3.1. Materials and Methods
In this section, the authors present a long-term approach that aims to evaluate the electricity costs produced by floating offshore wind farms in Italy, considering effects of climate change on wind energy production and further costs optimization due to development and large diffusion of this technology. The difficulty in conducting studies of this type arises from the multidisciplinary approach that must be applied. In order to set up a comprehensive long-term predictive methodology, it is necessary to consider the interactions between the development of the technology, the costs and the market, the climate, and sociopolitical dynamics (the latter is out of the scope of this study) of the next decades. Forecasting the evolution of all these issues is not simple, given their complexity [
25].
Concerning the costs, in
Section 2 an extensive assessment at 2030 was reported. In this section, the economic benefits beyond 2060, such as long-term progressive decrease of costs and weighted average cost of capital (WACC) due to the learning curve, and reaching large diffusion of floating wind technology, are considered [
26].
Climate change has potential impacts on the whole energy sector, but especially on renewables. Several studies provide risks analyses and predictions of future resource availability [
27]. The focus of the present study is on the change in spatial distribution of wind resources, and possible effects on wind energy productivity and, consequently, on economic sustainability of wind farms in Italian seas. Obviously, a reduction in mean wind speed generally leads to a drop in power production and, accordingly, to an increase in LCoE. Otherwise, a higher resource is favored in terms of economic sustainability.
Current resource distribution reproduced in wind charts and atlases is based on historical data recordings and on climatic models fitting those data. Uncertainties arise because of an incomplete knowledge of atmospheric processes and because of lack of data, especially offshore, where there are very few measurement stations.
Future spatial and temporal variations of meteorological parameters, such as temperature, precipitation, and mean wind speed, are estimated by simulations carried out with regional climate models (RCMs), whose boundary conditions (which include CO
2 emission pathways, as provided by Intergovernmental Panel on Climate Change, IPCC) are derived from global climate models (GCMs). Uncertainties related to these models are often softened by running different simulations constituting an “ensemble” or by validation through comparison with empirical data sets [
25,
28].
Concerning the possible impact of climate change on the production expected from wind power plants in the future, the studies are focused on mapping the wind speed and identifying the areas that, in long-term scenarios, will benefit from an increased wind potential, and areas that will be negatively impacted [
28]. Since wind energy density increases as the cube of the speed, even small variations in wind speed are amplified in the production calculation [
25].
Moreover, to depict a complete frame about economics, the trend of other LCoE drivers, such as CAPEX, OPEX, and WACC were investigated.
In order to carry out this analysis, a parametric tool similar to the one presented in
Section 2 has been developed. For the second generation of floating wind parks to be installed in 2060, the starting base case was the 30-year life cycle—1000 MW reference wind power plant already used in the 2030 analysis, and the central values of the previous cost analysis were applied. Moreover, for the baseline, no wind resource variation was supposed, since reference literature examined in this study do not show significant modifications of wind speed patterns over the area of interest in the next decade.
According to ATLAEOLICO [
21], Italy’s wind power productivity maps, shown in
Figure 2, indicate Apulia, Sardinia, and south Sicily’s coasts as the more attractive regions for installation of new offshore wind plants. In fact, in these areas, the average wind speed is closer to the North Sea or Atlantic one, where higher mean wind speeds occasionally exceed 12 m/s.
Firstly, as an exemplary demonstration, three hypothetical case studies were arranged in these three marine areas, about 40 km from the coastlines, at less than 500 m water-depth. The ATLAEOLICO performance calculator tool was employed to find out the power productivity of a theoretical parametric floating wind farm, given the geographic coordinates of the site and the hub height of the selected wind turbine. This tool is applied in the reference case and for all of the other production evaluations. Wind speed at hub height is automatically extrapolated by the software through vertical gradient wind intensity equation.
