
are retrieved from the literature, directly or introducing specific as­ 3. Case studies
sumptions, as shown in Table 2.
Using the values defined for the boundary limits in Table 2, the ab­ Two sites were selected to carry out case-studies aimed at demon­
solute indicators are re-worked to non-dimensional figures between strating the applicability and the results obtained from the developed
0 and 1 with a directionality (0 – undesired, 1 - desired). methodology. The selected locations are characterized by intense
Weighing and aggregation are then carried out on the normalized offshore extraction activities, the North Sea being the cradle of several
indicators. Weighing is performed using the Analytic Hierarchy Process oil&gas fields, while the Northern Adriatic Sea hosts plenty of NG re­
(AHP) method, which considers the trade-offs among the different serves. Furthermore, Italy is the country which mostly employs WECs
metrics according to four different social perspectives [49]. An overall prototypes in the Mediterranean area, with the Northern Adriatic being
Aggregated Sustainability Index (ASI) [-] is thus obtained through Eq. the test basins of several offshore devices [54], while North Sea is an
(15) and Eq. (16), where X identifies the normalized indicator and w the attractive region for WEC installation due to the high wave potentials (5
associated weight: to 10 kW/m [55]).
ASI = wLEE • XLEE + wecon Xecon + wLGHG • XLGHG + wLHI • XLHI (15) A preliminary assessment of the wave energy potential in the test
sites evidenced important differences, along with the site characteristics
where: [55].

7
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. 2. Bivariate distributions of yearly occurrences of (Tp, Hs) in 2017 in: (a) the Adriatic Sea test site; (b) the North Sea test site. The color scale represents the
contribution of each bin to the total incident energy, as a percentage. The wave iso-power curves are also reported.

In both locations an existing extraction rig is considered for the Table 3 summarizes the main data concerning the two sites for the
possible installation of the G2P system providing a back-up to a 2 MW reference year selected (2017). In the case-studies carried out, the
nominal capacity WF. The probability of correct dispatching is set equal reference year was selected on the basis of the availability of the
to 80% for the reference case assessment (scenario A), thus fulfilling the required input data.
requirements of step 0 of the proposed methodology (see Fig. 1). Fig. 2 reports the energy contribution of each bin to the total yearly
The data collection (step 1 of the methodology) is then performed. wave energy according to actual data obtained for the two sites. The
average wave potentials calculated are of 1 kW/m and 7.8 kW/m for the
Adriatic and North Sea cases, respectively. These values are coherent
Table 4
with the findings of Mørk et al. 2010 [55] (<5 kW/m for the Adriatic Sea
Performance parameters obtained for the five WECs selected in the test-cases
based on the available reference year meteo-marine data. The referenced sour­
and 5–20 kW/m for the North Sea). Further information about the sites
ces report the technical specification of the WEC devices considered in the considered is reported in Appendix B.
present study. WECs data is retrieved for the devices that could be installed in both
sites, i.e., Pelamis, Aqua BuOY, Lysekil WEC (either with 2 m or 4 m
Pelamis Aqua Lysekil Lysekil Wave
BuOY WEC WEC Dragon strokes) and the down-scaled device 1:1.5 of the Wave Dragon (WD).
with 2 m- with 4 m- prototype Some assumptions were introduced in the assessment of the case-
stroke stroke 1:1.5 studies:
Reference [50] [50] [65] [65] [50;66]
Adriatic Sea • the nominal capacity of the hybrid energy plants is selected equal for
AEPWEC [MWh] 11.4 5.4 1.4 1.4 83.0 the two sites in order to allow a significant comparison;
PWEC [kW] 4.7 2.2 0.6 0.6 34.1
• the same GT model, plant size and forecast horizon are selected in the
CF 0.01 0.01 0.02 0.01 0.04
CW[m] 4.7 2.2 0.6 0.6 33.9
two sites in order to reduce the number of diverging input parame­
Overall 0.09 0.06 0.13 0.06 1.00 ters and to capture the influence of the exploited RESs;
performance • the rig, the well and the measurement buoy are considered close to
ranking each other. The actual distances between the buoy and the WF
North Sea
installation are supposed to have a negligible influence on the vari­
AEPWEC [MWh] 794.5 199.6 28.9 32.0 1959.4
PWEC [kW] 90.7 22.8 3.3 3.6 223.7 ables related to wave motion across the same geographical area;
CF 0.12 0.09 0.12 0.08 0.23 • the lifetime considered for the two hybrid energy generation systems
CW[m] 11.6 2.9 0.4 0.5 28.6 is 20 years;
Overall 0.36 0.08 0.07 0.00 1.00 • for the sake of simplicity, the technological and economic perfor­
performance
mances in the reference year are assumed constant for the entire
ranking

8
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. 3. Scatter diagrams reporting the power production frequencies obtained for the application of the selected Wave Dragon device in the test sites: a) Adriatic Sea;
b) North Sea.

lifetime. Nevertheless, for more accurate evaluations, the application

Table 5
of a regional assessment considering the future conditions of the
Summary of the results of the conceptual design of the hybrid generation system
natural resources, as the one presented by Lira-Loarca et al. (2021)
in the two test sites.
[63], is suggested.
Adriatic Sea North Sea

In order to understand if any important modification of the results Selected WEC Wave Dragon prototype 1:1.5, Wave Dragon prototype 1:1.5,
with respect to the reference scenario A (80% probability of correct PWEC,nom = 0.96 MW PWEC,nom = 0.96 MW
PWF [MW] 1.9 1.9
dispatching) occurs when model parameters are changed, two further
NWECs 2 2
scenarios are introduced: Selected GT Microturbine Capston, Microturbine Capston,
PGT,nom = 200 kW PGT,nom = 200 kW
i) the declared dispatching accuracy is improved to 90% (scenario B); PGTP [MW] 2 2
ii) lower WEC costs are assumed due to full scale commercialization, as NGT 10 10

suggested by Sørensen and Friis-Madsen (2014) [64] (scenario C).

Nevertheless, as expected, the difference in the power output is
4. Results and discussion relevant between the two sites: the AEPWEC associated to the North Sea is
6.5 times higher than the one obtained in the Adriatic Sea. The power
4.1. Conceptual design of the hybrid generation system for the test cases production frequencies of the 1:1.5 WD prototype are reported in Fig. 3
in terms of operative bins (Hs , Tp ) for both sites. The figures allow the
Table 4 shows the values of the performance indicators obtained identification of the most frequent operating conditions. Notably, in the
from the simulation of the alternative WEC technologies considered in Adriatic Sea, the WD produces energy in a narrower range of Tp con­
the two test sites, as per step 2a of the methodology. The studies refer­
ditions and presents some powerful operative bins. On the contrary, in
enced in Table 4 report the technical specifications and the power
the North Sea, the same device operates in multiple sea conditions, but
matrices of each WEC, that were used in the performance evaluation.
its higher power contribution occurs with more powerful sea states (Hs
Table 4 also reports the performance indicators calculated for the
and Tp ranges are 1.5–2.5 and 6.5–8, respectively).
different WECs by the procedure described in section 2.2.1. As shown in
On the basis of the design capacity assumed for the WFs (2 MW), in
the table, the WD converter was found to be the best-performing WEC
each test site the hybrid energy system consists of 2 WECs (PWF = 1.92
technology in both sites.

