See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257679862

Life cycle assessment of wind power: Comprehensive results from a state-of-

the-art approach

Article in The International Journal of Life Cycle Assessment · January 2012

DOI: 10.1007/s11367-012-0445-4

CITATIONS READS

70 2,799

2 authors, including:

Peter Garrett
Vestas Wind Systems A/S
1 PUBLICATION 70 CITATIONS

SEE PROFILE

All content following this page was uploaded by Peter Garrett on 02 October 2015.

The user has requested enhancement of the downloaded file.

Life cycle assessment of wind power:
comprehensive results from a state-of-the-
art approach

Peter Garrett & Klaus Rønde

The International Journal of Life

Cycle Assessment

ISSN 0948-3349

Int J Life Cycle Assess

DOI 10.1007/s11367-012-0445-4

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer-
Verlag. This e-offprint is for personal use only
and shall not be self-archived in electronic
repositories. If you wish to self-archive your
work, please use the accepted author’s
version for posting to your own website or
your institution’s repository. You may further
deposit the accepted author’s version on a
funder’s repository at a funder’s request,
provided it is not made publicly available until
12 months after publication.

1 23
 Author's personal copy
Int J Life Cycle Assess
DOI 10.1007/s11367-012-0445-4

LCA FOR ENERGY SYSTEMS

Life cycle assessment of wind power: comprehensive results

from a state-of-the-art approach
Peter Garrett & Klaus Rønde

Received: 27 February 2012 / Accepted: 13 May 2012

# Springer-Verlag 2012

Abstract eutrophication potential 3.7 to 4.5 mg PO4-e, freshwater

Purpose The article presents the method and results of the life aquatic ecotoxicity 100 to 130 mg DCB-e, global warming
cycle assessments (LCAs) of the Vestas' 2-MW GridStrea- potential 7 to 10 g CO2-e, human toxicity potential 1,150 to
merTM wind turbines and outlines the state-of-the-art approach 1,400 mg DCB-e, marine aquatic ecotoxicity potential 1,100
adopted. For more than 10 years, Vestas has prepared LCAs of to 1,300 g DCB-e, photochemical oxidant creation 4 to 5 mg
wind power. However, since 2010, a step change in compre- ethene, terrestrial ecotoxicity potential 19 to 24 mg DCB-e,
hensiveness has been employed, for example, conducting the return-on energy 8 to 11 months and recyclability 81 to
LCA to individually assess all components within a wind 85 % of turbine mass.
turbine (being composed of around 25,000 parts). Being equipped with extensive facts and comprehensive
Methods Three LCAs have been conducted with the 2-MW LCA models provides Vestas the basis to further integrate
GridStreamerTM turbines in accordance with ISO 14040/44 environmental considerations into product marketing, de-
and critically reviewed by an expert. The goal was to eval- sign and research, procurement and to deliver transparent
uate potential environmental impacts and other non-impact information to stakeholders.
indicators per kilowatt hour of electricity generated for a Conclusions Overall, the article presents a case study of the
‘typical’ 50-MW onshore wind plant. LCA approach used to assess the potential impacts of 2-MW
The LCAs assessed all life cycle stages and were built GridStreamer™ turbines based upon comprehensive product
using GaBi DfX software. A significant quantity of primary knowledge and represents a state-of-the-art approach to LCA
data were gathered, for example, covering over 100 Vestas' modelling of wind power. The article discusses further appli-
sites for manufacturing, sales and servicing, as well as cations of LCA internally to direct product improvement and
establishing turbine use-phase performance (i.e. electricity for external communications and also highlights the LCAs'
generation, servicing, etc.) based on over 20,000 monitored aim to improve transparency and robustness of previous LCAs
wind turbines around the world, covering around 20 % of of wind power.
the current worldwide installed capacity.
Results and discussion The baseline results show that per Keywords Global warming . LCA . GWP . Recyclability .
kilowatt hour of electricity generated by the 2-MW Grid- Return on energy . Turbine . Wind
Streamer™ turbines have the following baseline perfor-
mance: ADP elements 0.44 to 0.58 mg Sb-e, ADP fossil
0.10 to 0.13 MJ; acidification potential 37 to 45 mg SO2-e, 1 Introduction

Public awareness and concern for the effects of climate

Responsible editor: Matthias Finkbeiner
change and other environmental impacts has dramatically
P. Garrett (*) : K. Rønde increased in recent years. Due to wind energy's significantly
Vestas Wind Systems A/S,
lower environmental burdens, compared with fossil
Hedeager 44,
8200 Aarhus, Denmark fuel-based electricity generation, wind is well placed to
e-mail: [email protected] contribute towards the global drive to help mitigate the
 Author's personal copy
Int J Life Cycle Assess

