Wind turbine generators: DFIGs or Rare Earth magnets?

wind turbine generators

Wind turbine generators can use doubly fed induction generators (DFIGs) or permanent magnet synchronous generators (PMSGs) based around Rare Earth metals. This data-file captures the trends in DFIGs vs PMSGs over time by tabulating 40 examples, as turbines have grown larger, and different wind turbine manufacturers have adopted different strategies.


How do wind turbines generate electricity? Both doubly fed induction generators (DFIGs) and permanent magnet synchronous generators (PMSGs) retain a balanced market share in modern wind turbines, converting rotational energy into electrical energy.

For an overview of the different designs, please see our overview of magnets, overview of electricity and overview of wind power-electronics.

This data-file tabulates data into forty recent wind turbine designs, sampling across all of the leading wind turbine manufacturers. We have tabulated the model, manufacturer, country, capacity (MW), year introduced (YYYY), rpm, gear system, typical voltages (kV, where available), generator system and other power electronic details.

Capacity trends. The maximum power capacity for DFIG turbines has been hovering around the 6-8MW mark for the past decade, while the trend towards larger 10-15MW turbines, especially offshore, is almost always associated with PMSG turbines.

Why do wind turbines use Rare Earth magnets? The short answer is that Rare Earth magnets have greater magnetic field strengths (flux densities), which opens up direct drive generator configurations with much lower gearing, and without requiring their own input power supply like the electromagnets in DFIGs do.

Advantages of PMSGs are cited by different operators, including higher >96% efficiency, fault ride-through, lower maintenance, more compact nacelles, easier installation.

Elimination of gears is an advantage for PMSGs as it can help to avoid maintenance issues, which typically costs $40/kW of capacity and 1-2c/kWh on levelized cost. For geared turbines, the gearbox is by far the largest source of maintenance issues (data here).

About two-thirds of the PMSGs in our data-file use direct drive to impart rotation into their power-dense magnets (no gears). Almost all of the IGBT turbines use gears, to step up the rotational speed by around 100:1, on average. Some PMSGs retain gears in order to lower the amount of Rare Earth materials/magnets required by up 90% (e.g. Vestas).

More compact turbines are associated with PMSGs, and this is an advantage, as it lowers the costs associated with wind turbine installation, and faster commissioning.

Fault ride-through is one of the most commonly cited issues for using Rare Earth PMSGs rather than DFIG turbines. A DFIG requires a power connection to magnetize the electromagnets in its rotor. If the power input drops, then the power output also drops.

Chinese manufacturers have been gravitating towards PMSGs especially since 2016. For example when Goldwind switched from DFIGs to PMSGs, the company noted it was able to eliminate 13 gears and “hundreds of parts”.

Vestas and GE have been relatively vocal about benefits of permanent magnets. GE back to 1998 (!). Vestas is now using Rare Earth elements in all new grid turbine models (per the Vestas website here).

Conversely, two other European manufacturers stand out as they have been more reluctant to use permanent magnets, and more focused on DFIG designs, especially in smaller turbines and onshore turbines.

Our forecasts for Rare Earth magnet use in future wind turbine’s bill of materials, both in mass terms (kg/kW) and in cost terms ($/kW) is modelled here.

Underlying data behind these observations is set out in the data-file of wind turbine generators. If you want to understand how magnets in wind turbines work, we recommend our overview of magnets. All of our broader wind research is here.

Offshore vessels: fuel consumption?

This database tabulates the typical fuel consumption of offshore vessels, in bpd and MWH/day. We think a typical offshore construction vessel will consume 300bpd, a typical rig consumes 200bpd, supply vessels consume 150bpd, cable-lay vessels consume 150bpd, dredging vessels consume 100bpd and medium-sized support vessels consume 50bpd. Examples are given in each category, with typical variations in the range of +/- 50%.

Wind power: energy costs, energy payback and EROEI?

Wind power energy paybacks

Wind power energy paybacks? This data-file estimates 3MWH of energy is consumed in manufacturing and installing 1kW of offshore wind turbines, the energy payback time is usually around 1-year, and total energy return on energy invested (EROEI) will be above 20x. These estimates are based on bottom-up modelling and top-down technical papers.