9
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table 6 4.2. Sustainability assessment of the test-cases

Technological, economic, environmental and inherent safety performances of
the hybrid generation plants for the two case-studies considering scenario A. Once consolidated the design of the hybrid system, step 3 of the
TECHNOLOGICAL PERFORMANCE methodology is applied. Table 6 shows the values of the sustainability
Adriatic North indicators calculated for the two test cases analyzed. As far as the
Sea Sea technological performance is concerned, the limited operating time and
Wave Farm Available wave MWh 1143 8875
the low energy production in the Adriatic Sea test site result in only 312
energy yearly
Nominal Power MW 1.92 1.92 Equivalent Full Load Hours (EFLH), with respect to the 2044 EFLH
Electrical energy MWh 598 3919 calculated for the North Sea test case. Therefore, although the intensive
produced yearly performance of the device in the Adriatic site during its few working
EFLH h 312 2044
hours (demonstrating an average η of 70%), the difference in the LEE
hWF h 1648 6450
Percentage hWF – 19% 74%
values is clear (53% vs 14%), being the operation of the WDs more
Average η in hWF – 70% 68% extended in the North Sea than in the Adriatic Sea. Even if the results are
GTs park Nominal Power MW 2.00 2.00 difficult to benchmark due to the specific power levels, the values ob­
Electrical energy MWh 54 415 tained are similar to those reported by the comprehensive study carried
produced yearly
out by Pecher (2013) [67].
EFLH h 27 207
hGT h 198 830 With respect to economics, the cost items for the down-scaled WDs
Percentage hGT – 2% 9% are retrieved from the specialized literature [68]. CAPEX and OPEX of
Average η in hGT – 31% 31% the 200-kW GTs amount to 3150 EUR/kW and 20 EUR/MWh, respec­
Hybrid System Nominal Power MW 3.92 3.92
tively [69]. All cost items were updated to the reference year (2017)
Electrical energy MWh 651.6 4333.6
produced yearly
using the Consumer Price Index for the European zone. It should be
Thermal Energy MWh 172 1300 remarked that no emission costs are charged, as the two plants are
input classified as “small emitters” (i.e. < 25 kton of CO2 per year) according
LEE – 14% 53% to the European Emission Trading System [70]. The low energy gener­
ation in the Adriatic Sea test site is responsible for a LCOE value as high
ECONOMIC PERFORMANCE as 3960 EUR/MWh, about six times higher than the value obtained for
Adriatic North the North Sea test site (610 EUR/MWh). The LCOE figures obtained for
Sea Sea
Revenues from renewable energy incentives EUR 134,135 344,848
the North Sea test site are aligned with the findings of de Andres et al.
RI (2016), who reported values from 500 to 1600 EUR/MWh for large wave
Revenues from sales at base market price EUR 46,085 171,301 farms in the North Sea [71], as well as with the values reported by
RBPS Foteinis and Tsoutsos (2017) [13] for small wave farms.
Total from sale A = RI + RBPS EUR 180,220 516,149
Clearly enough, wave energy-based hybrid systems still have a too
Revenues / Costs Runb+ EUR 9600 0
from unbalances C-unb EUR − 850 0 high LCOE to be considered competitive with respect to other generation
-
B = Runb+ + Cunb EUR 8750 0 sources, especially in the case of the Adriatic Sea. However, the LVOE
Emissions CGHG EUR 0 0 results are promising, highlighting the presence of favorable pricing
penalties rules concerning wave energy-based power systems connected to the
Total = A + B-C EUR 188,970 516,149
LVOE EUR/ 290 119
local grids considered in the test cases. The LVOE is higher of a factor of
MWh about 2.5 in the Adriatic Sea test case, as a consequence of the higher
LCOE EUR/ 3960 610 incentives applied in Italy.
MWh It is interesting to compare the performance of the hybrid system to
that of the sole wave farm considering three main parameters: the LCOE,
ENVIRONMENTAL PERFORMANCE the LVOE and the dispatchability of energy. The complete results con­
Adriatic North cerning the performance of the WF in the absence of the GTP are re­
Sea Sea
ported in Appendix C. As expected, the exclusion of the back-up GTP
Total emissions tonCO2eq 31.9 240.5
LGHG kgCO2eq/ 48.9 55.5 implies a slightly improved economic performance in the test cases
MWh considered. Indeed, the LCOE decreases to 3600 EUR/MWh and 563
EUR/MWh respectively in the Adriatic and North Sea. The reduction in
INHERENT SAFETY PERFORMANCE the LCOE highlights that the GTP installation requires an additional cost
Adriatic North in front of a reduced contribution to the energy generation, since the
Sea Sea GTP operation strategy aims at the increase of the energy dispatch­
HHI m2/y 6.15E-03 3.75E-
ability, not at the increase of energy generation. This is clearly shown in
02
LHI m2/MWh 1.14E-04 9.03E- Fig. 4 that reports the power trends of both the WF and the GTP in the
05 North Sea test case on a sample time period.
Differently, the LVOE increases to 306 and 128 EUR/MWH in the
Adriatic and North Sea, respectively, when only the WF installation is
MW). The Pr and Pf curves are obtained as discussed in step 2a of the considered. The increase in the LVOE obtained for a fully-renewable
methodology. Consequently, step 2b can be applied for the quantifica­ system is due to the power pricing patterns, that are not only depen­
tion of the dispatching errors ξ(t), in order to derive the Pd curves in each dent on the amount of energy generated, but are rather influenced by
site. Based on the dispatch schedules defined, the potentialities needed both the adjustment of generated power to the declared dispatching
from the back-up systems are found. The results show that both in the curve and the pricing that the TSO assigns hourly to the energy dis­
Adriatic and North Sea, 2-MW GTPs are needed. A total of 10 aero- patched to the grid.
derivative machines rated 200 kW each were selected for each GTP Finally, it should be remarked that the reduced LCOE and the null
[23], as summarized in Table 5. values of LGHG and LHI associated with the WF in the absence of the
GTP come in front of a reduced dispatchability of the energy generated.
Actually, the WF alone fails to comply with Probd , the probability of
correct dispatching, which was set to assure a stable power supply in the

10
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. 4. Forecast and dispatching power curves obtained for the North Sea hybrid system (scenario A) along with the power generation trends from the WF and the
GTP for a sample period of the reference year.

grid. In the Adriatic Sea case, the reduction of 9% in the LCOE obtained different values in the two case studies. It should be remarked that XLCOE
excluding the GTP occurs with a slight decrease of the probability of is zero for the application in the Adriatic Sea since its non-normalized
correct dispatching (from 78% to 77%). In the North Sea test case, the value exceeds the maximum considered for the normalization range,
LCOE drops by 8%, but the probability of correct dispatching decreases that is, it exceeds the maximum LCOE expected for WECs nowadays.
from 77% to 71%. This result is consistent since the LCOE is calculated on the basis of the
Concerning the environmental impacts, the equivalent GHG emis­ costs of energy generation, thus it is strongly dependent on the perfor­
sions depend on the GTP dynamic operation, which in turn depends on mance of the renewable energy system in each location. Differently, the
the need of compensation between the dispatched and real power curves upper limit considered for the normalization is the maximum figure
hour by hour. The LGHG is similar for the two sites, with the Adriatic Sea available in the literature for commercially deployed wave energy
test case performing better than the North Sea one because of the slightly projects, thus it concerns projects where the economic viability was
more accurate forecast. The figures obtained are lower than the values confirmed.
expected for micro-gas turbines [72], due to the definition of LGHG, Overall, the North Sea test case demonstrates a higher performance
which distributes the equivalent CO2 emitted on both the renewable and than the Adriatic Sea test case, as confirmed also by the values of the
fossil-based generations. four ASI indicators, reported in Fig. 7-(a), which were calculated ac­
Similarly, the inherent safety of the two hybrid plants is influenced cording to the four weighing approaches considered.
only by the operation time of the GTs, since the features and sizes of the The ranking of the alternatives is unchanged when applying different
two GTPs are equal for the two test cases. The Adriatic Sea test site, due social perspectives (i.e. different weighing modes) to the calculation of
to the fewer operation hours of the GTP, features a lower inherent risk the overall sustainability indicator, ASI. Fig. 7-(b) reports the ASI met­
(see the absolute inherent risk indicator HHI in Table 6). However, the rics obtained for scenario B, i.e. when the improvement of the correct
LHI ranking is reversed because of the higher energy generation dispatching to 90% is applied, together with their percentage variations
occurred in the North Sea test case, as per Eq. (14). More in general, the with respect to the baseline scenario. The ranking of the test sites is
LHI values are aligned with those obtained in similar studies targeting unchanged. In the case of the Adriatic Sea, the ASI always increases for
the conversion of offshore platforms [46]. any social perspective considered. In the North Sea test case, the
The improvement of the probability of correct dispatching from 80% aggregated results are worsened when considering the Individualist and
to 90% (scenario B in section 3) causes the increase of ξd in every month Equal Weight ASI. The lower ASI values obtained by these weighing
of the year. However, this phenomenon has different effects in the two approaches are mainly due to both the decrease of the LEE and the in­
test cases analyzed, as shown in Fig. 5-(b). In the North Sea test site, crease of the LCOE. The former is caused by the prolonged intervention
almost negligible variations affect the cost items due to the presence of of the GTP, thus by the higher influence of the GT efficiency on the LEE.
relevant forecast errors. In the Adriatic Sea test case, the enhanced The latter is related to the reduced energy delivery, as a consequence of a
forecasting causes the dispatching plan to diverge to a larger extent from more measured back-up system generation due to the optimized forecast
the generation curve due to real-weather conditions, with two opposite and dispatching.
effects: higher revenues caused by the positive unbalances but also The North Sea test case provides the best performance, even in sce­
higher costs due to the negative unbalances. nario C (reduction in the commercial prices of the WF), as shown in
With respect to the possible future reduction of WEC costs (scenario Fig. 7-(c). In both sites, only positive variations of ASI are recorded. Also
C in section 3), Fig. 5-(c) shows that, as expected, a relevant reduction of in this case, the Adriatic Sea test case shows a higher relative
CAPEX and OPEX (68% and 73% respectively) will have a favorable improvement. The Individualist and the Equal Weight weighing ap­
impact on the process economics: the LCOE drops to 1300 EUR/MWh proaches are influenced by cost reductions to a larger extent than the
and 248 EUR/MWh for the Adriatic Sea and North Sea sites respectively, other weighing modes, since these approaches give a notable consider­
with a higher relative benefit for the first site. The different impact of the ation to the stakeholder’s economic convenience.
reduced OPEX in the two sites is due to the different operation of the Fig. 8 shows the outcome of the sensitivity analysis, aimed at veri­
wave farms in the two locations. fying the robustness of the sustainability ranking based on the ASI
The normalized indicators calculated for the two test cases consid­ indicator.
ering scenario A are displayed in the radar plot of Fig. 6. Apart from In detail, the Monte Carlo method was applied to some critical input
XLGHG and XLHI , which are aligned, XLEE , XLVOE and XLCOE score very parameters for 106 iterations, assuming their variation according to both