effects of climate change and other potential environmental This article presents the LCA results for the 2-MW Grid-
impacts. StreamerTM turbines (Vestas 2011a, b, c) and discusses how
This article presents the life cycle assessment results for LCA considerations are being integrated into Vestas' business
the 2-MW GridStreamerTM turbines and underscores the systems. The 2-MW GridStreamerTM turbines represent a
approach adopted by Vestas Wind Systems A/S to help fulfil modern technology, geared-drive turbine which, in the
their on-going sustainability agenda and to provide publicly Mk 8 version, integrates the GridStreamerTM technology. The
available and transparent assessments of wind power. Vestas GridStreamerTM technology allows higher grid stability by
is the world's largest wind turbine manufacturer selling both synchronising power production with grid requirements. The
onshore and offshore turbines for electricity generation, 2-MW platform of turbines is installed globally, consisting of
having installed more than 46,000 wind turbines in 66 over 9,300 turbines (or 18.6 GW), which represents around
countries on six continents with a total installed capacity 9 % of the current worldwide-installed wind capacity (WWEA
of about 50 GW, representing around 25 % of total worldwide 2011). As such, the 2-MW platform represents a substantial
installed wind capacity (WWEA 2011). Vestas operates over proportion of modern-installed wind capacity. The LCAs pre-
100 sites around the world for manufacturing and sales and sented in this article build upon the entire bill of materials for
offers installation, repair and maintenance services to its the turbine, as well as including all main components of the
customers. complete power plant (up to existing grid connection), repre-
For more than 10 years, Vestas has prepared life cycle senting current technology for wind plant site connection and
assessments of various wind power plants (Elsam 2001, electricity transformation.
2004; PE 2011a; Vestas 2006a, b, 2011a, b, c). The life
cycle assessments (LCAs) are conducted in accordance with
ISO standards (ISO 2006a, b) and assess all main parts of 2 Methods
the different wind power plants over the entire life cycle.
In the recent article by Davidsson et al. (2012), which 2.1 Aim and scope
reviewed 12 available life cycle studies of wind energy systems
with the aim to provide a comparative analysis of current LCA Three LCAs of the 2-MW GridStreamerTM turbines (Vestas
practice, one of the key recommendations is to mitigate the lack 2011a, b, c) have been conducted in accordance with ISO
of transparency regarding LCAs of wind power, relating to 14040/44 standards (ISO 2006a, b) and have been critically
underlying assumptions, calculations and conversions, as well reviewed by an external expert. This indicates the objectiv-
as purporting to international collaboration to agree on LCA ity of the studies and compliance with LCA standards. The
approaches and methods. The LCAs presented in this article reviewer additionally indicated the high level of maturity of
were not included in that review, although one previous LCA the studies performed and the use of state-of-the-art methods
conducted by PE International of the V112 3-MW turbine was (Vestas 2011a, b, c).
assessed (PE 2011a). This article helps to address some of these The goal of the LCAs was to evaluate the potential envi-
issues, both by outlining the LCA methodology undertaken in ronmental impacts associated with the production of electric-
the present LCAs, showing the on-going improvement of ity from a 50-MW onshore wind plant comprised of 2-MW
LCAs of wind power, and by providing a high degree of GridStreamer™ turbines. The studies assess a ‘typical’ wind
transparency of externally reviewed studies that are available plant layout based on Vestas' global supply chain. The system
publicly (Vestas 2011a, b, c). Further consideration is given boundary of the wind plant assessed is shown in Fig. 1.
throughout the article to some of the outcomes of Davidsson et The LCAs assess all stages in the life cycle from cradle to
al. (2012) in relation to the present LCAs. grave, including raw materials; production of all parts of the
Since 2010, Vestas has adopted a step change in compre- wind plant; Vestas' manufacturing (including over 100 global
hensiveness for conducting LCAs whereby the wind tur- production factories for casting, machining, tower production,
bine's entire bill of materials is assessed, accounting generator production, nacelle assembly and blades produc-
individually for around 25,000 parts that make up the wind tion, as well as sales and servicing activities); all transport
turbine and, in total, around 99.95 % of the total mass of the stages; wind plant installation and erection; servicing, replace-
entire power plant. Also, in the manufacturing stage, the ment parts and operations; use-phase electricity generation;
LCA considers information from all Vestas' global sites, and and end-of-life treatment. Figure 2 shows a diagram of the
the use phase relies on Vestas' real-time performance data of basic wind plant components.
over 20,000 monitored wind turbines around the world,
covering around 20 % of the current worldwide installed 2.2 Product description
wind capacity (Vestas 2011a). This level of detail and reli-
ability in data represents a state-of-the-art approach to LCA The 2-MW GridStreamer™ range of turbines is Vestas' most
modelling of wind power. proven turbine platform with over 9,300 turbines installed
 Author's personal copy
Int J Life Cycle Assess

Fig. 1 System boundary of the wind power plant

globally (Vestas 2012a, b). The platform provides new 2.3 Functional unit
benchmarks of reliability, efficiency and competitive cost
of energy for 2-MW turbines. The 2-MW GridStreamer™ The functional unit used for the LCA study is 1 kWh of
range includes the onshore V80 2.0 MW, V90 1.8/2.0 MW electricity delivered to the grid by a wind power plant,
and V100 1.8/2.0 MW, known collectively as the ‘2-MW which includes all losses up to the existing grid point. This
platform’. The turbines share a significant number of common is based on the turbine design lifetime of 20 years and total
components (around 90 % of total weight), for example the electricity generation based on average wind speeds in the
nacelle, tower and all site parts (cabling, transformer, etc). The respective International Electrotechnical Commission (IEC)
primary difference between the turbines relates to the blade wind class (refer to Table 1).
size (80, 90 or 100 m in diameter), the generator output and
the ‘hub and nose cone’ module which has some differences
in construction. Figure 3 shows an open view of the nacelle of 2.4 Electricity generation and plant lifetime
a 2-MW GridStreamer™ turbine.
The Mk 8 series of turbines are assessed in the LCAs. Electricity production of the wind turbine is substantially
Also, a new upgrade to the 2-MW platform (Mk 9) will affected by the wind plant siting and site-specific wind
allow the V90 and V100 turbines to operate in stronger conditions that the turbine operates under, so it is not always
winds therefore generating more energy for the same mega- the size of the rotor or the generator rating (in megawatt)
watt rating of turbine. From a LCA perspective, the physical that determines the electricity production of the turbine. But
change to the upgraded turbine is minimal, relating to an wind class is a dominant factor. Nonetheless, electricity
upgraded gearbox of similar size and weight. The Mk 9 production is very accurately measured for Vestas' turbines
turbine was assessed in sensitivity analysis in the LCA. when the wind speed and conditions are known.

