Evaluation and development of normal year correction methodologies for icing climatology in wind farm applications

• Main Author: Klintström, Rebecka
• Format: Other/Unknown Material
• Language: English
• Published: Lunds universitet/Institutionen för energivetenskaper 2012
• Subjects: ice modeling normal year correction icing of wind turbines mesoscale meteorological models; Technology and Engineering; Northern Sweden
• Online Access: http://lup.lub.lu.se/student-papers/record/3054826

Wind power developments in cold and icing conditions pose new challenges to the wind power industry. Icing of wind turbines cause safety hazards, production losses and increased loads. This means that there is a need to develop a technique to quantify the icing climatology and corresponding eects on wind turbines to enable large scale wind power developments in the windy, sparesly populated areas of northern Sweden. This thesis is focused on evaluating the possibilities to estimate the long term icing climatology using mesoscale meteorological models in combination with existing ice models. The basis for the study were long term, low resolution time series from WRF and COAMPS in combination with high resolution, short term time series for the same geographical area, with the aim to nd a method for normal year correction of the icing climatology. A number of dierent methodologies for normal year correction of icing climatology were developed and evaluated using a few criteria: plausibility of the results, site dependency, sensitivity to data set used and sensitivity to the length of the time series used as reference. The ice modeling was performed using an implementation of the Makkonen algorithm for ice accretion. The normal year correction was performed using dierent theoretical reasoning and at dierent steps in the modeling. Linear regression of time series of standard meteorological parameters as well as of active ice hours and ice loads were tested. A method using iterative height adjustment of the parameters was developed as well as a basic method directly relating the high resolution modeled ice load to the low resolution modeled ice load. A method using averaged values for the dierent ice model input parameters was also tested. The evaluation showed promising results for the linear regression method. The dierent implementations showed feasible results and the outcome was not sensitive to the length of the correction basis. The height adjustment method was dicult to evaluate and was sensitive to the time ...