The average wind energy project has an energy intensity of 3MWH/kW, which is repaid after c1-year, for a total energy return on energy investment above 20x, over a 20-25 year operating life.

One observation from reviewing technical papers is that many have rough methodologies. Some are still basing numbers upon small, <1MW turbines, which are no longer representative. Conversely, others are incomplete, and have not fully captured materials costs.

Hence we have built up our own bottom-up estimates for the energy intensity of wind power, and the EROEI of wind turbines.

Our bottom-up estimates for the energy costs of wind turbines are based on a full bill of materials, economic models of those materials (e.g., glass fiber, carbon fiber, epoxies, steel, copper), data into the vessel days per turbine, and the fuel consumption of different vessels.

Our bottom-up estimates for wind power EROEI also captured power transmission, curtailment considerations and maintenance requirements.

The largest individual contributors to the up-front energy costs of wind turbines are transporting materials to the site (0.75MWH/kW), steel (0.6MWH/kW), other materials (0.3MWH/kW), large offshore vessels that install foundations and turbines (0.3 MWH/kW) and the tail of 20-40 smaller vessels that support offshore operations (data here).

The average CO2 intensity of wind turbines is suggested at 10-20g/kWh (0.01-0.02kg/kWh). This coheres with the technical papers that we reviewed, and our own bottom-up estimates.

Wind power energy paybacks will vary with individual project parameters, and we think that a realistic range for offshore wind projects is 15-30x EROEI.

The most important parameter is the location of the project, which will determine energy generated per year, but also transportation distances and steel requirements.

Comparable data for solar assets is linked here.

Renewables: how much time to connect to the grid?

time to connect to the grid

Is the power grid becoming a bottleneck for the continued acceleration of renewables? The median approval time to connect to the grid for a new US power project has climbed by 30-days/year since 2001; and has doubled since 2015, to over 1,000 days (almost 3-years) in 2021. Wind and solar projects are now taking longest to inter-connect, due to their prevalence, lower power quality and remoteness. This data-file evaluates the data, looks for de-bottlenecking opportunities, and wonders about changing terms of trade in power markets.


Accelerating wind and solar are a crucial part of our roadmap to net zero. But we have also been worrying about bottlenecks, especially in power grids. Project developers are increasingly required to fund new power transmission infrastructure, before they are allowed to interconnect, usually costing $100-300/kW, but sometimes costing as much as the renewables projects themselves (data here). If there is one research note that spells out the upside we see in power grids and electrification, then it is this one. We also see upside in long-distance transmission, HVDCs, STATCOMs, transformers, various batteries.

Other technical papers have also raised the issue of rising interconnection times and power grid bottlenecks for wind and solar. And the US’s Lawrence Berkeley National Laboratory has also started tracking the ‘queue’ of power projects waiting to inter-connect. We have downloaded their database, spent about a day cleaning the data (especially the dates), and aimed to derive some conclusions below.

Methodological notes. The raw LBL database contains a read-out of over 24,000 US power projects that sought to inter-connect to a regional power grid, going back to 1995. However, 13,000 of these applications were withdrawn, 8,000 are still active/pending and 3,500 are classed as operational. 2,500 of the projects have complete data on (a) when they applied for permission to inter-connect to the grid and (b) when they were ultimately granted that permission, allowing us to calculate (c) the approval time (by subtracting (b) from (a)). But be warned, this is not a fully complete data-set. And some States, which have clearly constructed large numbers of utility-scale power projects, seemed not to report any data at all. Nevertheless, we think there are some interesting conclusions.

The median time to receive approval to inter-connect a new US power project to the grid has risen at an average rate of 30-days per year over the past two decades and took over 1,000 days in 2021, which is 2.8 years. This has doubled from a recent trough level of 500 days in 2015 (chart below) and a relatively flat level of 400-days in the mid-2000s.

