The energy sector is the largest contributor to the global greenhouse gas (GHG) emissions [1
]. This has led to the development of renewable energy resources with lower carbon footprints. Wind energy, for example, became the second largest power generation source in Europe, just after gas, with an installed capacity of 169 GW in 2017 [2
]. However, wind energy comes with intermittency and unpredictability, due to variable wind conditions [3
], requiring substantial support technologies for load balancing. Wind power is also perceived as noisy and visually interfering [4
]. Tidal streams, on the other hand, are easily predicted [5
], since they are generated by the rise and fall of the sea level due to gravitational forces of the moon, sun, and earth. Yet, an important factor limiting the application of tidal streams is their relatively slow speed, and the problem associated with constructing turbines with a sufficiently low cut-in speed for starting electricity generation. The first generation tidal farms require 2.5 m/s of current to start [6
], making them economically viable in restricted geographical places.
According to Clarke, et al. [7
], tidal current power stations could be used as base load technologies if these stations are distributed in an efficient way. Furthermore, wave and tidal stream energies have the potential to meet up to 20% of the UK’s electricity demand, representing a 30 to 50 GW installed capacity [8
]. However, only a small portion (57 MW) of tidal capacity potential is expected to be installed in Europe by 2020 [9
Given the promising potential of large-scale adoption of tidal current power generation technologies, the development, though, should be taken in a precautious manner. This development could lead to negative environmental impacts during the full life cycle (material extraction and production, manufacturing, use, and end-of-life) of the power plants. The life cycle assessment (LCA) has been widely used to assess the environmental impacts of energy technologies, where a literature review for LCAs of wind power, for example, analyzed more than 30 studies already in 2012 [10
]. The majority of the emissions associated with the life cycle of wind electricity generation occurred in the production of various components.
Such studies are still scarce for marine energy technologies in general, and tidal current ones in specific. Douglas, et al. [11
] were the first to study the environmental profile of a tidal energy technology (SeaGen) using LCA. While the study aimed at including the details of all the life cycle phases, the operations and maintenance (O&M) phase was not fully modelled. The replacement of the powertrain was not included, which resulted in a minor contribution from the O&M phase to the total impact. In fact, Arvesen, et al. [12
] found that previous LCA studies on offshore wind farms, which require similar operations to tidal energy farms, usually underestimated the impacts of offshore operations (through ships) and spare parts. According to that study, offshore operations, including installation and maintenance, amounted to 28% of total greenhouse gas emissions, of which half were from maintenance alone. Uihlein [13
] studied the impact of different tidal technologies by performing LCA on 83 tidal devices retrieved from the Joint Research Centre’s (JRC) ocean energy database. Although the database includes various information such as weights, dimensions, and material, a lot of the data had to be estimated to fill the gaps. For example, 90% of the inventory data for structural components in horizontal axis turbines were estimated based on averages of different technologies. Furthermore, several parameters were excluded, such as the electrical losses from the connections (cable, connector, and hub), the substations, and the spare parts.
Deep Green is a tidal technology which solves a major issue in other technologies. Having its cut-in speed at only 1.2 m/s, it is capable of functioning in slower currents, while operating at depths between 60 and 120 m [14
]. The potential impacts of the Deep Green on local marine ecosystems have already been studied [15
], where it was found that phytoplankton dynamics and benthic communities were unlikely to be altered by its operation. A preliminary LCA was also done on Deep Green at an early planning stage [17
]. However, prototypes with improved design have been deployed after that, and better inventory data could be obtained. Thus, an updated LCA study on this technology would complement other studies and give a more accurate and comprehensive picture of its environmental profile.
The purpose of this study was to assess the environmental impact of an array of 24 DG500 prototypes at the current state of development through the entire life cycle. A prospective LCA approach was applied on a 12 MW rated power array of DG500, planned to be established 8 km off the shore of Holyhead, UK. Comparisons are also made with previous assessments of other energy technologies with a focus on tidal technologies.
The Array of Powerplants
represents a schematic of the model of the tidal energy array assessed in this study, and Table 1
contains its main system specifications [14
]. Each DG500 kite was connected to its gravity base foundation by a tether (containing the umbilical system) through the bottom joint. Every group of six DG500s was connected to a tidal marine substation (TMS) located on a floating buoy, which was connected to the onshore substation. It was assumed that the array would consist of 24 installed DG500s and four TMS buoys, one of which would also act as a substation. Furthermore, an additional two spare DG500s were assumed to be located onshore to facilitate efficient service and maintenance. The substation in turn was connected to the UK national electricity grid. The array was assumed to have a lifetime of 25 years.