Fig. 2 Scope of wind plant components

 Author's personal copy
Int J Life Cycle Assess

Fig. 3 Main components in the

nacelle of 2-MW Grid-
Streamer™ turbine

Vestas' turbines are designed to meet different functional The power plant lifetime is also a dominant factor when
requirements both in terms of onshore and offshore loca- determining the impacts of the electricity production per
tions, as well as the wind classes for which they are kilowatt hour from the wind plant. The LCAs use a lifetime
designed to operate. The wind classes range from low wind of 20 years which matches the standard design life; however,
(IEC III), medium wind (IEC II) and high wind (IEC I), as the wind turbine industry is still young (starting for Vestas in
defined by the International Electrotechnical Commission 1979), and few turbines have ever been disposed, reaching
(IEC 2005). Table 1 shows the electricity generation of the operational lives of 30 years and over, for other Vestas' turbine
turbines assessed for the wind classes they operate within. models. It is often wear or fatigue of the load-bearing

Table 1 Wind plant electricity generation (in megawatt hour)

Wind speed 7.0m/s 8.0m/s 9.25m/s

Wind class IEC III Low IEC II Medium IEC I High

Per turbine Per 50MW plant Per turbine Per 50MW plant Per turbine Per 50MW plant
per year per 20years per year per 20years per year per 20years

MK 8—GridStreamer™
V100 2.0-MW MWh 6,707 3,756,000 – – – –
V90 2.0-MW MWh 6,257 3,129,000 – – – –
V80 2.0-MW MWh – – – – 8,269 4,135,000
MK 9—GridStreamer™
V100 2.0-MW MWh 6,909 3,454,000 8,274 4,137,000 – –
V90 2.0-MW MWh 6,257 3,129,000 7,632 3,816,000 9,131 4,566,000

Figures include all losses in wind plant and up to the point of grid connection. Source: Vestas internal data for the electricity production of the wind
plant. This is based upon actual turbine test data for a typical power production curve and using analysis software (based on T-CAT) of the specific
turbine performance. The annual energy production is reported in increments of 0.25ms−1 within the different wind classes and total electricity
production is determined over the range of 0ms−1 to 25ms−1 of the entire power curve for the specific turbine
 Author's personal copy
Int J Life Cycle Assess

components (such as tower fatigue) which limits the overall LCSA framework discussed by Guinée et al. (2011), based
turbine lifetime. Many components can be routinely replaced on a multidisciplinary approach and collection of analysis
as part of maintenance, except for the fixed parts (such as the methods.
tower, foundation and site cables, etc.) which are generally not Previous LCAs of Vestas' turbines (Elsam 2001, 2004;
replaced. Vestas 2006a; PE 2011a) show that manufacturing and
Vestas operates sophisticated real-time diagnostic tools end-of-life are the most significant stages in the turbine life
and sensors which measure individual turbine performance cycle. Conversely, turbine operation is not significant for total
and fatigue, and it is possible to predict the lifetime of environmental impacts, except that electricity generation and
specific components for specific site conditions. Also, these turbine lifetime are critical factors when assessing potential
systems monitor the current ‘health’ status of the turbine and impacts per kilowatt hour of electricity generated. Data col-
record its power production. These systems operate on over lection focused on procuring precise data for the production,
20,000 wind turbines around the world (of over 46,000 in use and disposal stages.
total), which covers 20 % of the current worldwide installed
wind capacity, providing very detailed and reliable data of Raw materials and production Modelling of the life cycle
turbine performance. begins with a bill of materials, which contains a part tree of
If comparisons of wind turbine performance from one all parts in the wind turbine. This is extremely extensive for
turbine to another will be made, then it is important that this a 2-MW GridStreamer™ turbine, consisting of around
is made on an equivalent functional basis and should be 25,000 parts for a standard turbine model. In the LCA
compared within the same IEC wind class at equal average model, a material dataset, manufacturing process and coun-
wind speeds. This will help to standardise comparisons of try of origin are assigned to each part. This builds a com-
wind turbine performance. prehensive product model of the entire turbine. The
Some LCAs of other wind plants, for example, as dis- component and material data are extracted from Vestas'
cussed by Davidsson et al. (2012), use a theoretical capacity internal master-data systems, 3D-CAD system used for
factor of the wind turbine to estimate the energy produced component design, and from supplier declarations and tech-
from the particular wind power plant. The capacity factor is nical specifications. For the turbine itself, over 99.5 % of the
the ratio of the actual energy produced (in a given time total mass was mapped (gaps related to missing material
period) divided by the hypothetical maximum possible, for data). All other parts in the entire power plant are included,
example, if the wind plant is running continuously at full such as foundations, cabling and transformer station, which
rated power, etc. It should be noted that the capacity factor is covers over 99.95 % of the total plant mass. Consequently,
a derived figure and one which is location specific based on the LCA model for one power plant consists of over
many wind plant characteristics, such as site wind conditions, 500,000 components, as well as additional production and
plant layout, availability, wake losses, electrical losses, main- end-of-life processes. This allows results to be analysed to a
tenance schedules, etc. The present LCAs do not use an very high degree of detail for individual components, mate-
assumed capacity factor (i.e. to estimate energy output), but rials and manufacturing processes.
specific data of turbine performance and wind plant losses are Modelling this many components “conventionally” in
used to very accurately assess the electricity production from LCA is not practicable. However, using GaBi DfX software
the wind plant. If other studies use an estimated capacity factor (PE 2011b) allows the complete bill of materials to be
(i.e. one that is not derived from power plant performance, but imported into the LCA model, and materials and manufac-
estimated), then the authors recommend that extra sensitivity turing processes are mapped to individual components to
analysis in the LCA is required to show appropriate consider- create a complete ‘product-model’. Secondary datasets for
ation of this uncertainty. production and processing are taken from recognised sour-
ces, such as Worldsteel (2010), Eurofer (2009), Plastics
2.5 Data collection and product modelling Europe (2010) and European Aluminium Association
(2008), covering over 90 % of the turbine mass. Figure 4
The LCAs adopt an attributional modelling approach to shows a material breakdown of the V80 2.0-MW Grid-
accurately portray the product life cycle of the wind plant. Streamer™ turbine.
Other approaches may also potentially be relevant but have Amongst other materials, the turbine also contains
not been applied, such as input–output-based LCA, hybrid quantities of rare earth elements (i.e. neodymium and
LCA or a consequential approach (Finnveden et al. 2009; dysprosium) used in the permanent magnets of the turbine
Guinée et al. 2011), or these may be useful to fulfil missing generator, as well as in magnets within the turbine tower for
inventory data. Future developments may also see the in- attaching internal fixtures. The production of these materials is
clusion of wider sustainability considerations, pertaining to based on production datasets for their sourcing from Europe
environmental, social and economic aspects, such as the and Asia (PE 2010).
 Author's personal copy
Int J Life Cycle Assess

(i.e. the decreased electricity generation capacity of the

turbine that is downwind of another turbine) and transmission
losses, described in Section 2.4.

Turbine servicing Serviced parts, such as oil, filers and

replaced components (due to wear and tear of moving parts
within the lifetime of a wind turbine), are included.