Renewables projects now take longer to receive approval to connect to the power grid. Wind projects have always taken longer to receive approvals. And recent wind projects continued taking 30% longer than the total sample of approved projects in 2019-21. More interestingly, however, solar projects have gone from taking 50% less time to receive grid connection approvals in the mid-2000s to taking 10% longer than average, especially in 2020 and 2021. Why might this be? We consider five factors…

#1. Project quantity is probably the largest bottleneck. The numbers of different projects receiving permission to connect to the grid are tabulated below. A surge in wind projects in 2005-2012 correlates with the first peak in inter-connection approval times on the chart above. And a more recent peak in utility-scale solar, battery and wind projects correlates with the recent peak of approval times in 2020-21. This suggests a key reason it is taking more time to approve new inter-connections is that grid operators are backlogged. It would be helpful to resolve the backlog. And we wonder if the result might be a change in the terms of trade: favoring grid operators more, favoring capital goods companies more, and requiring project developers to be more accommodating?

#2. Project sizing does not directly explain inter-connection approval times. The average utility-scale solar project has become larger over time (now surpassing 150MW). But wind projects have always been larger than the average power project seeking approval to connect to the grid. And there are many small gas, coal and nuclear projects that take longer to receive connection approval than large ones. So we do not think there is a direct link between power project size and the time needed to approve an inter-connection. However, there may be an indirect link. It is clearly going to take longer to study the impacts of connecting 10 x 100 MW solar projects (in 10x separate locations), than 1 x 1,000 MW nuclear plant, even though both have the same nameplate capacity.

#3. Connection voltage does not explain inter-connection approval times. The median project in the database is connecting into the grid at 130kV. The median wind project is at 145kV. The median gas project is at 135kV. The median solar project is at 110kV. The median battery project is connecting at 140kV. Although we do think that moving power over longer distances is increasingly going to favor higher voltage transmission and also pull on the transformer market.

#4. Power quality seems to explain relative approval times and increasingly so. Another interesting trend is the difference in interconnection approval times between different types of power projects. Wind and solar projects now take 30% and 10% longer than average to receive approvals. Whereas gas, batteries and hydro now take 15%, 50% and 90% less time than average to receive approvals. We think this is linked to power quality. On a standalone basis, wind and solar may tend to reduce the inertia, frequency regulation, reactive power compensation and balancing of power grids. Whereas gas power plants, batteries and hydro typically help with these metrics (each in their own way). We think this adds evidence in support of our power grids thesis.

#5. Remote projects take longer to approve, as they will likely require more incremental transmission lines. The shortest interconnection times across all power projects were in Texas, which already has a very large power grid, arguably the best energy endowment and infrastructure in the world. But other more densely populated states (Michigan, Illinois) tended to have 50% lower times to approve inter-connections than some of the least densely populated states (the Dakotas, Iowa, Montana), where we think new power generation likely needs to be moved further to reach demand centers. Location matters for levelized cost of electricity. Again, we think this evidence also supports our power transmission thesis.

Overall the data suggest that there are growing bottlenecks to inter-connect renewables to power grids; especially in areas with a surge of activity, where power quality is increasingly important, and in more remote areas that require new transmission infrastructure. We think this trend will continue. It would be helpful to debottleneck the bottlenecks, to sustain the upwards trajectory of wind and solar. But we do think the terms of trade are shifting in favor of grid operators, power electronics, transmission infrastructure, developers that can use their own power and consumers that can demand shift.

Offshore wind: installation costs by vessel?

Offshore wind installation cost

An offshore wind project is likely to cost $2,500/kW, of which c$1,500/kW is turbines and $1,000/kW is offshore wind installation costs. This data-file aims to estimate the breakdown by vessel type, day-rates and costs per turbine.


Estimates range from 25 to 100 vessels being deployed in the installation of a typical offshore wind project, across different studies and technical papers.

Our base case estimate is that each turbine will require 10 days to install, consuming 100 “vessel days”, of which c10% are highly specialized vessels (data here).

The highest day-rates will be commanded by the most specialized vessels, and we think this includes wind turbine installation vessels, cable lay vessels (both intra-wind farm and for the export cable to shore), and foundation vessels that drive monopiles or jackets/piles into the seabed.

Offshore installation costs are sensitive and can realistically range from $800-2,500/kW, depending on project parameters.

