In the absence of easily exploitable, mild-climate onshore wind resources, numerous cold-climate sites around the world offer great wind-energy potential, with good wind resources and low population density. From 69GW at the start of 2013 — 24% of global wind-energy capacity — cold-climate sites have now reached 100GW, in line with or exceeding predicted growth, which forecast 119GW by 2017.
Yet, in demanding winter climates, many sites suffer from icing, which can cause the wind-turbine blades to reduce energy yield, the mechanical lifetime of turbines to fall, noise emissions to increase and safety risks to arise through ice throw, among other challenges. While solutions have been developed, such as antiand de-icing systems, more needs to be done to further cut the cost of wind energy in cold climates.
The decision on whether a wind farm is worth developing is based on the analysis performed during site prospecting and of the planning/site assessment phases, which represent only about 7% of all lifetime costs. It is very important to determine a realistic annual energy production (AEP) and define all major risks for the project prior to turbine acquisitions.
Standard turbines can typically operate down to -10ºC and have a survival temperature of -20 degsC, while low-temperature turbines may operate down to -30ºC and survive down to -40ºC. After the turbine purchase agreement is signed, the cost of changing components, such as installing retrofit ice protections systems, becomes extremely high.
The first step is to check if the site is both a low-temperature climate site and an icing-climate site or just one — active icing rarely occurs below minus 25ºC — according to the definition (see box) by the International Energy Agency (IEA) Wind Task 19 Wind Energy in Cold Climates, a research group tasked with safely and cost efficiently increasing wind power deployments in cold climates.
Some countries have publicly available icing maps, which are typically the first source of information about site icing severity. An icing site should be defined under the IEA ice classification to establish the severity of icing and its consequences. More detailed analyses is recommended in a later phase of the project.
While meteorological icing can be modelled numerically with mesoscale weather prediction models, the best way to obtain data for site classification is to measure the meteorological and instrumental icing directly at the site and, if possible, compare icing measurement of nearby wind farms for production experiences in icing conditions.
When only considering ice-induced production losses, two primary sources cause uncertainty: icing measurements, which can cause quite substantial uncertainty, even within the same wind farm; and turbine performance and control strategy in icing conditions, which can vary substantially from one turbine manufacturer to the next. For these two reasons, a range of average AEP losses including uncertainty is preferred in expert icing analysis.
Finland case study
To assess uncertainties in icing assessment and turbine control strategies, VTT Technical Research Centre of Finland analysed an operational Finnish wind farm. It used three icing maps to emulate a site-prospecting portfolio icing-risk analysis - as if the project had not been built yet. Two of the maps were based on mesoscale weather modelling and one on research from long-term observations of in-cloud icing.
The analysis produced a maximum scatter of results: the site was initially categorised into IEA ice classes 2, 3 and 4 for the different meteorological icing maps.
Two turbine manufacturers (A & B) were then compared in almost identical atmospheric conditions at the site. Ice-induced production losses were calculated using Task 19's free T19IceLossMethod software, which uses standard-turbine Scada data as input and calculates production losses due to icing.
Finally, meteorological icing was measured with a Labkotec LID ice detector, and instrumental icing was measured with several heated and non-heated anemometers at hub height.
The results showed that the site-prospecting icing map portfolio analysis connected to IEA ice classification overestimates measured production losses, although it is limited to Scada data for the past two winters.
The range on measured instrumental icing from more than three anemometer pairs and production losses due to icing within the same wind farm was almost +/-50% of the turbine A wind-farm average.
Turbine B had substantially smaller production losses due to icing than turbine A due to a different control strategy during iced turbine operation. We can conclude that it is therefore very challenging to be able to define a representative range of production losses from icing using only available icing maps and connecting them to the IEA ice classification. And, different turbine control strategies for turbines in icing conditions causes further uncertainties.
To minimise the uncertainty of estimating production losses due to icing, VTT recommends four actions:
- Adopt in-site prospecting, using portfolio analysis techniques with multiple icing maps if possible.
- If IEA ice class is greater than class 1 in the site prospecting, the resource-assessment phase should include at least one fully heated and one non-heated anemometers for one full winter season or more to define the instrumental icing duration at hub height.
- Perform long-term correction to short-term icing measurements as icing has an extremely high interannual variation.
- Lower turbine-specific uncertainties on standard turbines by demanding proof of performance and control strategy in icing conditions from the turbine manufacturer. For turbines with ice protection systems, assess their benefits, remembering that ice protection systems can reduce AEP uncertainties and so potentially reduce the cost of financing.
With these recommendations in mind, developing projects in cold-climate sites can be more attractive and the financial risks can be better managed. With proper planning and knowledge, one can harness offshore-like cold-climate wind resources at near-onshore costs.
Meteorological icing Suitable for ice accretion/formation
Instrumental icing Ice remains at a structure and/or an instrument remains disturbed by ice
Incubation time Delay between start of meteorological icing period and instrumental icing period
Recovery time Delay between end of meteorological period and end of instrumental icing - ice remains but is not formed
Definition of meteorological icing and instrumental icing
Ville Lehtomaki is a wind-power research scientist at the VTT Technical Research Centre of Finland. He is speaking at the Windpower Monthly Cold Climate Forum, 10-11 December in Helsinki, Finland