Back to the programme printer.gif Print

Thursday, 13 March 2014
11:15 - 12:45 Innovative concepts for drive train components
Science & Research  

Room: Ponent
Session description

New developments in wind turbines need innovation and advances in technology in the field of wind turbine drive trains. This session focuses on topics related to transmissions and generators

Lead Session Chair:
Emilio Gomez-Lazaro, Universidad de Castilla-La Mancha. Renewable Energy Research Institute, Spain
Kerri Hart University of Strathclyde, United Kingdom
Kerri Hart (1) F P Alasdair McDonald (1) Henk Polinder (2) Edward Corr (1) James Carroll (1)
(1) University of Strathclyde, Glasgow, United Kingdom (2) Delft University of Technology, Delft, The Netherlands

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Kerri Hart MSci, graduated in 2011 from the University of Strathclyde with a Masters degree in Physics. She is currently in her third year of her PhD with the University of Strathclyde’s Centre for Doctoral Training in Wind Energy Systems. Her research focus is optimum drivetrain design for large offshore wind turbines with a particular interest in cost of energy.


Improved Cost of Energy Comparison of Permanent Magnet Generator for Large Offshore Wind Turbines


This paper investigates geared and direct drive permanent magnet generators for a typical offshore wind turbine, providing a detailed comparison of various wind turbine drivetrain configurations in order to minimise the Cost of Energy. The permanent magnet generator topologies considered include a direct drive machine and single stage, two-stage or three-stage gearbox driven generators. The cost of energy calculations are based on initial capital costs, the costs of manufacture, installation, operations and maintenance, with particular focus on improved calculations of the annual energy yield with better availability estimations and gearbox loss modelling.


This paper builds on the work of Polinder et al. [1] and Bywaters et al. [2] but with an emphasis on a typical 6MW offshore wind turbine. This paper considers the four following generator topologies: the direct drive permanent magnet generator (DDPMG) and three geared permanent magnet generator topologies which include a single stage gearbox (PMG1G) (which was considered in [1]), a two-stage gearbox (PMG2G) and a three-stage gearbox (PMG3G).

The main design goal of a wind turbine design team should be to deliver the lowest possible Cost of Energy. The cost of energy (COE) [3] per MWh is the overall outcome of this study and is calculated as follows:


where FCR is the Fixed Charge Rate, ICC is the Initial Capital Cost, AOM is the Annual Operation and Maintenance [4] and AEP is the Annual Energy Production.

The cost of the drivetrain represents a fraction of the whole capital cost but all AEP is generated by it. The implication is that marginal increases in drivetrain capital costs have a smaller impact on turbine COE than a similar per unit change in drivetrain efficiency and availability. This drives the focus in this paper to focus on higher fidelity efficiency and availability drivetrain models.

Improved models for gearbox losses and drivetrain downtime are used to more accurately estimate the AEP of the candidate drivetrains. The gearbox losses are modelled based on ISO/TR 14179-1:2001 [5]. The study provides rotor speed dependant gearbox losses that may provide a more accurate loss model compared with other methods that use percentage losses of rated power per stage [6].

The availability figures of each configuration are based on a study by Carroll in [7] that calculates offshore availability figures based on onshore failure rates and predicted delay/travel times from particular sea conditions.

Main body of abstract

A baseline scenario is presented with typical drivetrain costs and ratings. Using the same designs two further scenarios are investigated. In one scenario the gearbox is replaced once in its lifetime for each of the geared drivetrains, and in another the cost of permanent magnets doubles. These represent credible risks that drivetrain designers must consider.

Table 1 – Wind turbine characteristics and loss model variables
Wind Turbine Characteristics:
Rated Grid Power (MW) 6
Rotor Diameter (m) 140
Rated Wind Speed (m/s) 11
Rated Speed (rpm) 12

Cost Modelling:
Single-stage gearbox (ratio 7) cost (kEuro) 489
Two-stage gearbox (ratio 40) cost (kEuro) 647
Three-stage gearbox (ratio 100) cost (kEuro) 970
Power electronics cost (Euro/kW) 40
Laminations cost (Euro/kg) 3
Copper cost (Euro/kg) 15
Permanent magnet cost (Euro/kg) 48
Rest of wind turbine cost (Euro) 6100

Gearbox Efficiency
The gearbox losses were modelled based on [5]:

Baseline scenario
Table 1 gives details of the 4 generator dimension constraints – these are typical for offshore turbines in the development stage (turbine ratings are in Table 2). Cost of energy was calculated for all the drivetrains in MATLAB and then minimised using a combination of published designs and by using an optimisation routine. Costs for the generator materials, gearbox and power converters were based on [1-6] and [8-9].

Results from the MATLAB model for the DDPMG are shown in Fig. 2. The results shown include voltage and current levels, the wind turbine power curve, the system efficiency and losses. The system efficiency takes into account bearing and cable losses as well as generator and converter losses (also includes gearbox losses where applicable). In this case the generator diameter was restricted to 7m. The significant losses in the system result from very high copper losses that account for over half of the total losses. This is due to the requirement that in order to produce a high torque there is a large number of coils. Availability is high: even though there are increased winding failures in the direct drive generator, there is no downtime due to a gearbox.