2. Materials and Methods
This study was a prospective LCA [18
] of a 12 MW array of 24 DG500 prototypes developed by Minesto AB [14
]. The LCA was done according to the ISO 14040:2006 standard [19
], covering scoping, life cycle inventory (data gathering), life cycle impact assessment, and interpretation of results. Furthermore, it followed the technical scope defined in the Product Category Rules (PCR) for electricity production technologies, as defined in the Environmental Product Declaration (EPD) System [20
]. The system was modelled using openLCA software [21
]. Most of the information needed for the downstream processes was obtained from Minesto and connected to the upstream data provided by the ecoinvent database version 3.3 [22
]. Data gaps, when found, were filled from other LCA studies or available EPDs. ReCiPe version 1.11 [23
] was used for impact assessment.
A brief literature review on available LCA studies of tidal power technologies was performed after commencing the LCA, to investigate the effect of various modelling parameter choices on the robustness of the results. Although most studies reported the results in terms of global warming potential (GWP), additional impact categories were applied to the DG500 array to enable a comparison with a study by Hertwich, et al. [24
], covering a wide set of different energy generation technologies, and a study by Uihlein on tidal technologies [13
2.1. LCA of the DG500 Array
2.1.1. Goal and Scope
This study aimed to assess the environmental impact of an array of tidal current power plants (Deep Green) planned to be installed at the Holyhead Deep site in Anglesey, UK. Since the power plant and the array were still in the prototype phase, assumptions about some technical requirements and performance metrics were taken. Different scenarios were covered to overcome these inherent uncertainties. One such important aspect was the power output, which was set to a projected hypothetical base scenario of 2 GWh/yr per kite at the selected site and an optimistic case of 3 GWh/yr.
The functional unit was chosen as one kWh of electricity delivered to the consumer connected to the UK grid. The study included the cradle-to-grave production and operation of the entire array, covering the extraction and production of raw materials, transportation, manufacturing, installation, operation and maintenance, and end of life (EOL). It also included the production of factories for the processing of materials, the manufacturing of subsea export cables, the building of an onshore substation and maintenance center, and the UK grid.
This is a site-specific study, which affects the natural conditions for the operation of the DG500, such as the tidal current speed. The site also affects the transportation distances for the materials, and the utility electricity used for the production. The grid distribution to the final user was based on the UK national grid. The raw materials, however, were produced in both the UK and Sweden; thus, utility electricity and transport distances were modelled accordingly. The expected lifetime of a DG500 is 25 years. However, the model did not take into account future technical developments (including the background system).
2.1.2. Life Cycle Inventory Analysis
The inventory analysis was based on the system’s flow chart (Figure 2
). Inventory data and processes used in the openLCA are provided in the Supplementary Materials
Kite and Umbilical System
The main component of the DG500 tidal current power plant is the tidal energy converter (“kite”). The kite consists of a wing, a turbine, a nacelle, a rudder, struts, and a top joint that connects to the tether (Figure 1
). The wing has a wingspan of 12 m and is used to create lift and propulsive force for the kite. The wing is a composite structure with metal inserts for attachment to the nacelle and the struts. Guided by the rudders, the kite sails the ocean current in a continuous infinity (∞) trajectory.
The tidal current energy is collected by a 1.5 m diameter turbine attached to the nacelle. The nacelle is 9 m long with a 0.9 m diameter steel housing that encapsulates the electronic equipment, including a generator, a gearbox, and two converters. This system transforms the kinetic energy of the turbine to electrical energy which is transferred through the umbilical (cable). The nacelle also includes sensors and electronics for controlling the rudder and providing kite performance control data. The front struts connect the wing to the top joint and transfer most of the tether load. The top joint connects the kite to the tether system. The kite measures 9.8 m from the top of the rudder to the top joint.
With 24 kites in the array, two additional complete kites are needed for swap-out during maintenance, totaling 26 kites produced for the power plant.
The tether system consists of tether fairings, a tether rope, and an umbilical cable including power and communication signal cables. Each tether is between 80 and 120 m in length. The rope carries the pull force of the kite. The tether system connects to the kite at the top joint and to the gravity based foundation (GBF) at the bottom joint. The tether fairing, made of various plastic materials, covers the rope and cables, while also reducing the flow resistance. Manufacturing of the tether involves significant energy demand during the polyurethane curing.