These installation costs come on top of the underlying turbine costs, which we think are usually going to be around $1,500/kW per our cost breakdown of a wind turbine.

Economies of scale are achieved by using larger turbines in larger projects. Conversely, costs re-inflate when moving further from shore and into deeper water.

Please see our broader research into wind and solar and the ramp-up of renewables in our roadmap to net zero.

Please download the data-file to stress-test the numbers and see our estimates for the breakdown of offshore wind installation costs by vessel type.

Goldwind: frequency response from wind turbines?

Goldwind frequency response

Goldwind is one of the largest wind turbine manufacturers in the world, having delivered over 50,000 turbines by early-2023. The company was founded in 1998, headquartered in Beijing, it has 11,000 employees, and shares are publicly listed.

The wind industry is increasingly aiming to mimic the inertia and frequency response of synchronous power generators. Goldwind has published some interesting case studies, boosting power by 6-10%, within 1-2 seconds, for 5-6 seconds. Hence this data-file reviews Goldwind frequency response patents to look for an edge?


Synchronous power generators have ‘inertia’. If the grid suddenly becomes short of power, due to a large new load switching on, or a large power source disconnecting, then all of these synchronized machines will slow down very slightly, and in unison. Harvesting their rotational energy provides more power to the grid. But it also lowers the ‘frequency’ of the grid. At least for a few seconds, until firing rates can be ramped up. Generally, the larger the synchronous power generators, the higher their ‘inertia’, the more power that can be drawn out as they slow down, and the less grid frequency will drop per unit of incremental power. This is all crucial to the functioning of a modern power grid.

Wind turbines do not inherently provide any inertia or frequency response. Interestingly, they have about the same amount of angular momentum as conventional power generators, as they spin. But it is not synchronized with the grid. A wind turbine spins at 15-20 revolutions per minute, whereas a typical grid at 50 Hz (or 60 Hz in the US) is completing 50 full AC cycles per second.

An increasing goal for the wind industry is to ‘mimic’ some of the inertia and frequency responses of synchronous generators. Recently, Goldwind has published interesting case studies at wind farms, e.g., in Australia, boosting power by 6-10%, within 1-2 seconds, for 5-6 seconds (chart below).

The goal of this patent screen is to evaluate whether Goldwind’s frequency response algorithms give it an edge over other wind turbine manufacturers. We were unable to de-risk this idea from reviewing the patents, for the reasons explained in the data-file.

There are three key challenges for implementing frequency responses, as a form of ‘synthetic inertia’ and wind power plants, which are also borne out by the patent analysis. It takes time to implement the response (certainly longer than a battery, capacitor bank or conventional generator). Second, it is challenging to coordinate the responses of dozens of turbines, each with a different operating state. Third, there is always a danger that the wind drops at the precise moment you want to implement your frequency response, due to natural wind volatility.

Goldwind frequency response? Goldwind discusses possible solutions to each of these challenges in its patents. The bright spot is that we think many of the solutions are software-side, which will lower their implementation cost, and not pull too hard on already-bottlenecked power electronics. However, as usual, we find it harder to de-risk algorithm-heavy patents.

Offshore wind: installation vessels and time per turbine?

Offshore wind installation vessel time per turbine

Wind turbine installation vessels are estimated to cost $100-500/kW in the breakdown of a typical offshore wind project’s capex. Total offshore construction time is around 10 days per turbine. Offshore wind installation vessel time per turbine averages around 5 days per turbine. Data from past projects are tabulated in this data-file.


The average offshore wind project comprises 90 turbines. Hence over the entire course of offshore construction, a full turbine is installed every 10-days on average.

However, over 50% of the time is spent on preparations, foundations, cable-laying and commissioning. The average wind turbine installation vessel installs a full turbine every 5-days.

The numbers vary by project. The best installation vessels operating under the best conditions can install an entire turbine upon pre-laid bases/foundations in a single day.

At the other end of the spectrum, installation can take over ten days per turbine, especially amidst bad weather, delays with parts, mechanical issues, and projects that are a long way from ports.

Offshore wind installation vessel time per turbine? Data on the times taken to install an offshore wind turbine are aggregated across 35 offshore wind projects in this data-file. While the data are interesting, there are few correlations to be drawn between simple headline metrics.