11
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. 6. Normalized sustainability indicators calculated for the two test sites
considered in scenario A.

4.3. Discussion

The proposed methodology uses the forecast-based dispatching

curve and a user-defined probability of correct dispatching to support
the conceptual design of offshore hybrid power generation systems
exploiting wave energy in combination with a natural gas fueled back-
up system aimed at valley filling. The proposed design procedure al­
lows the system to comply with the WF capacity and with the probability
of correct dispatching, thus fulfilling the main TSO requirements for the
supply of renewable energy in the grid. The wave resource exploitation
is designed by a specific step addressing the selection of the WEC
technology, following the simulation of the device performance over
real site-specific meteo-climatic data. Basing the back-up generation
system design on the probability of correct dispatching avoids over-
sizing of the system and allows a flexible operation of the GTP. The
final sustainability assessment, which resulted robust as per the sensi­
tivity analysis, provides an a posteriori evaluation of the conceptual
design proposed, that allows the performance comparison of alternative
options suitable for wave energy harvesting which have been previously
identified as viable based on a preliminary scouting.
Thus, the sustainability footprint and overall index obtained from the
method address the identification of the most sustainable configuration
of an offshore wave-based generation plant with respect to the wave
potential and to the best operational adaptability to the grid
requirements.
Fig. 5. Costs and revenues for the Adriatic Sea and North Sea test cases: a) Within the sustainability assessment, costs represent an important
scenario A; b) differences calculated when considering scenario B with respect issue. In the analysis of the test cases, no difference was introduced in
to scenario A; c) differences calculated when considering scenario C with the single cost items considered for the two test cases. Actually, even if
respect to scenario A (see Section 3).RI = total revenues from incentives, RBPS = two different geographical areas are compared (Adriatic and North Sea),
revenues from base price sale, Runb+ = revenues due to positive unbalances, it should be noted that they belong to the same economic area. In
Cunb− = costs due to negative unbalances. addition, the capital costs related to offshore projects undergo the dy­
namics of the global market, while operative costs, which are more
the uniform and the beta distribution (α = β = 2) in defined existence influenced by regional factors, have a lower influence on the results.
ranges. The most critical parameters are those which are subject to the However, in perspective, if sites in geographical areas having different
analyst decision and perception. Therefore, the two extremes of the levels of economic development are compared, the methodology should
normalization range as well as the aggregation weights for each sus­ be extended to include cost correction factors taking into account the
tainability metric are variated: the former in a range +/- 50% of the different offshore operational theaters considered for the installation of
original values, the latter with the constraint to sum up to 1. The CDF the hybrid system.
curve of the beta distribution was calculated for the differences between Nevertheless, the cost figures obtained in the test cases show that a
the North Sea (ASIN) and the Adriatic Sea (ASIA) aggregated indicators. hybrid generation system obtained coupling a renewable generation
In the baseline case (scenario A in section 3), this difference is positive. system to a GTP, operated with a strategy aiming at increasing the sta­
The obtained CDF curves confirm that a robust ranking was provided bility of the energy supplied to the grid and to minimize the energy
and that its inversion has a negligible probability of occurrence. generation from the GTP, results in a LCOE higher than that of the
Notably, the robustness of the results is also confirmed when consid­ renewable system alone. However, when calculating the LCOE of
ering, more conservatively, a uniform distribution for the variation of renewable energy generation systems, the overall costs of grid balancing
the target parameters (red curve in Fig. 8). are usually not considered. Further work will thus be needed to fully
understand the actual economic performance of hybrid systems when

12
M. Cipolletta et al. Applied Energy 347 (2023) 121410

taking into account the costs related to the grid balancing and to the
stabilization of energy supply.

5. Conclusions

A systematic methodology to support the conceptual design of wave

energy hybrid systems and to assess their sustainability performance
from the technological, economic, environmental and process safety
viewpoints is proposed. The method is demonstrated by its application
to two test cases, where hybrid solutions are tested for wave energy
valorization in the Adriatic Sea and in the North Sea. The hybrid gen­
eration systems defined applying the developed methodology perform
differently in the two sites, due to differences in the available wave
energy resource and in the forecasting accuracy. The annual energy
production resulted 6.5 times higher in the North Sea test-case.
Conversely, the low energy generation in the Adriatic Sea test site re­
sults in a levelized cost of energy of 3960 EUR/MWh, much higher than
the value of 610 EUR/MWh obtained for the North Sea. In both cases,
the installation of the gas turbine park impacts negatively on the cost of
energy production but is of crucial importance in meeting the design
value of the probability of correct dispatching.
The application of the multi-criteria method developed allows the
quantification of the system response under each sustainability aspect,
highlighting site-specific strengths and weaknesses. The sustainability
ranking obtained highlights the best performance of the North Sea test
site. The robustness of the sustainability ranking was confirmed by a
multi-variable Monte Carlo analysis. The developed methodology is
intended as a support for the development and design of hybrid gener­
ation systems based on the exploitation of wave energy. Its novelty is in
the inclusion of an appropriate interface modulation between renewable
generation and the grid owners, aimed at avoiding technical issues and
at maximizing the dispatching accuracy by the definition of a target
probability for correct dispatching. The assessment of the dispatching
plan and the back-up system management strategy approached by the
methodology also allow unveiling economic aspects related to the
operation of the system, thus providing a comprehensive support to the
decision-making process in the conceptual design stage of the project
lifecycle.

CRediT authorship contribution statement

Mariasole Cipolletta: Conceptualization, Investigation, Methodol­

ogy, Writing – original draft. Anna Crivellari: Conceptualization,
Investigation, Methodology. Valeria Casson Moreno: Conceptualiza­
tion, Methodology, Validation, Writing – review & editing. Valerio
Cozzani: Conceptualization, Writing – review & editing, Supervision.
Fig. 7. ASI results according to the four decision-making perspectives: a) sce­
nario A; b) scenario B; c) scenario C. Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Data availability

Data will be made available on request.