End-of-life The LCA model adopts an avoided impacts

approach, as compared to a recycled-content approach,
which is aligned with the approach supported by the metals
industry (Atherton 2006). As stated by Ekvall and Weidema
Fig. 4 Material breakdown of V80 2-MW GridStreamer™ (in (2004), whichever recycling approach is adopted, the
percent) inflows and outflows of recycled materials should be con-
sistent. For the wind power plant, environmental credits are
assigned for recycling of metals only, for example, based on
Vestas' production Over 100 Vestas' global factories and Worldsteel (2010) inventories for steel scrap value. Also, it
non-production sites are included in the LCA models. On has been assumed that all materials entering the production
a quarterly basis, Vestas' business units report all usage of system are sourced from primary material; however, for
materials, fuel and energy, as well as wastes and emissions iron, steel, aluminium and copper, the secondary (or scrap
(public data available at Vestas 2012a). Parameterised models metal) inputs to primary production have been adjusted to
were built in the GaBi software for each facility type, covering assign a burden to all secondary metal inputs (using primary
all inputs and outputs, production losses and internal recycling production or worldsteel ‘scrap value’ for these burdens).
loops. This standardised modelling and reporting procedure This provides a consistent approach to crediting for end-of-
allows easy inclusion of a large number of sites and update on life recycling. The use of an avoided impacts approach also
an annual basis. provides a very ‘usable’ business measure to drive up the
total recyclability of the wind turbine, which can be accu-
Transport Transport includes all stages for raw materials, rately measured using the LCA models. This in turn drives
turbine components, wind plant components, maintenance business activities, for example by focusing on recycling/
and end-of-life disposal. The LCA uses primary fuel con- reuse of non-metallic parts, such as composite blade materi-
sumption data collected by Vestas for truck and sea vessel als, controllers and polymers. It is also recognised that, from
transport of turbine components (i.e. tower, blades and a scientific perspective, both recycling approaches may be
nacelle). Specific transport distances reflect the actual Vestas' applied to wind turbines, as adopted by Crawford (2009)
supply chain, for example, tower manufacture reflects both where a recycled-content approach is used, for example.
Vestas' own production sites and supplier locations. Additionally, guidance already exists for preparing an En-
vironmental Product Declaration (EPD) based on ISO
Background electricity Choices of electricity mix selected 14025 (2006c) for electricity generation and distribution
for manufacturing steps can affect the LCA results. For via the Product Category Rules (Envirodec 2007, 2011). In
the current LCAs, country-specific average electricity general, the EPD guidance aligns with the current LCAs
mixes have been used based on the country producing (which are in compliance with ISO 14040/44), relating to
each component to reflect the actual supply-chain functional unit, system boundaries and general data quality
geographies. requirements, etc. However, for the EPD approach, end-of-
life and recycling credits are excluded from the boundary,
Turbine installation/decommissioning Specific data are meaning that a recycled-content approach is adopted in the
used for usage of cranes, onsite vehicles, diggers and gen- EPD. Given that both approaches can be applied, it may be
erators for laying the foundations, erecting the turbines, beneficial to show the consequences of both crediting
laying site cables, installing/erecting the transformer station approaches in future LCA results via sensitivity analysis.
and connecting to the existing grid. Provision of roads, Additionally, the end-of-life disposal of the wind plant
working areas and turning areas is included. has been modelled in a part-specific manner. The GaBi DfX
software has the function to disassemble the entire turbine
Turbine operation Use-phase electricity generation (or parts of it) into its source components and dispose of
accounts for wind turbine availability (i.e. the capability of these parts through different routes and disposal efficiencies.
the turbine to operate when wind is blowing), wake losses This allows for an extremely detailed end-of-life model to
 Author's personal copy
Int J Life Cycle Assess

be created that is part specific, for example, components & Primary energy from renewable raw materials (megajoule
that are easily dismantled and recycled (e.g. the steel net value)
tower) will receive higher efficiencies than other parts & Primary energy from resources (megajoule net value)
of the turbine. The entire power plant is broken down & Water use (mass)
into all major components or assemblies, and each & Turbine recyclability (percent turbine mass)
receives specific recycling and disposal rates based on & Return-on energy (months)
industry data.
Specifically, for water use, the present LCAs only ac-
As is apparent, the LCA models provide a tremendous
count for net freshwater use—meaning that only the total
level of detail for examining the performance of the wind
input of freshwater used for production and disposal pro-
turbine and its individual components. Furthermore, the
cesses is accounted for. Nonetheless, for this to be treated
LCA models include various ‘product options’ for the tur-
more thoroughly, further consideration should be made re-
bines, such as differing tower heights or foundation types,
garding types of water used, inclusion of local water scarcity
which may vary depending on customer or site-specific
as well as differentiation between watercourses and quality
requirements. These product options are not ‘used’ in the
aspects (Berger and Finkbeiner 2010), which will aid more
baseline LCA model of a typical wind plant; however, this
accurate decision making.
flexible approach to LCA modelling allows Vestas to create
tailored LCAs of a specific wind power plant, for example,
giving Vestas' customers the opportunity to obtain a detailed
(and externally reviewed) assessment of their own specific 3 Results for 2-MW GridStreamer™ wind turbines
wind plant.
The LCA provides accurate and robust results by utilising
2.6 Impact assessment the very detailed LCA models developed in the inventory
stage, which cover 99.95 % of the wind plant, and include a
The impact categories assessed in the LCA represent those detailed range of sensitivity analyses. For the V80 2.0-MW
likely to arise from a wind plant system, based on the CML GridStreamer™ turbine, the whole-life results are shown in
(2009) baseline characterisation factors for mid-point poten- Table 2, which show that the manufacturing phase domi-
tial impacts. A series of detailed sensitivity analyses were nates the life cycle, followed by end-of-life management
also conducted which explore the primary uncertainty in the where environmental credits are received for recycling of
LCA models. metals. The plant setup and operation stages have much less
The CML impact categories focus on the so-called ‘mid- significance.
points’ of the cause–effect chain. This means that the meth- Figure 5 presents the potential impacts for manufacture,
od aggregates data for the inventory flows and characterises plant setup and operation (i.e. all life cycle stages excluding
their potential impacts in various categories (e.g. global end-of-life). The results show that the nacelle, tower and site
warming, acidification, etc.) but do not go as far as to assess parts (primarily site cables) contribute most significantly
the endpoints, such as loss of biodiversity, damage to human across all impact categories. The next most significant com-
health, etc. caused by these impacts, as developed in meth- ponents relate to the turbine foundation and blades. Vestas'
ods such as Eco-Indicator 99 (Goedkoop and Spriensma factories contribute between 1 and 8 % across all impact
1999) or ReCipe (Goedkoop et al. 2009). As such, the categories. Transport is included for each part and is not
impact assessment results generated do not predict impacts disaggregated. The following sections describe the potential
on category end points, the exceeding of thresholds, safety impacts in more detail for the global warming potential,
margins or risks. return-on energy and turbine recyclability.
In general, a life cycle assessment cannot always address
all environmental concerns, such as the potential impacts of 3.1 Global warming potential
land use, noise or potential local impacts on flora and fauna.
As such, a LCA should not stand alone in the assessment of Table 3 shows the whole-life global warming potential per
technologies; but other environmental management techni- kilowatt hour of electricity produced by the various Grid-
ques, such as risk assessment and environmental impact StreamerTM turbines. Overall, the manufacturing stage dom-
assessment, for example, are valuable tools that address inates the life cycle, with the production of the tower (25–
these environmental concerns. These types of assessments 30 %), site cables (20 %), nacelle (15 %), blades (10–15 %)
are usually conducted as part of the local permitting and and foundations (10 %) which are the primary components
planning process for installation of a wind plant. contributing to global warming. Vestas' production and
Several additional, non-impact indicators were assessed, operations contribute around 3 % and end-of-life contributes
as follows: around −30 % for the environmental credit associated with
 Author's personal copy
Int J Life Cycle Assess

