Umeå in Sweden is a coastal city situated about 600km north of Stockholm and about 400km south of the Arctic Circle. This year, it hosted the annual Winterwind International Wind Energy conference. With its proximity to the Arctic, this is a very apt location for such a niche event which focuses on the challenges of generating wind in cold climates.
The conference comprised of 54 oral and poster presentations by wind energy professionals from universities, manufacturers, developers, consultants, investors and wind farm owners, each providing an element of insight into the wind industry in cold climates. DNV GL was well represented during the conference sessions leading four engaging conference sessions. The topics covered in the presentations discussed: The blockage effects in stable regimes associated with cold climate; Using 1Hz data to monitor turbine integrity; Improving short-term forecasting of turbine icing using machine learning; and a validation study of WICE, the method used to estimate production losses due to icing in site assessments.
It’s well documented that cold climates represent an encouraging opportunity for wind energy expansion, largely due to sparse populations and favourable wind conditions. But these cold climates pose a significant challenge: icing. The icing of rotor blades and other components can impact turbine production due to ice accumulation or other safety concerns. The icing effect on turbines can vary greatly depending in different regions throughout Northern Europe, creating significant challenges for operational turbines.
In the Nordics, ice loss estimates have formed an important part of an energy performance assessment for many years. In such a competitive market, a few percentage points of improvement can make all the difference as to whether a project goes ahead or not. Therefore, an accurate icing loss assessment is vital to validate pre-construction wind farm energy output predictions. But icing is still a relatively ‘new beast’ and mitigating its impact is a real test for a still relatively infant wind industry. It’s particularly infant in terms of the development and operation of wind farms in cold climates.
During the conference our experts spoke about our recent step forward in our ability to predict production losses when utilising a tool called ‘WICE’. WICE is a model that predicts production losses due to icing based on a combination of atmospheric modelling and machine learning. In the training phase, high fidelity mesoscale model data and SCADA data are used. When the model is applied, no local measurements are needed. All necessary atmospheric conditions are modelled in site specific mesoscale model studies, which is one of the strengths of the model. In parallel with the validation project, we’ve upgraded several parts of the model chain, namely the machine learning setup used, the processing of the production data, the addition of more training sites, and the long-term corrections method. When fully developed, the model will be applicable worldwide and able to predict the benefit of Ice Protection System in reducing icing losses.
A key take-away from the event was how the industry is advancing in this area With focused events like Winterwind bringing together an entire industry to challenge current industry insight and accelerating knowledge, the difficulties associated with developing and operating wind turbines will be overcome to ensure that the energy we all use comes from clean and sustainable power sources.