Appendix A:. Specific details of the methodology

Real vs forecast data

Fig. 8. CDF of the difference between the ASI index calculated for the North Real data are in-situ measurements derived from specific devices
Sea test case (ASIN) and that calculated for the Adriatic Sea (ASIA) resulting
located at the offshore site or in the close vicinity. Based on the type of
from the sensitivity analysis where error probabilities were simulated using a
instrument and recording technique used, wave climate parameters can
uniform (red curve) or a beta (α = β = 2, blue curve) distribution.

13
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table A1
Main scaling rules based on scaling factor λ for WECs at different sites [38].
WEC (λ = L1 /L2 )
Parameter Relation Scale dependence

WEC power production PWEC PWEC,1 /PWEC,2 λ7/2

Significant wave height Hs Hs,1 /Hs,2 λ
Wave period Tm or Tp Tm,1 /Tm,2 or Tp,1 /Tp,2 λ1/2

be obtained as average values of the wave parameters for short-term statistics (from seconds to hours according to the measurement principle of the
device) [67]. Bivariate distributions of occurrences corresponding to different combinations of Hs and Tm (or Hs and Tp ) – hereafter called bins - can be
also derived and visualized as scatter diagrams. A less cumbersome way to describe wave conditions is to group bins into a limited number of zones,
referred to as “sea states”, on the basis of Hs and Tp values of interest in a specific marine location [67]. Concerning wave forecasting, two major
methods can be distinguished: physics-based and time series models [73]. The comparison between physics and time series models showed statistical
methods are more accurate than physic models over limited time spans (1–4 h) [21], while for longer forecasts physics-based methods tend to produce
more reliable results, provided that the convergence point between the two techniques for comparable accurate results is around 6 h forecast horizon
[35].
Time horizons between 3 and 6 h enable to meet the variations in production and to control the capacity at the system operators’ disposal.
Furthermore, for an improved dispatching, the time horizon for forecasting should be consistent with the market operation constraints [74] that
vary worldwide, e.g. (about 6 h in the United States of America, Canada and UK [75], 4–5 h forecasts in Europe while 2–3 days ahead are used to
determine the available reserves for the day-ahead market [73]).

Theory and scaling of WECs

The knowledge of the performance of the WECs is essential.

Typically, a performance matrix is provided by the constructors reporting the produced power as a function of bins (Hs , Tm ) or (Hs , Tp ) [58]. Power
matrices may refer to the net electrical power output of the WEC (PWEC ) or to the power absorbed by the device (Pw,abs ); in the latter case, power-take-
off and generator efficiencies (ηPTO and ηgen ) should be considered to account for the conversion of absorbed power into rotating mechanical power and
for the transformation of rotating mechanical power into electrical power [76]. Power matrices are derived from experimental data or numerical
simulations for given operating conditions based on the working principle of the device [67].
During the primary conversion of marine energy inside the WECs, gravity and inertial forces are dominant and the effect of remaining forces (such
as kinematic viscosity) are negligible, thus mechanical similarity is achieved by the Froude’s scaling law between model and prototype devices [77].
The scaling factor λ, based on Froude similitude, is the ratio between the characteristic lengths L1 of the reference device and L2 of the scaled device: by
varying the capture width, the output power changes, but the device’s response keeps constant with the spread over the frequency range [38].
Table A.1 shows the scale dependence of all parameters of a down-scaled power matrix [39].

GTs part-load efficiency

Depending on the GT category, the efficiency reduction at part-load can be expressed as a function of the part-load ratio (power produced at part-
load, PGT , against PGT,nom ).
For aero-derivative GTs, the following correlation proposed in the literature [43] is used in the present methodology:
( )3 ( )2 ( )
ηGT PGT PGT PGT
= 0.7035 • − 1.91151 • + 2.0642 • + 0.1481 (A1)
ηGT,nom PGT,nom PGT,nom PGT,nom

For micro GTs, a proper correlation is derived in the present work by regressing the part-load efficiency curve of commercial micro GTs by
Capstone Turbine Corporation (C800 and C330 models) [78,79], expressed as follows:
( )3 ( )2 ( )
ηGT PGT PGT PGT
= 2.1812 • − 4.6655 • + 3.4475 • + 0.0584 (A2)
ηGT,nom PGT,nom PGT,nom PGT,nom

Among the possible control strategies proposed in the literature for GTs [80–84], the approach suggested by Guandalini et al. [43] is adopted in the
present study, which consists in managing each GT of the park in parallel at the same part-load. Starting from the full-load condition of the park, if the
load decreases all machines reduce equally their load up to the condition at which one machine can be switched-off and all remaining machines return
to operate with their PGT,nom . The strategy proceeds equally in the case of other load decreases, until one sole machine remains in operation before
reaching its minimum technical load, which is set at 50% of PGT,nom in order to meet the environmental limits on CO and NOx commonly imposed in the
technical specifications.

Details for the calculation of sustainability indicators

A4.1 LCOE

The LCOE is calculated starting from preliminary estimates of capital expenditures (CAPEX), operating expenditures (OPEX) of the hybrid plant,
thus of the wave farm and the associated back-up system. CAPEX and OPEX associated to the GT plant for aero-derivative and micro GTs are available
in the literature [85,86,87].

14
M. Cipolletta et al. Applied Energy 347 (2023) 121410

A4.2 LVOE

Dealing with electricity sale, the market prices for renewable power can vary depending on the pull mechanism adopted at national level for the
promotion of RESs participation into the grid (e.g. feed-in tariff, feed-in-premiums, quota-based tradable green certificates, investment subsidies or tax
cuts) that are added to the electricity base price. Clearly enough, eligibility requirements to receive support should be verified according to the current
national regulations.
Therefore, market base prices of electricity (Priceel,base ) are retrieved with particular attention to possible incentives launched at national level for
the promotion of RESs and related eligibility requirements. Furthermore, based on the local regulations to control power unbalances, producers may
receive incentives (Priceunb+ ) when power injection is higher than the declared one, but, on the contrary, may also pay fines (Priceunb− ) in case of
negative unbalances [43]. Thus, prices established by the local TSO for positive and negative power unbalances need to be collected.
Finally, the price associated to GHG emissions from the GTP is retrieved: it may be a carbon allowance total direct GHG emissions from specific
sectors in a cap-and-trade system (e.g. emission trading scheme or ETS) or a pre-defined carbon tax on GHG emissions based on the policy adopted by
local governments [88]. Information about the regional, national and subnational carbon pricing initiatives implemented, scheduled and under
considerations, including the associated prices, is published every year by the World Bank [89]. The revenue and cost items composing LVOE are
detailed in the following equations.
Rsale,t = Priceel,base • PGT,t + Priceel,RES • PWF,t = RBPS,t + RI,t (A3)

Rsale,t (in Eq. A3) is the total revenue from electric sale through grid feed-in. Within total revenues from sale, a distinction can also be made between
revenues from base price power sale RBPS,t and the ones only due to incentives RI,t . Priceel,base is the market price [EUR/MWh] of conventionally
generated energy and Priceel,RES [EUR/MWh] is the market price of each MWh of renewable energy supplied to the grid, comprehensive of possible
incentives.
Eq. A4 and Eq. A5 show the calculations for the hourly revenues and costs from positive and negative unbalances, respectively:
Runb+,t = Priceunb+ • Punb+,t (A4)

Cunb− ,t = Priceunb− • Punb− ,t (A5)

where Punb+ and Punb− are the average powers unbalanced in the t-th hour, respectively in excess and in defect with respect to Pd according to the
dispatching plan.
GHG emission costs CGHG,t at t-th hour depend on the monetary fines imposed over CO2eq emissions (PriceGHG [EUR/kgCO2,eq]) and the hourly
emissions from the GT park eGHG,GT [kgCO2eq] as per Eq. A6:
CGHG,t = PriceGHG,t • eGHG,GT,t (A6)

A4.3 LGHG

GHG emissions for the hybrid plant are estimated for the calculation of LGHG, provided that the renewable power plant’s contributions during
operation are null. Hourly emissions (eGHG,GT ), commonly expressed in units of equivalent CO2 (CO2eq) for a given period, can be evaluated from Pfuel
by assuming a typical emission factor per fuel and machine. For example, typical emission factor values are 202 kg/MWhfuel in the case of aero-
derivative GTs [43] and 185 kg/MWhfuel in the case of micro GTs powered by NG [72,90].