Table 2 Whole-life impacts of V80 2.0-MW GridStreamer™ by life cycle stage per kilowatt hour (in milligram, gram or megajoule)

Impact category Unit Manufacture Plant setup Operation End-of-life Total

Abiotic resource depletion (elements) mg Sb-e 0.53 0.00 0.05 −0.14 0.44
Abiotic resource depletion (fossils) MJ 0.12 0.01 0.01 −0.04 0.10
Acidification potential mg SO2-e 47 1 2 −13 37
Eutrophication potential mg PO4-e 3.8 0.2 0.2 −0.4 3.7
Freshwater aquatic ecotoxicity mg DCB-e 110 1 3 −11 100
Global warming potential g CO2-e 10.8 0.2 0.4 −3.8 7.7
Human toxicity potential mg DCB-e 2,600 10 140 −1,600 1150
Marine aquatic ecotoxicity potential g DCB-e 5,400 10 100 −4,400 1100
Photochemical oxidant creation mg Ethene 5.4 0.1 0.2 −1.7 4.1
Terrestrial ecotoxicity potential mg DCB-e 21 0.3 2.3 −4.5 19
Non-impact indicators
Renewable primary energy (net) MJ 0.021 0.000 0.002 −0.006 0.016
Non-renewable primary energy (net) MJ 0.14 0.01 0.01 −0.04 0.11
Water use g 46 0 2 −10 39

metal recycling. Transport in total over the life cycle contrib- the energy requirement over the whole life cycle of the wind
utes around 8 % to global warming potential. plant (i.e. to manufacture, setup, operate, service and
dispose) versus the electrical energy output of the wind
3.2 Return-on energy plant. The payback period is measured in months where
the energy requirement for the life cycle of the wind plant
Return-on energy provides an indication of the energy bal- equals the energy it has produced, using the following
ance of the power plant, showing the relationship between equation:

life cycle energy requirement of the wind plant ðin megajouleÞ  240 ðmonthsÞ
Energy paybackðmonthsÞ ¼
electrical energy output from the wind ðin megajouleÞ

The above measure may be calculated in two basic ways: the electrical energy output (either net or primary) from the
(1) based on the net electrical energy output of the wind wind plant and then multiplied by the power plant lifetime
plant or (2) based on the equivalent primary energy output in months.
of the wind plant. In both cases, the primary energy require- Table 4 shows the results for the 2-MW GridStreamerTM
ment for the whole life cycle of the wind plant is divided by turbines based on net energy which ranges from 8 to

Fig. 5 Impacts from

manufacture, setup and
operation of V80 2.0-MW
GridStreamer™ (in percent)
 Author's personal copy
Int J Life Cycle Assess

Table 3 Global warming po-

tential impacts per kilowatt hour Wind class Low Medium High
(grams CO2-e) IEC III IEC II IEC I

MK 8—GridStreamer™
V100 1.8-MW CO2-e per kWh 9.3 – –
V90 2.0-MW CO2-e per kWh 9.7 – –
V80 2.0-MW CO2-e per kWh – – 7.7
MK 9—GridStreamer™
V100 2.0-MW CO2-e per kWh 9.3 8.0 –
V90 2.0-MW CO2-e per kWh – 8.0 7.2