The data for the kite materials and assembly were mainly provided by Minesto. Additional data for the converter and the generator were taken from ABB EPDs of the ACS 600 frequency converter [25
] and DMI type DC machine [26
], respectively, both scaled to match the onboard components. Data (mainly VOC emissions) for the paint were taken from the Jotun data sheet [27
], and from a carbon fiber LCA by Romaniw [28
Gravity Based Foundation
The GBF anchors the kite to the seafloor. Three different GBF designs have been assessed. The first GBF design (base scenario) is a concrete block with steel reinforcement. The other two alternatives are “Bsteel”, a steel foundation consisting of a steel-based center node with 4 steel mooring chains and anchors, and “Bhybrid,” a hybrid foundation similar to the steel foundation; however, the center node is based on a concrete block with steel reinforcement.
Tidal Marine Substation and Sub-Hub
The tidal marine substation (TMS) was used to receive the electric energy from the kites (6 kites per TMS in this system) through the kites’ umbilical cables, and to step-up the voltage to 33 kV through a transformer. The TMS buoy is anchored with polyester mooring lines to a clump weights made of recycled steel. The transformer is connected to the sub-hub, which in our case was one of the TMSs, through a TMS–TMS cable. The sub-hub is connected to the onshore station through a subsea export cable. This is illustrated in Figure 3
. The transformers used were modelled according to the ABB EPD for a large distribution transformer [29
] and adjusted to the power levels of the DG500 devices.
There are four types of cables that transfer the electricity from the kite to the grid: the umbilical cable, the TMS–TMS cable, the export cable, and the onshore cable (Figure 3
The 500 m, low voltage umbilical cable connects to the kite at the top joint, runs through the tether, and continues beyond the bottom joint on the seafloor to the TMS. The 33 kV TMS-TMS cables connect three of the TMSes to the sub-hub. (The sub-hub is itself a TMS, and hence needs no additional TMS–TMS cable.)
There are currently three options concerning how the 33 kV subsea and onshore cables are going to be installed. For this study, the shortest offshore path was assumed with an 8000 m subsea export cable and a 4500 m onshore cable.
Data for the cables are based on the technical data sheet from Nexans (2013) [30
The main purpose of the onshore substation is to connect the power plant to the national grid. Various monitoring activities also take place there. Since the voltage is already at grid level after the TMS, no further transformation is needed. Next to the substation, there is a building for the maintenance of the kites, consisting of 2 bays for the current 12 MW array. The houses were modelled with the dimensions provided by Minesto and using a building dataset from the ecoinvent database. The electrical components (the reactor and earthing transformers) were modelled according to the ABB EPD for large distribution transformer [29
Transports and Power Plant Construction
The transports of different parts of the DG500 from the manufacturing sites to the Holyhead site were modelled by train, lorry, or ship, according to the location of the supplier of each component (see Supplementary Materials
). In the on-site construction phase, different vessels are used to install different parts. This includes towing and installing kite foundations, deploying the tether with bottom joint and umbilical, deploying the TMS, and deploying the export cable. Assumptions on standby times for the vessels, where diesel consumption was less, were also included. Vessels include different tugboats, multicats, and ships. This phase was modelled by the diesel combustion used by the vessels according to Jivén, et al. [31
]. The total diesel needed for the complete array was approximately 820 tonnes.
Operation and Maintenance
The kite is expected to start moving with the tidal current, but when this current is too low, reserve power stored in a battery (modelled in ecoinvent) is used to get it started. Once started, the kite should be able to move by itself with the tidal current. The maintenance was also modelled based on the maintenance scheme provided by Minesto. Diesel powered multicat vessels are mainly used to perform routine and non-routine inspections and maintenance of the kites, buoys, and cables. The diesel needed for all the maintenance trips across the lifetime of the system is around 1160 tonnes. More information about the vessel operations (function of each vessel, time spent in water, idle time, and diesel consumed) is found in the Supplementary Materials
. The maintenance phase also includes the production of replacement parts and transportation of those parts to the site. The assumptions of replacement rates and lifetimes for different parts were modelled according to Table 2
. All other parts were assumed not to need replacement during the 25 year life time of the power plant.
A scenario was also assessed where the life-time of the tether was doubled to 10 years to assess how this would affect the environmental performance. This could be done by coating the tether with protective material, for example. The onshore maintenance building was modelled as part of the onshore substation. The final step in the operation is the connection to the grid, which was modelled according to the UK grid in ecoinvent (including grid and transformation losses and the infrastructure of the grid).
The dismantling of the power plants was modelled as diesel used to remove different parts based on estimations by Minesto of vessel operations needed. The total diesel needed for the dismantling is around 420 tonnes. This phase also includes the transportation of different parts to different disposal facilities, which is assumed to be done using lorries for an average distance of 100 km.