Some eagle-eyed readers will have spotted, in the charts above, that some projects appear to have higher WTIV days per turbine than total construction days per turbine, which seems confusing. The reason is that recent large offshore wind projects will tend to have 2-3 installation vessels operating simultaneously. Thus they can achieve 2-3 WTIV-days of work per calendar day. More vessels will mean faster construction. And probably lower total costs. But it will also require booking out more vessels from the total fleet.

Overall, we think that installation vessels and support vessels could be a bottleneck for accelerating offshore wind. Amazingly, some of the largest projects in the history of the industry had 60 – 100 vessels operating at peak construction activity.

For perspective, recent offshore wind projects likely have total installation costs of $2,000-6,000/kW (offshore wind cost data here, breakdown of turbine cost here). Cost is saved by larger turbines. The other large bottleneck is in downstream power electronics for integrating wind with power grids.

Leading companies in offshore wind turbine installation vessels include DEME, MPI, Van Oord, Cadeler, Eneti (Seajacks), and increasingly, private-equity backed vessels gearing up for an expansion of offshore wind, especially in the US.

Cable installation vessels: costs and operating parameters?

database of cable installation vessels

This is a database of cable installation vessels for offshore wind and power transmission; tabulating costs (in $M), contract awards (in $/km), capacity (in tons), installation speeds (in meters per hour), power ratings (in MW), crew sizes, positioning systems and leading companies. But there is a paradox in the past decade’s cost data?


The paradox is that the world’s fleet of offshore cable installation vessels have become ever more sophisticated over time, and yet the costs of offshore cable laying do not appear to have risen to reward this build-out?

For example, at the cutting edge, Prysmian’s Leonardo Da Vinci vessel cost around $200M to build, has 21MW of total power, 180 tons of lifting capacity, and thus can lay a staggering 2.1 km of cables per hour, in water depths up to 3,000 meters, with ultra-redundant DP3 positioning. Similarly, Van Oord’s new Calypso vessel will be able to lay two cables at once.

Generally these cutting edge vessels also boast improved environmental performance, with less CO2, NOx, hybridization or readiness to run on biofuels. The details are noted in the data-file, vessel by vessel.

By contrast, many of the cable laying vessels built a decade ago only had 7-8MW of total power on average, 110 tons of lifting capacity, maximum lay rates of 1km of cables per hour, in water depths up to 800m, with DP2 positioning. So the industry has truly transformed its capabilities.

And yet the contract awards that we have aggregated for offshore cable installation have hardly changed. Laying the inter-cabling at an offshore wind project might cost $0.5M/km, offshore interconnectors (such as HVDCs) cost $1-2M/km, and complex projects with erratic seabed terrain cost as much as $3-5M/km.

Is it possible that attempts to accelerate offshore wind and renewables more broadly will pull on the supply chain for cable installation vessels, and rescue what has thus been a relatively challenging industry? If our energy supply-demand numbers are right, there could even be another offshore cycle in the conventional energy industry?

The full database of cable installation vessels covers assets owned by Prysmian, Van Oord, Nexans, Deme, NKT, Jan de Nul, Seaway, Boskalis et al; and contracts whose details have made it into the public domain.

Wind volatility: second by second output data?

Second by second volatility of wind power

We have aggregated the power output and power drops across a month of second-by-second wind data, from a 25MW onshore wind farm in Germany (12 x 2MW turbines). We want to use the data to understand the typical second-by-second volatility of wind power, and what kind of power smoothing will be needed to ramp wind power in the energy transition.


As background, our recent research has evaluated the feasibility of powering real-world grid loads using solar power using a similar data-set quantifying the second-by-second volatility of solar power.

We were hoping that the second-by-second volatility of wind power would be less volatile. However, if anything, wind power is more volatile than solar. Volatility events for wind and solar are similarly frequent, but for wind, they tend to be deeper and last longer.

In the average day, the onshore wind farm that we assessed saw 75 power volatility events, where output fell by over 10% from its trailing ten-minute average level (example below). 40% of these volatility events lasted less than 10-seconds. 80% of the volatility events lasted less than 60-seconds.