11 months return-on energy. This may be interpreted that the efficiency conversion of up to 45 % with latest technologies
wind plant will return between 22 and 30 times more energy for coal power and a global average of 28 % IEA 2006) and
back to society than it consumed over the wind plant life will therefore always be net consumers of energy, returning
cycle. Accounting for the return-on energy based on the less energy back to society than is consumed throughout the
equivalent primary energy approach gives a significantly life cycle. Conversely, data shown in Höök et al. (2012) for
lower payback period of between 2 and 4 months for all energy-return-on-energy investment only appear to account
turbines assessed. The latter (primary energy) measure gives for the capital energy investment of the power plants and
a relative theoretical return-on energy because the electrical exclude the inherent energy in the fuel to operate the plants,
energy from the turbine is converted to the primary energy which is why return-on energy for these fossil fuels ranges
of a typical electrical grid mix for different world regions. from around 20 to 100 times. However, this appears
This approach accounts for the efficiency of the electricity nonsensical, as the dominant energy contributor (i.e.
power stations when determining the primary energy. There the fuel itself needed to generate the electricity) is
is no distinction made here as to whether or not base-load missing from the equation; if included, the return-on
energy mix or marginal-load energy mix should be energy would at best be below 0.5 times—or 50-fold
assessed, but in this article, an average electricity mix to 60-fold lower than wind. This does not mean to
for Europe, Australia and USA were analysed. The net suggest that wind power is entirely independent of fossil
return-on energy approach does not include these relative energy in its own life cycle, conversely, but it does
conversions and therefore provides an absolute indication show the clear relationship of energy investment versus
of performance. return-on energy back to society.
In real terms, whichever approach is used, the return-on
energy provides a measure of the energy that is returned 3.3 Turbine recyclability
back to society from investing in the wind plant. This
provides a robust measure when choosing between where Recyclability measures the proportion of the turbine (in
to invest energy and the resulting payback to society as a percent mass) that can be usefully recycled at end-of-life,
whole. For example, fossil-based electricity sources, such as accounting for specific recycling rates of various compo-
coal, oil or gas, will consume the intrinsic energy of the fuel nents. The measure only relates to the turbine itself and
in generating electricity during operation (with energy excludes other wind plant components (i.e. foundation,

Table 4 Return-on energy of

the GridStreamerTM turbines Wind class Low Medium High
(months) IEC III IEC II IEC I

MK 8—GridStreamer™
V100 1.8-MW Months payback 11 – –
V90 2.0-MW Months payback 11 – –
V80 2.0-MW Months payback – – 9
MK 9—GridStreamer™
V100 2.0-MW Months payback 11 9 –
V90 2.0-MW Months payback 11 9 8
 Author's personal copy
Int J Life Cycle Assess

site cabling, transformer station, etc.) using the following resource use (both energy and material resources) associated
equation: Turbine recyclability ðin percentÞ ¼ sum of all turbine with the life cycle of that product or service. LCA methods to
  characterise resource depletion account for the geological
ðin percentÞmetal part mass ðin kilogramÞ
parts metal recycling ratetotal part massðin kilogramÞ The metal reserves and the anticipated depletion rates (CML 2009),
components of iron, steel, aluminium and copper contribute to and they are usually presented in mass of antimony equiva-
recyclability. Table 5 shows the results for the GridStreamer™ lents. The quantity of reserve may relate to, for example, the
turbines. By analysing the wind turbine to cover 25,000 parts ultimate reserve (every element in the earth's crust), extract-
that make up the turbine provides very detailed and reliable able reserve or the economic reserve. The current LCAs use
results. The primary reason for variation in recyclability the established CML3.6 (2009) method for resource deple-
between the 2-MW GridStreamerTM turbines relates to the tion, although some evidence shows that the methods to
tower construction, where a turbine operating in a higher wind measure resource use requires further consensus (Finnveden
class requires a higher strength tower, using more steel for the 2005). In the context of rare earth elements (REEs), there is
same height. often a discussion about the use of these metals for renewable
energy systems, such as wind turbines. They are also com-
monly used in many different industries and a wide range of
products, including a large number of common personal and
4 Discussion household products such as hybrid car motors, MP3 players,
computer hard drives, low-energy light bulbs, LCDs and
The article presents a case study for the application of LCA mobile phones, as well other industrial sectors such as medical
to a complex wind-turbine product and presents some of the imaging, catalysts to process crude oil and military hardware
results of these studies. This delivers value to Vestas and its such as radars and guidance systems (Delaney 2010). Also, it
various stakeholders, though being equipped with extensive is a common misconception that REEs are actually scarce
facts and state-of-the-art LCA models, which is providing materials—in reality, they are not very scarce at all and there
the basis to further integrate environmental considerations is an abundance in the earth's crust. Nonetheless, they may
into product communications, design and research optimi- present more of a political scarcity issue in so far as the vast
sation, procurement and to deliver transparent information majority of the world's supply is from China (Chen 2011),
to customers, consumers and governments. which potentially presents an issue of security of supply and
The article aims to present a transparent and robust ap- pricing. In relation to wind turbines, the use of REE actually
proach to the LCA modelling of wind power systems and to provides environmental benefits for the product, allowing
show improvement in the methods and comprehensiveness more compact and energy efficient turbines to be built, which
of LCA application from previous studies of wind energy. reduces the reliance on resource use. These elements are used
Some points highlighted in Davidsson et al. (2012), such as in the wind turbine generator, for permanent magnet genera-
the need for greater transparency and standardisation of tors, as well as in the tower magnets for fixing internal tower
methods and calculations, are comparatively addressed in components (as briefly outlined in the example below for
the article. Vestas aims to work with best available LCA tower magnets).
approaches to assess the performance of its products, as well Through identifying environmental ‘hotspots’, Vestas is
as to provide openly transparent and robust LCA studies. able to direct potential product improvements, for example,
Some further discussion about the characterisation of nat- the LCA of the V112 3.0-MW turbine has been used to set
ural resource use is also worth noting. As with all material- targets to reduce the turbine's global warming potential
based products or energy systems, there will be some kind of per kilowatt hour by 15 % and to increase the turbines

Table 5 Turbine recyclability

(in percent turbine mass) Wind class Low Medium High
IEC III IEC II IEC I

MK 8—GridStreamer™
V100 1.8-MW % turbine mass 81 % – –
V90 2.0-MW % turbine mass 82 % – –
V80 2.0-MW % turbine mass – – 85 %
MK 9—GridStreamer™
V100 2.0-MW % turbine mass 81 % 81 % –
V90 2.0-MW % turbine mass – 84 % 85 %
 Author's personal copy
Int J Life Cycle Assess