A4.4 LHI

The safety indicator used, the Levelized inherent Hazard Index, is defined as in Eq. A7 where HHIk is an index quantifying the risk for humans and
associated to the k-th process unit (here a gas turbine), PGT,t is the power output of the GT park at the t-th hour.
∑
NGT ∑
T NGT N∑
∑ LOCs ∑
T
LHI = HHI k / PGT,t = 2
(cf i,k • maxdi,j,k )/ PGT,t (A7)
j
k=1 k=1 k=1 i=1 k=1

More in details, HHIk [m2/y] is used to quantify the risk for the human target associated with the k-th process unit of the plant in terms of
potentially damaged area per year [48]. di,j,k [m] is the damage distance at which human targets undergo lethal effects when the j-th accident scenario
happens as a consequence of the i-th release mode from the k-th process unit. cf i,k [1/y] is the Credit Factor assigned to the i-th release mode, rep­
resenting the credibility of the event leading to the dispersion of hazardous substances. NLOCs represents the possible release modes of the k-th unit. In
order to correctly account for the intermittent operation of the turbines, affecting the credibility factors cf i,k , which are normally entailing continuous
operation during the year [91], in the newly-defined indicator LHI, the cf i,k are evaluated as in Eq. A8: CFi,k is the credit factor taken from specialized
literature and referred to full-year operation [91], hop the number of operation hours in the reference year and hyear the number of hours in the year.
hop,GT
cf i,k = CFi,k • (A8)
hyear

15
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table A2
Scores and weights assigned for aggregation of indicators based on different
perspectives.
Decision-making Criteria Indicators used for the MCDA
perspectives
LEE LCOE and/ LGHG HHI
or LVOE

Score Score 1–5 Score Score

1–5 1–5 1–5

Individualist Time 2 3 3 5
Space 3 4 3 5
Receptor 2 4 5 5
Sum 7 11 11 15
Weight 0.159 0.250 0.250 0.341
Egalitarian Time 5 3 5 1
Space 5 3 5 1
Receptor 4 1 5 1
Sum 14 7 15 3
Weight 0.359 0.179 0.385 0.077
Hierarchist Time 3.5 3 4 3
Space 4 3.5 4 3
Receptor 3 2.5 5 3
Sum 10.5 10 13 9
Weight 0.253 0.217 0.313 0.217

Normalization and weighing procedures

A5.1 the normalization procedure

Depending on the definition of the absolute metric, the performance improvement leads to the maximization or minimization of the indicator.
There follows the necessity to apply the most appropriate normalization approach to correctly obtain the non-dimensional indicator (XI ): Eq. A9 to
those indicators that increase their performance by maximizing their values (as LEE and LVOE), and Eq. A10 to the ones behaving in the opposite way
(LCOE, LGHG, LHI).
{
Iact − Imin XI = 1if Iact > Imax
XI = with (A9)
Imax − Imin XI = 0if Iact < Imin
{
Imax − Iact XI = 1if Iact < Imin
XI = with (A10)
Imax − Imin XI = 0if Iact > Imax

A5.2 The weighing procedure

The common method to extract trade-offs between indicators is the Analytic Hierarchy Process (AHP) method [49] which makes use of pair-wise
comparisons to evaluate the performance of the alternatives on indicators (scoring) and indicators among themselves (weighing) [92]. To limit the
intrinsic subjectivity of this process, a literature procedure for deriving importance coefficients is used, which bases the elicitation of trade-offs weights
on proper criteria (time–space-receptor) along with three decision-makers perspectives (individualist-egalitarian-hierarchist) [93]. Equal weighting is
further added to the archetypes of decision-makers.
Before scoring and weighing, the indicators proposed for the assessment are classified in terms of time, space and receptor criteria based on their
definition.
Being a measure of resource use, LEE is considered important for a long-term perspective and on a global scale since improvements may lead to a
better resource utilization and to lower emissions, with reduced costs. The ecosystem is evaluated as the main receptor by this indicator, but also
humans may be influenced since resources’ use and costs are human-related aspects.
LCOE and LVOE exhibit mainly short-term, local/regional and anthropocentric perspectives. However, they can be considered unimportant over
time since externalities (e.g. available resources, sociopolitical variations and other local/regional factors) are internalized into the cost/price as­
sessments.
LGHG is considered a very important long-term and global-scale concern, even though its effect may be on the short-term and at local scale based
on the incidence of the weather. Thus, it is evaluated as neutral on time and space criteria. Both humans and ecosystems are sensitive receptors.
LHI is an indicator quantifying the likelihood of severe accident scenarios for human targets. It is short-term with respect to the time horizon.
Moreover, it is local in terms of space perspective, as it quantifies the inherent hazard within the area containing the chemical plant/facility. This
indicator is consistent with respect to association with humans as opposed to the ecosystem.
For each archetype of decision-makers, given scores in a five-level Likert scale are assigned to the indicators based on these three criteria. The
overall score to each indicator is estimated as the sum of the scores given with respect to each criterion. The relative importance of the indicators is
determined as the ratio of the associated overall score to the sum of overall scores. All values of assigned scores and weights are reported in Table A.2.
These weights are used to derive the pair-wise comparison matrix, the evaluation matrix and the trade-off weights among indicators according to the
AHP method.

16
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. C1. Costs and revenues obtained in the case of 80% and 90% correct power dispatching in the Adriatic Sea test site (a) and in the North Sea test site (b).

Fig. C2. Normalized metrics compared for the two offshore sites in the case of 80% correct and 90% correct dispatching (A stands for the Adriatic Sea, N for the
North Sea).

Appendix B:. Detailed data concerning the sites of the case studies

The site in the Adriatic Sea

The northern area of the Adriatic Sea is characterized by offshore extraction platforms close to decommissioning. Porto Corsini MW C platform was
selected (44◦ 51′ N, 12◦ 37′ E) since in 2017 it was harvesting NG from 4 wells out of the 12 originally linked, with an average production of 4960 Sm3/h
and a gas composition mainly consisting of methane after dehydration [58].
From a financial and economic viewpoint, day-ahead prices for the Italian case are retrieved from the site of the Italian body of Electric Market
Management “GSE” which provides hourly prices; market data referred to Northern Italy are selected to comply with the hybrid energy plant location
[94].
In the Italian case, the participation in the low auction is optimistically supposed to lead to the minimum discount on the base tariff for incentives
(2%).
The Italian electricity market also evaluates the positive and negative unbalances by assigning prizes and costs. The allocation of prizes or costs not
only depends on the unbalance sign with respect to the dispatch power, but also on the zonal signs and sale/purchase dynamics, which are defined
hourly by the national TSO TERNA [33]. Thus, referring to the Northern Italy region and to the reference year 2017, historical hourly data are
retrieved from the public dataset “SunSet” where aggregated zonal signs, average sale and purchase prices are available monthly [95].
In the reference year, incentives on RE were regulated by a decree of the Italian Ministry of Economic Development in force since 2016. The decree
foresaw indirect access to incentives by means of a low auction on the base tariff (discounted from 2 to 40%) for those hybrid plants whose total
capacity fell below the threshold value set for the nominal potential of the sole RE generators; moreover, the produced energy eligible for incentives,
defined as the total generation diminished of the thermal energy in input, was to be at least the 5% of the totality. In the case of hybrid energy facilities
based on ocean energy, the threshold was 5 MW and the base incentive tariff amounted to 300 EUR/MWh [96].

17
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Fig. C3. Costs and revenues obtained in the case of sole wave farms applied in
the test-sites, compared with the respective results obtained for scenario A
(hybrid systems).

Fig. C4. Normalized sustainability indicators calculated for the two test sites
according to the scenario of sole wave farms operating.

Fig. C5. ASI results according to the four decision-making perspectives: a)

scenario A; b) scenario B; c) scenario C.