Modelling the waste management of novel offshore energy technologies requires assumptions, since almost no existing similar power plants (offshore wind farms for example) have yet reached their respective end-of-life stage [32
]. For the DG500 tidal array, all the iron (including steel) is assumed to be recycled at a rate of 95% and copper at a rate of 90%, in line with Haapala and Prempreeda [33
]. The only exceptions are the steel used for the clump weights and ballast anchor shackles for the TMS, and ballast steel in the Steel GBF scenario, which is not assumed to be recycled due to high uncertainty of recyclability of the lower quality steel made from scrap. The recycling was modelled as end-of-life recycling with avoided burden [34
], crediting the system by avoiding production of virgin material, and including the impact of waste collection and treatment. Other waste materials are assumed to either be sent to landfills or incinerators, but no credits (or burdens) have been modelled due to the uncertainties.
Since the GBF contributes significantly to GWP, a scenario was defined where the concrete GBF was re-used once (i.e., the GBF life-time was extended to 50 years) in another system, based on suggestions by Andersen, Eriksson, Hillman, and Wallhagen [32
]. Given that GBF has passive, durable components, extending its lifetime will most likely contribute to environmental benefits with no trade-offs [35
Sensitivity and Scenario Analysis
To study the robustness and sensitivity of some parameters on the results, seven alternative designs were modelled. Those included extending the lifetimes of impactful components, such as the tether and the gravity base foundation (BT and BTR), different gravity base foundation alternatives (Bsteel and Bhybrid), different power outputs (optimistic), and different site locations (favorable site). In a favorable site, the power rating of the kite is higher due to faster current speeds; thus, less kites in total are needed. Besides, such a location would be closer to the shore, and no TMSes are needed to connect the kites, where each kite is directly connected to the onshore substation. Table 3
summarizes the different scenarios analyzed in this study.
2.2. Literature Review
A brief literature review was performed to analyze previous LCA studies of tidal energy technologies. The keywords used for the search were (“LCA” OR “life cycle assessment” OR “life cycle analysis”) AND (“tide” OR “tidal”) using Scopus database on January 3, 2019. That returned 29 articles, three conference papers, two review papers, and one article in press. Of those, only 10 studies were related to tidal energy generation, and only six were chosen for analysis in this study, as summarized in Table 4
. Of the 10 related articles, Khare [36
] was excluded because it was a hybrid solar-tidal system, and it was not possible to separate the impacts of each part. Uihlein [13
] is an extensive assessment of available tidal technologies, but the average of the technologies was presented, and data were taken from a database. This study was not part of the literature review due to the aggregation of data of different tidal turbines into main categories, but used later in the discussion of the results (Section 3.3
). Amponsah, et al. [37
] and Walker and Howell [38
] performed comparisons of different technologies, but were excluded, as they used results directly from other LCA studies [11
] which are included in our review. Other studies had the LCA performed on the same technologies, but with varying scopes and different results; thus, they were included in the table (SeaGen in [40
] and [11
]; Severn barrage in [39
] and [41
]). Although some studies included other impacts in addition to GWP, only GWP was used for the literature review to keep it consistent. The GWP impact from the base case of the DG500 array study done in this paper is included in Table 4
This study assessed the environmental impact of tidal power generation by performing a prospective LCA of an emerging tidal current technology (Deep Green) based on the Holyhead Deep site in Anglesey, UK. The base-case 12 MW array of DG500 prototypes had a GWP of 26.3 g CO2-eq/kWhe. The processes contributing the most to the various emissions were the production of the foundations, production of spare parts, and the maintenance operations. EOL recycling also lead to significant reduction of net impacts.
The results were compared to previous LCA studies of related energy technologies. It was seen that tidal technology has much less impact than fossil-based technology, and is in the same range as other renewables. Thus, from a life cycle perspective, this technology was found to be environmentally competitive. The comparison with other studies on tidal technologies revealed that the maintenance phase is often neglected, while our results indicate its significance. Furthermore, the EOL phase is rarely clearly expressed, which hinders the possibility of reproducing the study and having fair comparisons.
The environmental impact of the power plant (to the point of delivery to the grid) is essentially linearly dependent on the capacity factor. The actual energy output depends on local conditions, and in this case the capacity factor was assumed to be 46% (2 GWh/year per DG500 kite) for the base case scenario based on prototype testing and tidal current conditions measurements at the Holyhead site. A set of scenarios was also modelled with variations in design options and various sensitive parameters. The results show moderate variations of net results, with up to 16% (22.0 g CO2/kWhe) lowered impact compared to the base case in all scenarios assuming the same capacity factor.