MATLAB simulations of the performance of the PMG1G are shown in Fig. 3. The gearbox only has a single stage, so the gearbox losses are the lowest of the three geared designs. The copper losses in the generator and greatly reduced compared to the direct drive machine because the torque rating is reduced by a factor of 7. There is a modest increase in the iron losses due to the higher frequency. Availability is similar to the direct drive machine – downtime due to generator increases but is replaced by gearbox downtime.

The PMG2G generator size is considerably smaller than the single-stage gearbox design and where the higher frequency sees iron losses exceeding copper losses. Because of the increased gearbox ratio the losses in the gearbox are greater than those in the generator. This lightweight generator would be advantageous to developers during installation procedures.

The increased gear ratio in the PMG3G does not make a significant improvement in generator size, cost or efficiency, but does lead to increased gearbox cost and losses.

Table 2 – Annual Energy Production and Cost of Energy for the 4 drivetrains

The Cost of Energy results are shown in Table 3. The preliminary results indicate that the most cost (of energy) effective generator configuration is the PMG1G with overall lowest losses and highest energy yield. The PMG3G appeared to be the least appealing choice as the large, expensive gearbox resulted in high losses in the gearbox and overall highest COE.

Gearbox replacement scenario:
There have been some wind farms where there have been a significant number of gearbox replacements required [10]. What happens to the cost of energy if there are one or two gearbox replacements per turbine over the lifetime of the wind farm? This can be modelled by modifying equation (1) and adding a replacement cost for one gearbox replacement. Table 4 shows the change in cost of energy.

Permanent magnet cost increase scenario:
If the cost of permanent magnets increases by 100%, then the effect of the topologies considered is shown in Table 4.

Table 4 – Effect on COE for two different scenarios
COE (Euro/MWh) 111.6 108.9 110.7 115
COE with one gearbox replacement (Euro/MWh) (111.6) 111.1 113.6 119.4
COE with magnet prices increased by 100% (Euro/MWh) 113.4 109.3 111.7 115.8


This study provided a comparison between four different drivetrain configurations using permanent magnet generators. The drivetrains were modelled to assess which drivetrain configuration offers the lowest cost of energy solution for a large offshore wind turbine.

Both the PMG1G and PMG2G perform well in terms of both energy yield and COE with the PMG1G slightly in front. The single stage gearbox concept outperforms the other three designs, reinforcing the benefits of the “Multibrid” design. This implies that the COE for these geared concepts could be much greater than indicated in this study. In the scenario with one gearbox replacement, there was little difference between the COE for the DDPMG and the PMG1G.

From the initial results presented in Table 3, it can be seen that the PMG3G is the most expensive in terms of ICC and leads to the highest COE. This is due to having an expensive gearbox and high total losses with the gearbox losses accounting for almost half of the losses. Due to the generator’s compact design, there are reductions in copper losses, but not significant enough to improve the performance. The DDPMG is the second most expensive design in terms of ICC. It does however, have the lowest energy yield which could be due to the fact that the design could not be fully optimized beyond a diameter of 7 m as this proves too unfeasible in terms of transport and installation. This then highlights the fact that in order to be an efficient design, DDPMGs must have very large diameters in order to produce a high energy yield that is required of large wind turbines.

The DDPMG COE was the most sensitive to changes in permanent magnet costs: a doubling of the magnet price increased the COE by 1.6% (for the PMG1G 0.4%, for the PMG2G 0.9% and PMG3G 0.7%).

Learning objectives
This will enable a design team of a typical offshore wind turbine:
 To come to an informed decision on choice of drivetrains with permanent magnet generators when minimised cost of energy is the goal.
 To model gearbox losses in wind turbines with a higher degree of accuracy.
 To incorporate drivetrain availability into the cost of energy calculation.

[1] H. Polinder, F. F. A. van der Pijl, G. J. de Vilder, and P. J. Tavner, "Comparison of direct-drive and geared generator concepts for wind turbines," IEEE Transactions on Energy Conversion, vol. 21, pp. 725-733, Sep 2006.
[2] G. Bywaters, et al., "Alternative Design Study Report: WindPACT Drive Train”, NREL, Colorado, report No. NREL/SR-500-35524, Oct 2004.
[3] L. Fingersh, M. Hand, and A. Laxson, "Wind Turbine Design Cost and Scaling Model", NREL Technical Report, NREL/TP-500-40566, Dec 2006.
[4] Anon., "Value breakdown for the offshore wind sector", Renewables Advisory Board, Feb 2010.
[5] Anon., “Gears -- Thermal capacity -- Part 1: Rating gear drives with thermal equilibrium at 95 °C sump temperature”, ISO/TR 14179-1:2001, International Organization for Standardization, Geneva.
[6] Cotrell, J. "A preliminary evaluation of a multiple-generator drive train configuration for wind turbines." 21st American Society of Mechanical Engineers (ASME) Wind Energy Symposium. 2002.
[7] Carroll, J. McDonald, A., “Drivetrain Availability in Offshore Wind Turbines”, Abstract submitted to EWEA 2013
[8] Anon.,"A Guide to an Offshore Wind Farm", The Crown Estate, 2010.
[9] Anon., "Offshore Wind - Forecast of Future Costs and Benefits", Renewables UK, Jun 2011.
[10] S. Alshibani, V. G. Agelidis, "Issues regarding cost estimation of permanent magnet synchronous generators for mega-watt level wind turbines" IEEE International Electric Machines & Drives Conference IEMDC, Sydney, Austrailia, 2011.