The site in the North Sea

The North Sea hosts several oil and gas reserves, divided according to blocks as displayed in the Dutch Oil and Gas portal [61]. In the North Sea, the
L11 block was considered and well L11B-A-08 was selected out of the two still open in 2017, being the extraction end foreseen for 2018. L11B-A-08 is
afferent to the near L08-D gas field, which is a common gas–water charged structure located at 50 km from the shore with a sub-sea extension of 37
km2 [60].
Regarding economics in the North Sea location, the Dutch electricity market is based on day-ahead prices, available hourly on the website of the
European Network of Transmission System Operators for Electricity [97]. Incentive tariffs for Dutch renewable generation are established by the

18
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Table C1
Absolute indicators compared for the two offshore sites in the case of 80% and
90% correct power dispatching.
Indicators Adriatic Sea North Sea

80% 90% 80% 90%

LEE % 14 14 53 52
LGHG kgCO2eq/MWh 48.9 40.5 55.5 39.1
LVOE EUR/MWh 290 296.1 119 121.5
LCOE EUR/MWh 3960 4015 610 627
LHI m2/MWh 1.14E-04 1.12E-04 9.03E-05 8.74E-05

Table C2
Technological, economic, environmental and inherent safety performances of
the sole wave farms operating in the test cases.
TECHNOLOGICAL PERFORMANCE
Adriatic North
Sea Sea
Wave Farm Available wave MWh 1143 8875
energy yearly
Nominal Power MW 1.92 1.92
Electrical energy MWh 598 3919
produced yearly
EFLH h 312 2044
hWF h 1648 6450
Percentage hWF – 19% 74%
Average η in hWF – 70% 68%
GTs park Nominal Power MW 0 0
Electrical energy MWh 0 0
produced yearly
EFLH h 0 0
hGT h 0 0
Percentage hGT – 0% 0%
Average η in hGT – NA NA
Hybrid System Nominal Power MW 1.92 1.92
Electrical energy MWh 598 3919
produced yearly
Thermal Energy MWh 0 0
input
LEE – 13% 50%

ECONOMIC PERFORMANCE
Adriatic North
Sea Sea
Revenues from renewable energy incentives EUR 134,135 344,848
RI
Revenues from sales at base market price EUR 43,645 155,075
RBPS
Total from sale A = RI + RBPS EUR 177,779 499,923
Revenues / Costs Runb+ EUR 9600 0
from unbalances C-unb EUR − 3389 0
B = Runb+ + C-unb EUR 5210 0
Emissions CGHG EUR 0 0
penalties
Total = A + B-C EUR 182,990 499,923
LVOE EUR/ 306 128
MWh
LCOE EUR/ 3600 563
MWh

ENVIRONMENTAL PERFORMANCE
Adriatic North
Sea Sea
Total emissions tonCO2eq 0 0
LGHG kgCO2eq/ 0 0
MWh

INHERENT SAFETY PERFORMANCE

Adriatic North
Sea Sea
2
HHI m /y 0 0
LHI m2/MWh 0 0

19
M. Cipolletta et al. Applied Energy 347 (2023) 121410

Netherlands Enterprise Agency through “SDE+” incentive schemes, to which plant operators can apply in four rounds. Application for incentives was
supposed to be carried out in year 2016, while plant production in 2017. Thus, the reference incentive for free flowing energy and wave energy is 12
EUR/MWh, equal to the 2016 average values [98]. This value has to be diminished by a correction amount determined by the energy prices of the
current year, thus to enable a good Dutch market representation, the correction amount was taken from “SDE + 2018” where the provisional one (3.8
EUR/MWh) is quantified from the real 2017 price tenors [99]. Moreover, according to “SDE + 2016”, the maximum RE eligible for incentives is
equivalent to 3700 full load hours and no prizes are assigned to those generation hours following at least 6 h of null energy supply to the grid.

Appendix C:. Detailed results of the scenarios considered

See Figs C1-C5 and Tables C1 and C2.

Scenario B: Hybrid system with 90% probability of correct dispatching
Fully-renewable generation system: A parallel scenario
The following results derive from the application of the proposed sustainability assessment in the case only the wave farms are the installed systems
in the considered test-cases, therefore excluding the possibility of supporting the generation with a compensating back-up system.