Secondly, we introduced parameters to represent long-term scenarios at 2060. To evaluate possible variations in wind resources due to climate change, the authors mainly refer to the above mentioned study [
28] that analyzes the results, and the level of accord of six different simulations performed with three RCM (KNMI-RACMO2, HIRHAM5, RCA3) and four GCM (BCM, ECHAM5, ECHAM5-r3, HadCM3Q16). General conclusions of the paper are that, although some delimited regions could experience an increase in mean wind speed, the wind resource in the Mediterranean area will generally undergo a modest decrease (especially in the Central Mediterranean) or remain constant. A comparable tendency is extrapolated for Italy: models agree on a slight reduction (up to − 0.5m/s) over Sicily and Calabria’s coasts, while there is not an obvious trend for the rest of the Italian seas. This picture is also confirmed by the simulations conducted with a RSE tool, called CLIMED [
29], aimed to assess the impacts of climate change on the electricity system. Even in a scenario with high emissions, the decrease in wind speed on Italian seas is limited to −0.4 m/s and seems concentrated around Sicily. Nevertheless, even though some general patterns may be distinguished, it must be accounted that perceptible discrepancy still exists between long-term projections provided by different sources. Accordingly, a conservative approach was applied in determining long-term resource decrease estimates. Chosen values should be intended as the worst scenario that could incur based on the sources analyzed in this section. We set for case study 1 (Apulia) a null variation of wind speed, for case study 2 (Sicily), due to the scarce level of agreement of predictive models, two −0.25 m/s (less pessimistic) and −0.5 m/s (pessimistic) hypothesis and, finally, for case study 3 (Sardinia), a decrease of −0.25 m/s. The new values of wind speed were re-implemented in ATLAEOLICO and, for each case study, future estimates of wind power productivity were obtained as reported in the following
Table 5,
Table 6,
Table 7 and
Table 8.
Regarding economic inputs, projections of the main parameters, CAPEX and OPEX, at 2060 were extracted from NREL reports [
30,
31], which contain mid and low cost trends, and allowed to develop more or less optimistic scenarios. Distance-from-shore-dependent costs, such as grid connection expenses, were considered in the 2030 scenario. In order to represent more fortunate evolutions to 2060 of the floating wind sector, the WACC was gradually lowered to 2.5% [
24].
Wind farms and turbine sizes are expected to change as well. To describe this increasing trend, we introduced a second reference 2000 MW wind power plant, composed by 100 20-MW turbines. To compute power productivity of these hypothetical turbines, we assumed (in a very rough first approximation) that the power curve is twice the curve of the reference 10-MW turbine. Therefore, performance calculation for this layout was simply run considering twice as many turbines. Another factor was the higher hub height of the supposed 20 MW turbine, which was set at 168 m a.s.l. Wind velocity at that altitude was calculated using the “wind shear” equation—α equal to 0.11 (already discussed in
Section 2).
In order to outline the range of potential evolutions of floating wind LCoE in Italy, data inputs are clustered in different scenarios, highlighting optimistic, medium, and conservative inputs combinations:
“Baseline 2030”, as already reported, considered the performance of 1000 MW reference farm without assuming any change in the resource compared to current values. This constitutes a reference point for following scenarios.
“CC 2060” (climate change at 2060) is a scenario that shows only wind resource rearrangement, without variations on cost and technology variables. Thus, modifying only wind speed inputs, climate change effects were isolated.
“CC + TEC(I) 2060” (climate change + low technology development at 2060) combines both climate change effects on wind velocities and an initial conservative hypothesis about costs and WACC (–10% for CAPEX and –5% for OPEX, WACC reduced from 7% to 5%).
“CC + TEC(II) 2060” (climate change + mid technology development at 2060) examines a more optimistic but still realistic framework in which, besides resource change, favorable reductions of WACC (reduced to 3.75%) and an increase in farm size (2000 MW) are assumed.
“TEC(II) 2060” (technology development II at 2060) simulates the same pattern of inputs of the previous scenario except for the fact that resource variation is not considered.