recyclability to 85 % by 2015. Vestas has also set stringent conditions and production supply chain. This will provide
targets for increased use of renewable energy and renewable customers with transparency and key facts for their business
electricity across all Vestas' business units and factories (e.g. requirements and stakeholders.
use of 100 % renewable electricity by 2012). The impact of
achieving these targets is included in the product LCA
5 Conclusions
models, allowing traceability and management at both a
business-unit and product level.
Overall, the article presents a case study of the LCA ap-
The LCA approach is also being used within other parts of
proach used to assess the potential impacts of the 2-MW
the Vestas' business, such as the environmental assessment of
GridStreamer™ turbines, which represents a state-of-the-art
alternative technologies or supply chain sourcing options
approach to LCA modelling of wind power. The LCAs are
which aim to simultaneously lower the cost of energy for wind
based on highly detailed LCI component models (of around
and reduce environmental impacts. Two examples include
25,000 parts per turbine and covering the entire wind power
& Magnets are used in a turbine tower for the fixation plant), allowing detailed and transparent product analysis, as
internal components (e.g. ladders, fixtures, etc.). On well as using industry-published LCI datasets and current
the V112 3.0-MW turbine with an 84-m tower, the use best practice for methodological choices for life cycle mod-
of magnets will result in a saving of around 10 tonnes of elling, such as recycling approach, and the application of
steel per turbine versus the traditional solution of weld- established mid-point impact characterisation (CML 2009)
ing because the thickness of the steel plate in the tower models. Alternative approaches to LCA modelling (for ex-
can be reduced. This equates to avoiding around 8.0 ample, input–output LCA, hybrid LCA or future LCSA
tonnes of CO2 equivalents per turbine over the life cycle methods) may also be useful, as well as some additional
(or about 0.3 % of total wind plant GWP), as well as areas where further sensitivity analysis could improve pre-
significant cost savings and improved safety for manu- sentation of results, e.g. to show results for avoided-impact
facturing and assembly operations. versus recycled-content crediting approaches at end-of-life
& Sulphur hexafluoride (SF6) gas is a very potent green- and to include additional sensitivity analysis for other stud-
house gas which may be used in the turbine switchgear. ies where only ‘estimated’ capacity factors have been used.
The switchgear protects and isolates the electrical com- Additionally, to improve standardisation of LCA studies, we
ponents of the turbine and contains around 7 kg of SF6 recommend that wind turbines are only compared within
gas per turbine. The gas acts as an electrical insulator equal (IEC) wind classes, as well as including appropriate
and allows for compact design of switchgear units. But sensitivity analysis in the LCAs in order to assess primary
if improperly disposed at end-of-life whereby the gas is assumptions and uncertainties. In terms of non-traditional
released to the atmosphere, then this will contribute very LCA impact results, the measure presented here of net-
significantly to the total global warming burdens of return-on energy provides a real-life indicator for the elec-
about 160 tonnes of CO2 equivalents per turbine or trical energy performance of the wind turbine, which corre-
about 10 % of total wind plant GWP (PE 2011a). lates to a contribution between 22 and 30 times more energy
Consequently, Vestas has established procedures and is back to society as a whole, than is consumed by the wind
working in partnership with customers and suppliers to plant over the life cycle. The current LCAs also demonstrate
assure the safe operation, disposal and selection of switch- an improvement in LCA application over previous studies
gears used by Vestas. of wind power, from a perspective of both transparency, as
well as comprehensiveness of LCA modelling. Lastly, the
Vestas' newest and 100 % dedicated offshore turbine, the
article discusses some further applications of using LCA as
V164-7.0 MW, has undergone a screening LCA as an inte-
a tool for both internal and external business processes, for
gral part of the product development process. The LCA was
example, to direct product improvements and for external
based on other more detailed LCA models and was prepared
communications, as well as highlighting improved transpar-
to enable managers and designers to include environmental
ency and robustness.
considerations in decision making. The LCA has identified
areas of the turbine performance where potential improve-
ments can be made earlier in the design phase. Acknowledgments We thank Prof. Dr. Matthias Finkbeiner (chair-
man of Sustainable Engineering, Department of Environmental Tech-
As mentioned, by creating such comprehensive LCA nology, Technische Universität Berlin and chairman of ISO TC207
models described in this article, Vestas can now offer cus- SC5 Life Cycle Assessment) for his expert review of the LCAs in
tomers the option to acquire a detailed and tailored LCA of accordance with paragraph 6.2 of ISO 14040/44. He acted as a con-
their own wind power plant. The LCA can be developed sultant in this capacity and did not represent his employer. We also
thank the reviewers during the submission of the manuscript to the
at an early stage in the wind project before the plant is International Journal for LCA for their valuable and constructive
operational, taking into account turbine type, local site comments.
 Author's personal copy
Int J Life Cycle Assess