References [21] Reikard G, Robertson B, Bidlot JR. Combining wave energy with wind and solar:
short-term forecasting. Renew Energy 2015;81:442–56. https://doi.org/10.1016/j.
renene.2015.03.032.
[1] Our World in Data. Share of electricity production from renewables 2021. https://
[22] Guo Z, Zhang R, Wang L, Zeng S, Li Y. Optimal operation of regional integrated
ourworldindata.org/grapher/share-electricity-renewables?
energy system considering demand response. Appl Therm Eng 2021;191:116860.
tab=chart&stackMode=absolute&time=earliest..latest&country=~OWID_
https://doi.org/10.1016/j.applthermaleng.2021.116860.
WRL&region=Europe (accessed February 25, 2021).
[23] Capstone Turbine Corporation. C200S ICHP Microturbine. 2021.
[2] Our World in Data. Annual change in renewable energy generation 2021. https://
[24] du Toit M, Engelbrecht N, Oelofse SP, Bessarabov D. Performance evaluation and
ourworldindata.org/grapher/annual-change-renewables?
emissions reduction of a micro gas turbine via the co-combustion of H2/CH4/CO2
tab=chart&time=2001..2019&country=~OWID_WRL (accessed February 25,
fuel blends. Sustain Energy Technol Assessments 2020;39:100718. https://doi.org/
2021).
10.1016/j.seta.2020.100718.
[3] IRENA. Renewable Energy Finance Flows 2021. https://www.irena.org/Statistics/
[25] Kolian SR, Godec M, Sammarco PW. Alternate uses of retired oil and gas platforms
View-Data-by-Topic/Finance-and-Investment/Renewable-Energy-Finance-Flows
in the Gulf of Mexico. Ocean Coast Manag 2019;167:52–9. https://doi.org/
(accessed February 25, 2021).
10.1016/j.ocecoaman.2018.10.002.
[4] Department of Social and Economic Affairs of United Nations. The 2030 Agenda for
[26] van Elden S, Meeuwig JJ, Hobbs RJ, Hemmi JM. Offshore Oil and Gas Platforms as
Sustainable Development 2021. https://sdgs.un.org/goals (accessed March 10,
Novel Ecosystems: A Global Perspective. Front Mar Sci 2019;6. https://doi.org/
2021).
10.3389/fmars.2019.00548.
[5] Boshell F, Roesch R, Salgado A, Hecke J. Unlocking the potential of Ocean Energy:
[27] Ou TC, Lu KH, Huang CJ. Improvement of transient stability in a hybrid power
from megawatts to gigawatts. EnergypostEu 2021.
multi-system using a designed NIDC (Novel Intelligent Damping Controller).
[6] The world Factbook. Field listing - coastline. CIAGov 2021. https://www.cia.gov/
Energies 2017;10. https://doi.org/10.3390/en10040488.
the-world-factbook/field/coastline/ (accessed February 20, 2021).
[28] Oliveira-Pinto S, Rosa-Santos P, Taveira-Pinto F. Electricity supply to offshore oil
[7] IRENA. Renewable energy employment by country 2021. https://www.irena.org/
and gas platforms from renewable ocean wave energy: Overview and case study
Statistics/View-Data-by-Topic/Benefits/Renewable-Energy-Employment-by-
analysis. Energy Convers Manag 2019;186:556–69. https://doi.org/10.1016/j.
Country (accessed February 26, 2021).
enconman.2019.02.050.
[8] d’Amore-Domenech R, Leo TJ, Pollet BG. Bulk power transmission at sea: Life cycle
[29] Dallavalle E, Cipolletta M, Casson Moreno V, Cozzani V, Zanuttigh B. Towards
cost comparison of electricity and hydrogen as energy vectors. Appl Energy 2021;
green transition of touristic islands through hybrid renewable energy systems. A
288:116625. https://doi.org/10.1016/j.apenergy.2021.116625.
case study in Tenerife, Canary Islands. Renew Energy 2021;174:426–43. https://
[9] Irena. Innovation outlook: Ocean energy technologies. Abu Dhabi 2020.
doi.org/10.1016/j.renene.2021.04.044.
[10] Trivedi K, Koley S. Performance of a hybrid wave energy converter device
[30] Dincer I, Cozzani V, Crivellari A. HYBRID ENERGY SYSTEMS FOR OFFSHORE
consisting of a piezoelectric plate and oscillating water column device placed over
APPLICATIONS. Joe Hayton: HYbrid Ene; 2021.
an undulated seabed. Appl Energy 2023;333:120627. https://doi.org/10.1016/j.
[31] Nasrollahi S, Kazemi A, Jahangir M-H, Aryaee S. Selecting suitable wave energy
apenergy.2022.120627.
technology for sustainable development, an MCDM approach. Renew Energy 2023;
[11] Li Y, Ma X, Tang T, Zha F, Chen Z, Liu H, et al. High-efficient built-in wave energy
202:756–72. https://doi.org/10.1016/j.renene.2022.11.005.
harvesting technology: from laboratory to open ocean test. Appl Energy 2022;322:
[32] Jahangir MH, Mazinani M. Evaluation of the convertible offshore wave energy
119498. https://doi.org/10.1016/j.apenergy.2022.119498.
capacity of the southern strip of the Caspian Sea. Renew Energy 2020;152:331–46.
[12] Chen W, Lu Y, Li S, Gao F. A bio-inspired foldable-wing wave energy converter for
https://doi.org/10.1016/j.renene.2020.01.012.
ocean robots. Appl Energy 2023;334:120696. https://doi.org/10.1016/j.
[33] Terna S.p.A. Terna - driving energy 2021. https://www.terna.it/it (accessed May
apenergy.2023.120696.
27, 2021).
[13] Foteinis S, Tsoutsos T. Strategies to improve sustainability and offset the initial
[34] Sun P, Xu B, Wang J. Long-term trend analysis and wave energy assessment based
high capital expenditure of wave energy converters (WECs). Renew Sustain Energy
on ERA5 wave reanalysis along the Chinese coastline. Appl Energy 2022;324:
Rev 2017;70:775–85. https://doi.org/10.1016/j.rser.2016.11.258.
119709. https://doi.org/10.1016/j.apenergy.2022.119709.
[14] Hu J, Zhou B, Vogel C, Liu P, Willden R, Sun K, et al. Optimal design and
[35] Reikard G, Pinson P, Bidlot JR. Forecasting ocean wave energy: The ECMWF wave
performance analysis of a hybrid system combing a floating wind platform and
model and time series methods. Ocean Eng 2011;38:1089–99. https://doi.org/
wave energy converters. Appl Energy 2020;269:114998. https://doi.org/10.1016/
10.1016/j.oceaneng.2011.04.009.
j.apenergy.2020.114998.
[36] Fernández-Chozas J, Kofoed JP, Jensen NEH. User guide – COE Calculation tool for
[15] Kluger JM, Haji MN, Slocum AH. The power balancing benefits of wave energy
Wave Energy Converters ver. 1.6. Allborg - Denmark: 2014.
converters in offshore wind-wave farms with energy storage. Appl Energy 2023;
[37] Babarit A, Hals J, Muliawan MJ, Kurniawan A, Moan T, Krokstad J. Numerical
331. https://doi.org/10.1016/j.apenergy.2022.120389.
benchmarking study of a selection of wave energy converters. Renew Energy 2012;
[16] Jiang X, Gao D, Hua F, Yang Y, Wang Z. An improved approach to wave energy
41:44–63. https://doi.org/10.1016/j.renene.2011.10.002.
resource characterization for sea states with multiple wave systems. J Mar Sci Eng
[38] de Andres A, Maillet J, Todalshaug JH, Möller P, Bould D, Jeffrey H. Techno-
2022;10. https://doi.org/10.3390/jmse10101362.
economic related metrics for a wave energy converters feasibility assessment.
[17] Coe RG, Ahn S, Neary VS, Kobos PH, Bacelli G. Maybe less is more: considering
Sustain 2016:8. https://doi.org/10.3390/su8111109.
capacity factor, saturation, variability, and filtering effects of wave energy devices.
[39] O’Connor M, Lewis T, Dalton G. Techno-economic performance of the Pelamis P1
Appl Energy 2021;291:116763. https://doi.org/10.1016/j.apenergy.2021.116763.
and Wavestar at different ratings and various locations in Europe. Renew Energy
[18] López I, Andreu J, Ceballos S, Martínez De Alegría I, Kortabarria I. Review of wave
2013;50:889–900. https://doi.org/10.1016/j.renene.2012.08.009.
energy technologies and the necessary power-equipment. Renew Sustain Energy
[40] Vannucchi V, Cappietti L. Wave energy assessment and performance estimation of
Rev 2013;27:413–34. https://doi.org/10.1016/j.rser.2013.07.009.
state of the artwave energy converters in italian hotspots. Sustain 2016:8. https://
[19] Reikard G. Integrating wave energy into the power grid: Simulation and
doi.org/10.3390/su8121300.
forecasting. Ocean Eng 2013;73:168–78. https://doi.org/10.1016/j.
[41] Nilsson D, Westin A. Floating wind power in Norway: Analysis of opportunities and
oceaneng.2013.08.005.
challenges 2014.
[20] Armstrong S, Mollaghan D, Blavette A, O’Sullivan D. An initialisation methodology
[42] Wang F, Pei Y, Boroyevich D, Burgos R, Ngo K. Ac vs. dc distribution for off-shore
for ocean energy converter dynamic models in power system simulation tools. Proc
power delivery. IECON Proc (Industrial Electron Conf 2008:2113–8. https://doi.
Univ Power Eng Conf 2012. https://doi.org/10.1109/UPEC.2012.6398685.
org/10.1109/IECON.2008.4758283.