“CC + TEC(III) 2060” (climate change + high technology development at 2060) finally includes a wind farm operating over the most favorable techno-economic conditions: CAPEX and OPEX are assumed to decline by, respectively, 35% and 50%, compared to the 2030 baseline. WACC is fixed at 2.5%, a value corresponding to mature technology with established market.
3.2. Results and Discussion
The results of this assessment are reported in
Table 5,
Table 6,
Table 7 and
Table 8. Every table contains the input parameters discussed in the previous section, mentioned by classes (geophysical and environmental parameters, wind farm specifications, costs, financial aspects and losses). Note that only some of these parameters plays a role in LCoE calculation. Mean wind speed (whereby gross production and capacity factor are derived), CAPEX, OPEX, and WACC are the main drivers of LCoE, which was variated. LCoE and LCoE percentage variation compared to the baseline for each case study are reported in the last lines of the tables.
The benchmark scenario at 2030 indicates, as in the previous section, a mean LCoE near to 100 €/MWh, a value compatible with the estimates reported in [
11] for floating wind turbines at commercial stage. The baseline highlights differences in wind resource between case study sites, which recur along successive scenarios. As noticeable in ATLAEOLICO wind speed maps, all selected sites present attracting wind conditions. In the Apulia site, the resource (7.79 m/s) is slightly lower than the other sites, but is still assessed as profitable in terms of electricity production. Marine surroundings of Sicily present more suitable wind velocities while Sardinia’s west coast exhibit even better wind conditions. Sicily and Sardinia sites currently present, respectively, wind speeds of 8.05 m/s and 8.92 m/s. Some projects of wind farms for these zones have already been submitted, but are still waiting for authorization [
32].
For clarity’s sake, the results are summarized in the following bar plot (
Figure 3), which shows the differences between case studies along the different scenarios.
Specific scenario related comments are:
In CC 2060, following the approach used by [
25], the first scenario at 2060 was assembled by focusing only on the resource change. Costs and technological parameters were not then modified in order to isolate climate change effects on capacity factors and consequently on LCoE. Results show an increase smaller than 10% in the LCoE at 2060 for case study 2 (Sicily) under a −0.50 m/s speed decrease hypothesis, and an increase smaller than 5% in the LCoE at 2060 for case studies 2 (Sicily) and 3 (Sardinia) under a −0.25 m/s speed decrease hypothesis. For case study 1 (Apulia) no resource variation was applied; therefore, LCoE did not experience any variation. Given these results, obtained without considering any technological improvement or costs decrease, it is inferable that climate change could play an antagonistic role in reducing LCoE of floating wind technology, although its impacts does not appear to be dramatic.
CC + TEC(I) 2060 shows that, even when a conservative hypothesis of costs and WACC reduction is applied, the LCoE positive trend of the previous scenario is largely balanced. Decrease in LCoE, which involves both resource change and a low technological development contributes, is placed between −10.9% (case study 2 under −0.50 m/s hypothesis) and −18.4% (case study 1).
In the CC + TEC(II) 2060 scenario, a more optimistic pattern of economic and technological inputs is applied, including higher turbines, which access greater wind speeds. LCoE reductions is more evident: Apulia’s site, which does not suffer climate change effects, experiences a reduction of −28.4% in LCoE. Sicily shows reductions of −25.4% (optimistic hypothesis) and −22.1% (pessimistic hypothesis), while Sardinia LCoE drops by 25.6%. Absolute values of LCoE decrease to about 70–80 €/MWh range.
In the TEC(II) 2060, the deviation between LCoE decreases if resource variations are taken or not into account is negligible and obviously maximal in the −0.50 m/s hypothesis for case study 2 (−6.2%). This, fortunately, confirms again the weak effects of climate change on wind energy production in the analyzed sites.
Lastly, in the hyper-optimistic scenario, CC + TEC(III) 2060, very low LCoE values (below 50 €/MWh) are reached by considerable percentage reductions (−55% to −60%) compared to 2030. Although very optimistic, this scenario could be achieved if a large diffusion of floating wind technology will take place in the next decades following, i.e., the learning curve of the onshore wind energy sector.