References Höök M, Li J, Johansson K, Snowden S (2012) Growth rates of global

energy systems and future outlooks. Nat Resour 21(1):21–41
IEC (2005) INTERNATIONAL STANDARD, IEC 61400-1: wind
Atherton J (2006) Declaration by the metals industry on recycling turbines—part 1: design requirements. Third edition 2005-08.
principles. Int J Life Cycle Assess 12(1):59–60 International Electrotechnical Commission, Geneva, Switzerland
Berger M, Finkbeiner M (2010) Water footprinting: how to address International Energy Agency (2006). Focus on clean coal. International
water use in life cycle assessment? Sustainability 2(4):919–944 Energy Agency, Paris, France. www.iea.org/work/2006/gb/
Chen Z (2011) Global rare earth resources and scenarios of future rare papers/focus_on_coal.pdf. Accessed March 2012
earth industry. J Rare Earth 29(1):1–6 ISO (2006a) ISO 14040. Environmental management—life cycle as-
CML (2009) CML 3.6 developed by the Centre for Environmental Studies sessment—principles and framework (second edition, 2006-07-
(CML). November 2009. University of Leiden, The Netherlands. http:// 01). Geneva, Switzerland
cml.leiden.edu/software/data-cmlia.html. Accessed October 2011 ISO (2006b) ISO 14044. Environmental management—life cycle as-
Crawford RH (2009) Life cycle energy and greenhouse emissions sessment—requirements and guidelines (first edition, 2006-07-
analysis of wind turbines and the effect of size on energy yield. 01). Geneva, Switzerland
Renew Sustain Energy Rev 13(9):2653–2660 ISO (2006c) ISO 14025:2006 environmental labels and declarations—
Davidsson S, Höök M, Wall G (2012) A review of life cycle assess- type III environmental declarations—principles and procedures.
ments on wind energy systems. Int J Life Cycle Assess. Geneva, Switzerland
doi:10.1007/s11367-012-0397-8 PE International (2010) Life cycle inventory datasets for neodymium
Delaney K (2010) Rare earth supply and national security & clean and dysprosium metal production. PE International, Leinfelden-
energy, TREM 10. Understanding the fundamentals, Hilton Wash- Echterdingen, Germany
ington Embassy Row, March 17, 2010. Rare Earth Industry and PE International (2011a) Life cycle assessment of electricity produc-
Technology Association (REITA) tion from a V112-3.0 MW wind plant—February 2011. PE Inter-
Ekvall T, Weidema BP (2004) System boundaries and input data in national, Leinfelden-Echterdingen, Germany. www.vestas.com/
consequential life cycle inventory analysis. Int J Life Cycle As- en/aboutvestas/sustainability/wind-turbines-and-the-environment/
sess 9(3):161–171 life-cycle-assessment-(lca).aspx. Accessed October 2011
Elsam (2001) Life cycle assessment of turbines. Report 01-488 (no. PE International (2011b) GaBi DfX software. PE International,
138094) from July 2001. Elsam Engineering A/S, Denmark Leinfelden-Echterdingen, Germany
Elsam (2004) Life cycle assessment of offshore and onshore sited wind Vestas (2006a) Life cycle assessment of electricity produced from
farms. Doc. no. 200128. Elsam Engineering A/S, Denmark. www. onshore sited wind power plants based on Vestas V82-1.65 MW
vestas.com/en/about-vestas/sustainability/wind-turbines-and-the- turbines. Vestas Wind Systems A/S, Alsvej 21, 8900 Randers,
environment/life-cycle-assessment-(lca).aspx. Accessed October 2011 Denmark. www.vestas.com/en/about-vestas/sustainability/wind-
Envirodec (2007) Product Category Rules (PCR) for preparing an turbines-and-the-environment/life-cycle-assessment-(lca).aspx.
Environmental Product Declaration (EPD) for electricity, steam, Accessed October 2011
and hot and cold water generation and distribution. PCR CPC 17. Vestas (2006b) Life cycle assessment of offshore and onshore sited wind
Version 1.1, 2007-10-31 power plants based on Vestas V90-3.0MW turbines. Vestas Wind
Envirodec (2011) Product Category Rules (PCR) for preparing an Systems A/S, Alsvej 21, 8900 Randers, Denmark. www.vestas.com/
Environmental Product Declaration (EPD) for electricity, steam, en/about-vestas/sustainability/wind-turbines-and-the-environment/
and hot and cold water generation and distribution. PCR CPC 17. life-cycle-assessment-(lca).aspx. Accessed October 2011
Version 1.1, 2007-10-31 Vestas (2011a) Life cycle assessment of electricity production from a V80-
Eurofer (2009) Life cycle inventory study on stainless steel production 2.0MW Gridstreamer wind plant—December 2011. Vestas Wind
in the EU, November 2009. European Confederation of Iron and Systems A/S, Alsvej 21, 8900 Randers, Denmark. www.vestas.com/
Steel Industries, Eurofer: www.eurofer.org cited in PE Interna- en/about-vestas/sustainability/wind-turbines-and-the-environment/
tional (2011) GaBi DfX software. PE International, Leinfelden- life-cycle-assessment-(lca).aspx. Accessed October 2011
Echterdingen, Germany Vestas (2011b) Life cycle assessment of electricity production from a V90-
Europe P (2010) PlasticsEurope's cycle inventory datasets cited in PE 2.0MW GridStreamer wind plant—December 2011. Vestas Wind
International (2011) GaBi DfX software. PE International, Systems A/S, Alsvej 21, 8900 Randers, Denmark. www.vestas.com/
Leinfelden-Echterdingen, Germany en/about-vestas/sustainability/wind-turbines-and-the-environment/
European aluminium association (2008) EAA environmental profile life-cycle-assessment-(lca).aspx. Accessed October 2011
report for the EU aluminium industry, EAA, April 2008 cited in Vestas (2011c) Life cycle assessment of electricity Production from a
PE International (2011) GaBi DfX software. PE International, V100-1.8MW GridStreamer wind plant—December 2011. Vestas
Leinfelden-Echterdingen, Germany Wind Systems A/S, Alsvej 21, 8900 Randers, Denmark. www.vestas.
Finnveden G (2005) The resource debate needs to continue. Int J Life com/en/about-vestas/sustainability/wind-turbines-and-the-
Cycle Assess 10(5):372 environment/life-cycle-assessment-(lca).aspx. Accessed October 2011
Finnveden G et al (2009) Recent developments in life cycle assess- Vestas (2012a) Vestas signs 56MW order in Italy. Vestas Wind Systems A/
ment. J Environ Manage 91(1):1–21 S, Alsvej 21, 8900 Randers, Denmark. www.vestas.com/Defaul-
Goedkoop M, Spriensma R (1999) The eco-indicator 99—a damage t.aspx?ID010332&action03&NewsID03013. Accessed March 2012
oriented method for life cycle impact assessment. PRe´ Consul- Vestas (2012b) Vestas by numbers. Vestas Wind Systems A/S, Alsvej
tants, Amersfoort, The Netherlands 21, 8900 Randers, Denmark. http://data.vestas.com. Accessed
Goedkoop M, Heijungs R, Huijbregts MAJ, De Schryver AM, Struijs October 2011
J, van Zelm JR (2009) ReCiPe 2008—a life cycle impact assess- Worldsteel (2010) Worldsteel life cycle inventory study for steel in-
ment method which comprises harmonised category indicators at dustry products. World Steel Association. www.worldsteel.org
the midpoint and the endpoint level, 1st ed.; Report I: Character- cited in PE International (2011) GaBi DfX software. PE Interna-
isation, 2009. Available at www.lcia-recipe.net/ tional, Leinfelden-Echterdingen, Germany
Guinée JB, Heijungs R, Huppes G, Zaagmo A, Masoni P, Buonamici WWEA (2011) World wind energy report 2010. World Wind Energy
R, Ekvall T, Rydberg T (2011) Life cycle assessment: past, Association. April 2011. www.wwindea.org/home/images/stories/
present, and future. Environ Sci Technol 45(1):90–96 pdfs/worldwindenergyreport2010_s.pdf. Accessed October 2011

View publication stats