*2.3. Sustainability Factors of Power Generation Technologies*

The sustainability assessment carried out in this investigation relies on a series of emission, economic and employment factors defined for the life cycle of each of the generation technologies that

compose the national electricity mix. This section describes the methodologies followed to define these factors.

#### 2.3.1. Environmental Dimension

Environmental emissions associated with individual power generation technologies were extracted from Ecoinvent v3.1 [30]. The inventory data in these datasets cover the following life cycle stages: (i) extraction and processing of raw materials employed in the construction of power generation infrastructures, (ii) construction of power plants, end of life of construction materials, extraction and processing of fuels (where required); and (iii) operation of power plants and power transmission. When more than one dataset was available for any given technology, a weighted combination of the situation describing the Spanish electricity system was employed. Since no background data was available for CSP plants, and due to the fact that its contribution to the Spanish electricity system is limited (4.0% in 2015), the emissions associated with this technology were not considered.

Environmental impact assessment calculations were carried out using the ILCD 2011 Midpoint+ method [31], except for the human toxicity category for which the ReCiPe 2016 Midpoint (H) v1.03 method [18] was used. This latter method was favoured over ILCD in the human toxicity category as it provided an aggregated approach that included both cancerous and non-cancerous effects.

The environmental categories and the impact units considered in this investigation include the global impacts: climate change (kg CO2 eq), fossil depletion (kg oil eq) and ozone layer depletion (g CFC-11 eq), and the more locally focused human toxicity (kg 1.4 DB eq), terrestrial acidification (kg SO2 eq) and photochemical ozone formation (kg NMVOC eq).

#### 2.3.2. Economic Dimension

The levelized cost of electricity (LCOE) was employed to evaluate the economic sustainability of the Spanish electricity. This indicator has a life cycle approach that is calculated by dividing the discounted cost of power generation (including investment, operation and maintenance, fuel expenditures and decommissioning) by the discounted rate of power generation, as shown in Equation (1):

$$\text{LCOE}\left(\frac{\text{€}}{\text{MWh}}\right) = \frac{\text{discounted lifetime costs}}{\text{discounted power generation}} = \frac{\sum\_{t=1}^{n} \frac{I\_t + M\_t + F\_t}{(1+r)^t}}{\sum\_{t=1}^{n} \frac{E\_t}{(1+r)^t}} \tag{1}$$

where *It*, *Mt* and *Ft* represent investment, operations and maintenance and fuel expenditures in the year *t*, and *Et* represents the power generated in the same year *t*. The value *r* represents the discount rate assumed for the power generation project and *n* its expected lifetime.

The LCOE considered for each of the technologies considered in the Spanish electricity mix were obtained from the International Energy Agency [20] for scenarios prior to 2016. LCOE values were calculated assuming a discount rate of 7.0% and had a national specificity. In cases where this information was not available for Spain (e.g., coal, natural gas combined cycle and nuclear), the cost values were calculated as the average of those applicable to countries within the European Union. Additional information about other key parameters (e.g., technology type and lifetime, average capacity factors) employed to calculate the LCOE may be found in [20].

The future cost of power generation technologies is a matter of debate [32–37]. For the purpose of this investigation, a dynamic approach has been applied based on a series of factors applicable to the reference costs proposed by the International Energy Agency [20]. The transformation factors used for the period 2015–2030 were those proposed by [33] as follows: coal (−5.43%); natural gas (+46.02%); nuclear (+9.51%); hydro (−27.49%); wind (−54.30%); PV (−55.28%) and CSP (−56.95%). Reliable transformation factors for the period 2015–2050 were only available for wind (−69.93%) [34] and PV (−64.22%) [35], which are the most dominant technologies in all the 2050 scenarios (except for ST, which incorporates a high proportion of natural gas). In the absence of dependable data for other

technologies, the costs assumed for nuclear, natural gas and hydro in 2050 were the same as in 2030. In view of past trends (increasing costs of fossil resources and reduced costs for renewables), these assumptions are likely to represent an underestimation in the cost of natural gas and nuclear power.

Figure 6 shows the LCOE applied to different power generation technologies in Spain, according to the procedures described in [20]. The figures show that CSP, waste incineration and biomass have the highest costs. The cost of renewables (wind, PV and hydro) is comparable to conventional fossil fuels and nuclear energy, with coal power being the cheapest. The cost of fossil technologies is dominated by the operation phase, due to the expenses associated with the extraction and processing of fuels, while the cost of renewables is dominated by the construction of the infrastructures allocated to the capital costs. Certain technologies (biomass, CHP, biogas) benefit from heat credits due to the combined generation of power and thermal energy. To avoid the results being affected by international currency policies, a fixed exchange rate was used to convert monetary data published by IEA from USD to Euro. The exchange rate considered was the average value for the core assessment dates (2010–2015) as reported by the European Central Bank at 1 USD = 0.77 €.

**Figure 6.** Economic performance of different technologies for power generation, as applicable to Spain in 2015 in terms of LCOE (adapted from [20]).

This economic analysis does not take into consideration external costs in the form of carbon taxes or carbon emission credits. The incorporation of these levies would be particularly detrimental to the economic interest of the scenarios with a higher contribution of fossil technologies.

#### 2.3.3. Socio-Economic Dimension

The socio-economic performance of the power system in Spain was evaluated using the direct employment generated by the technologies participating in the electricity mix as the indicator. The methodology employed was published by the Institute for Sustainable Futures at the University of Technology Sydney (ISF-UTS) and follows a life cycle approach that takes into consideration four stages: extraction of raw materials and manufacturing of components; construction and installation of additional capacity; operation and maintenance of power plants; and extraction and refining of fuels (where necessary).

Data from the original report published in 2010 [38] was used to quantify employment in the Spanish electricity sector prior to 2010, data published in a subsequent update from 2012 [39] was used to quantify the period 2012–2013 and data from the latest report of 2016 was used to quantify the period 2014–2016. As technology becomes more mature, employment requirements decrease. The employment reduction factors proposed by [21] were used to quantify the situation in the projected scenarios of 2030 and 2050. This methodology takes into consideration the geographic location of the energy projects and time factors that account for expected deviations in future scenarios (due to learning curves and economy of scale). Figure 7 shows the employment factors used for the calculation of jobs in the Spanish electricity systems, as extracted from the references cited above.

**Figure 7.** Employment factors used to quantify direct jobs created from different power generation technologies (adapted from [21,38,39]).

The construction periods selected for different technologies in these calculations were as follows: 10 years for nuclear power plants, five years for coal, two years for natural gas, oil, biomass, hydroelectric, wind, CSP and combined heat and power (CHP), and one year for PV, as reported by [21].