Second by second volatility of wind power

What is challenging is that the remaining 20% of volatility events (by number) could often last for hours or even days (see below). Indeed, our data-set includes a four-day timeframe with effectively zero wind power generation (examples in the data-file). By contrast, the sun does tend to rise every day, and generates some power even on cloudy days in the dead of winter.

Second by second volatility of wind power

Wind volatility is also more binary than solar volatility, as is visible in the first chart above. Some striking numbers are that solar output tends to drop off by over 90% from its trailing ten-minute average level an average of 0.02 times per day (i.e., very rarely). Whereas wind output dipped by over 90% below its trailing ten minute average level, an average of 10 times per day (i.e., the wind can stop blowing, entirely, and this occurs relatively often).

Arguably, our data-set is also under-estimating volatility. Some days in our data-file showed very few volatility events, simply because wind output was precisely zero for several consecutive hours (example above). Yes, from a purely statistical perspective, this means that a wind power plant can produce the exact same power output for several hours at a time. But so can a potato.

On the other hand, some of the days with high wind output were significantly more volatile than the average day in our sample. For example, the chart below shows a day where output reached nameplate capacity of 25MW. But this day had 163 ‘volatility events’.

Second by second volatility of wind power

We clearly want to ramp both wind and solar power up as much as possible in the energy transition. But powering real world grid loads with the volatile output from both wind and solar is going to create opportunities and challenges in power grids. We think the best solution that is going to emerge is in power infrastructure, inter-connecting large areas, with resilient and long-distance transmission lines.

Definitions. Our wind analysis defined a “power drop” as a >10% reduction from the trailing 10-minute average wind power being generated at the asset. The power drop’s duration is defined as the number of seconds until output returns to 90% of that previous 10-minute average power output. The average power drop is the average reduction in power output during the duration of the power drop, compared to the prior trailing ten minute average. And the peak drop is the maximum reduction in power output during the power drop, compared to the prior trailing ten minute average.

Data sources. It is surprisingly hard to find good, second-by-second data for large, modern wind farms. However, this analysis has been based on an excellent technical paper and underlying data-set into short-term volatility of renewables. Source: Anvari, M., Lohmann, G. Wächter, M., Milan, P., Lorenz, E., Heinemann, D., Tabar, M. R. & Peinke, J. (2016). Short term fluctuations of wind and solar power systems. New J. Phys. 18, 063027.

Wind power: operating costs?

Wind power operating costs

Wind power operating costs. A breakdown of wind turbine opex is build up in this data-file, using granular data from technical papers, and our other models.

We think a typical wind turbine costs $40/kw per year to run and maintain, equivalent to 1-2c/kWh of opex, depending on the load factor.


The data-file shows how costs can vary, as a function of inputs. A sensible range is $25-75/kW-year. And the single largest way to lower costs is through up-scaling, spreading relatively fixed costs across larger turbines and larger overall assets. Our recent note on up-scaling wind turbines is linked here.

Wind turbine maintenance market? We estimate that around $25/kW of our cost build-up can be described as ‘maintenance’. Hence across 825GW of global wind capacity (at the end of 2021), the wind turbine maintenance market would be worth >$20bn per year.

This maintenance ‘after-market’ is typically controlled by wind turbine manufacturers. Larger operators can lower costs by self-managing, albeit this comes at the risk of lower uptime.

The data-file includes a build-up of maintenance costs across 19 different categories — drives, generators, hydraulic systems, blades, hubs electricals (chart below)– multiplying their annual failure rates by the sum of their replacement costs (i.e., parts) and the time spent by technicians.

Wind power operating costs

Another surprisingly large cost component is lubricants, around $1-2/kW-year.

Full details on wind power operating costs are split out across 6 tabs, including our highlights from the best technical papers that crossed our screen in the ‘notes’ tab. We have also drawn on other data-files assessing wind turbine capex costs, typical plant sizes, land use, and land costs.

Key conclusions on the opex of a wind power project were also highlighted in a research article sent out in August-2022 to our distribution list.

Copyright: Thunder Said Energy, 2019-2023.