20
M. Cipolletta et al. Applied Energy 347 (2023) 121410

[43] Guandalini G, Campanari S, Romano MC. Power-to-gas plants and gas turbines for [71] de Andres A, Guanche R, Vidal C, Losada IJ. Adaptability of a generic wave energy
improved wind energy dispatchability: Energy and economic assessment. Appl converter to different climate conditions. Renew Energy 2015;78:322–33. https://
Energy 2015;147:117–30. https://doi.org/10.1016/j.apenergy.2015.02.055. doi.org/10.1016/j.renene.2015.01.020.
[44] Tugnoli A, Landucci G, Salzano E, Cozzani V. Supporting the selection of process [72] U.S. Department of Energy. Combined Heat and Power Technology Fact Sheet
and plant design options by Inherent Safety KPIs. J Loss Prev Process Ind 2012;25: Series - Microturbines 2016:1–4. https://doi.org/DOE/EE-1329.
830–42. https://doi.org/10.1016/j.jlp.2012.03.008. [73] Reikard G, Robertson B, Bidlot JR. Wave energy worldwide: simulating wave
[45] Cipolletta M, Moreno VC, Cozzani V. Inherent safety assessment for two solar- farms, forecasting, and calculating reserves. Int J Mar Energy 2017;17:156–85.
based fuels production processes: Methanol via co2 catalytic hydrogenation and https://doi.org/10.1016/j.ijome.2017.01.004.
biodiesel from microalgal oil. Chem Eng Trans 2020;82:85–90. https://doi.org/ [74] Ackermann T. Wind Power in Power Systems. John Wiley & Sons, Ltd; 2005.
10.3303/CET2082015. https://doi.org/10.1002/0470012684.
[46] Crivellari A, Bonvicini S, Tugnoli A, Cozzani V. Multi-target Inherent Safety Indices [75] Pinson P, Reikard G, Bidlot JR. Probabilistic forecasting of the wave energy flux.
for the Early Design of Offshore Oil&Gas Facilities. Process Saf Environ Prot 2021; Appl Energy 2012;93:364–70. https://doi.org/10.1016/j.apenergy.2011.12.040.
148:256–72. https://doi.org/10.1016/j.psep.2020.10.010. [76] Penalba M, Ringwood JV. A review of wave-to-wire models for wave energy
[47] IEA, NEA. Projected Costs of Generating Electricity 2015:1–215. converters. Energies 2016;9. https://doi.org/10.3390/en9070506.
[48] Crivellari A, Tugnoli A, Cozzani V, Macini P. Systematic methodology for inherent [77] Hughes SA. Advanced Series on Ocean Engineering: Volume 7 - Physical Models
safety indicators assessment of early design stages of offshore oil & gas projects. and Laboratory Techniques in Coastal Engineering. World Scientific Publishing Co
Chem Eng Sci 2018;67:691–6. https://doi.org/10.3303/CET1867116. Pte Ltd; 1993. https://doi.org/10.1142/asoe10.1142/2154.
[49] Saaty TL. The analytic hierarchy process. New York: McGraw-Hill International [78] Capstone Turbine Corporation. Capstone microturbines 2019.
Book Co; 1980. [79] do Nascimento MAR, Rodrigues LDO, E. C. Dos Santos EEBG, Dias FLG, Velásques
[50] Silva D, Rusu E, Soares CG. Evaluation of various technologies for wave energy EIG, Carrillo R a. M. Micro Gas Turbine Engine: A Review 2014:107–42. https://
conversion in the portuguese nearshore. Energies 2013;6:1344–64. https://doi. doi.org/10.5772/54444 13.
org/10.3390/en6031344. [80] Barelli L, Ottaviano A. Supercharged gas turbine combined cycle: An improvement
[51] OECD. Renewable energy feed-in tariffs. OECDStats 2020. https://stats.oecd.org/ in plant flexibility and efficiency. Energy 2015;81:615–26. https://doi.org/
Index.aspx?DataSetCode=RE_FIT (accessed February 16, 2020). 10.1016/j.energy.2015.01.004.
[52] European Environment Agency. Greenhouse gas emission intensity of electricity [81] Liu Z, Karimi IA. Simulation and optimization of a combined cycle gas turbine
generation in Europe 2020. https://www.eea.europa.eu/data-and-maps/ power plant under part-load operation. Comput Aided Chem Eng 2018;44:2401–6.
indicators/overview-of-the-electricity-production-3/assessment (accessed https://doi.org/10.1016/B978-0-444-64241-7.50395-5.
February 17, 2020). [82] Trapani K, Millar DL. Proposing offshore photovoltaic (PV) technology to the
[53] Metropolis N, Ulam S. The Monte Carlo Method. J Am Stat Assoc 1949;44:335–41. energy mix of the Maltese islands. Energy Convers Manag 2013;67:18–26. https://
[54] Dialyna E, Tsoutsos T. Wave energy in the mediterranean sea: Resource doi.org/10.1016/j.enconman.2012.10.022.
assessment, deployed wecs and prospects. Energies 2021:14. https://doi.org/ [83] Kim TS. Comparative analysis on the part load performance of combined cycle
10.3390/en14164764. plants considering design performance and power control strategy. Energy 2004;
[55] Mørk G, Barstow S, Kabuth A, Pontes MT. Assessing the global wave energy 29:71–85. https://doi.org/10.1016/S0360-5442(03)00157-9.
potential. Proc Int Conf Offshore Mech Arct Eng - OMAE 2010;3:447–54. https:// [84] Li Y, Zhang G, Bai Z, Song X, Wang L, Yang Y. Backpressure adjustable gas turbine
doi.org/10.1115/OMAE2010-20473. cycle: a method to improve part-load efficiency. Energy Convers Manage 2018;
[56] ARPAE. DEXT3R 2021. https://simc.arpae.it/dext3r/ (accessed May 27, 2021). 174:739–54. https://doi.org/10.1016/j.enconman.2018.07.077.
[57] EMODnet physics. ERDAPP 2021. https://erddap.emodnet-physics.eu/erddap/ [85] Boyce MP. An Overview of Gas Turbines. Gas Turbine Eng. Handb. 4th ed., Oxford,
index.html (accessed May 8, 2021). United Kingdom: Butterworth-Heinemann Elsevier Inc.; 2012. https://doi.org/
[58] Gamna V. Eni apre le porte della piattaforma Garibaldi C. AffariitalianiIt 2019:5. 10.1016/C2009-0-64242-2.
[59] Ministero dello Sviluppo Economico. Piattaforme e teste pozzo sottomarine. 2020. [86] U.S. DOE. CHP Technology Factsheet Series: Gas Turbine 2016:4.
[60] Offshore Technology. L8-D field, North Sea 2010. https://www.offshore- [87] U.S. Department of Energy. Combined Heat and Power Technology Factsheet
technology.com/projects/l8dfieldnorthsea/ (accessed September 15, 2020). Series: Gas Turbines 2016:1–4.
[61] NLOG. Dutch Oil and Gas portal 2021. https://www.nlog.nl/en/map-boreholes [88] World Bank Group. Report of the High-Level Commission on Carbon Prices 2017:
(accessed November 10, 2020). 1–69.
[62] Marine Traffic. Platform L11-B 2021. https://www.marinetraffic.com/no/ais/ [89] World Bank, Ecofys. State and Trends of Carbon Pricing 2018. Washington DC,
details/ships/shipid:4795744/mmsi:245651000/imo:0/vessel:PLATFORM_L11B USA: International Bank for Reconstruction and Development - The World Bank;
(accessed July 12, 2020). 2018. https://doi.org/10.1596/978-1-4648-1292-7.
[63] Lira-Loarca A, Ferrari F, Mazzino A, Besio G. Future wind and wave energy [90] U.S. Environmental Protection Agency. Greenhouse Gas (GHG) Verification
resources and exploitability in the Mediterranean Sea by 2100. Appl Energy 2021; Guideline Series: Natural Gas-Fired Microturbine Electrical Generators 2002.
302:117492. https://doi.org/10.1016/j.apenergy.2021.117492. [91] Uijt de Haag PAM, Ale BJM. The “Purple book” – Guidelines for quantitative risk
[64] Sørensen HC, Friis-Madsen E. Wave Dragon 1.5 MW North Sea Demonstrator - assessment. Publ Ser Danger Subst (PGS 3) 2005:237.
Phase 1. 2015. [92] Belton V, Stewart T. Multiple criteria decision analysis: an integrated approach.
[65] Hong Y, Eriksson M, Boström C, Waters R. Impact of generator stroke length on Boston: Kluwer Academic Publishers; 2002.
energy production for a direct drivewave energy converter. Energies 2016;9. [93] Hacatoglu KA. Systems Approach to Assessing the Sustainability of Hybrid
https://doi.org/10.3390/en9090730. Community Energy Systems. University of Ontario Institute of Technology; 2014.
[66] Miguel Sagaseta de Ilurdoz Cortadellas A, Miguel Ángel Guerra Rodríguez B, [94] GSE. Esiti MGP - prezzi 2021. https://mercatoelettrico.org/It/Esiti/MGP/
Raquel Ramos Pereda C, Cuesta Moreno DPD. Preliminary study for the EsitiMGP.aspx (accessed May 27, 2021).
implementation of the “wave dragon” in Gran Canaria, Canary Islands, Spain. [95] SunSet. Storico segno giornaliero. MyTerna 2021. https://myterna.terna.it/
Renew Energy Power Qual J 2011;1:1111–6. https://doi.org/10.24084/ SunSet/Public/Pubblicazioni (accessed May 27, 2021).
repqj09.560. [96] Ministero dello Sviluppo Economico. Ministerial Decree 23 June 2016. Incentive
[67] Pecher A. Performance Evaluation of Wave Energy Converters. 2013. https://doi. for electric energy produced from renewable sources apart from PV. 2016.
org/10.13052/rp-9788792982278. [97] ENTSO-E. Day-ahead prices. Eur Netw Transm Syst Oper Electr 2021. https://
[68] Sørensen HC, Friis-Madsen E. Wave Dragon 1.5 MW North Sea Demonstrator - transparency.entsoe.eu/transmission-domain/r2/dayAheadPrices (accessed May
Phase 1 2014:2012–4. 27, 2021).
[69] U.S. DOE. Combined Heat and Power Technology Fact Sheet Series: Microturbines. [98] Netherlands Enterprise Agency. SDE+ 2016. 2016.
2016. [99] Netherlands Enterprise Agency. SDE+ Spring 2018. 2018.
[70] European Commission. EU ETS Handbook. Clim Action 2015:138.

21