More Related Content
Featured (20)
Greenbyte - The Effects of Cold Weather on Wind Data Quality
- 1. The Effects of Cold Weather on Wind Data Quality An empirical study on how data produced from met mast and SODAR is affected by cold weather Oscar Winter, CTO Greenbyte AB Winterwind 2012, Skellefte奪 2012-02-08
- 2. Introduction Wind measurements are performed in order to estimate production levels in wind power projects. Properly performed wind measurements decrease risk in investment decisions. Wind measurement equipment is affected by cold climate. This can result in long periods of unreliable data that increases risk in bankable data and uncertainty in production estimations.
- 3. Table of contents 1. Summary 2. Purpose 3. Results 4. Why are cold climate measurements problematic? 5. SODAR case study 6. Met mast case study 7. Key Takeaways
- 4. Summary (1/3) A total of 107 sites all over Sweden have contributed data to the study The data covers 2010 and 2011 Data from both SODAR and met masts have been analyzed
- 5. Summary (2/3) Sweden Northern Sweden Total 37 sites 62 属 north Southern Sweden Total 70 sites
- 6. Summary (3/3) Met masts Total 19 sites SODARs Total 88 sites Images are unrelated to the actual data.
- 7. Purpose Bring awareness concerning problems in cold climate wind measurements Demonstrate cold climate problems that can occur during a wind measurement Discuss SODAR and met mast specific problems Discuss if and how cold climate problems can be minimized or prevented Discuss if and how data affected by cold climate should be post processed to increase reliability and accuracy
- 8. Data from all sites 2010-2012 120 30 100 20 Data availability(%) Temperature (属C) 80 10 60 0 40 -10 20 -20 0 -30 Januari Februari Mars April May June July August September October November December Met mast DA 99,45 98,95 99,94 99,7 99,85 99,81 100 99,99 99,99 99,45 98,49 99,44 Met mast DA icing 86,95 88,73 95,31 98,61 99,85 99,81 100 99,99 99,99 98,68 91,77 86,49 SODAR DA 88,77 85,78 95 96,1 96,75 97,24 97,92 95,8 96,11 96,11 96,82 95,06 Temperature -5,25 -7,41 -0,64 4,67 8,79 13,4 16,72 14,09 10,58 5,74 1,47 -3,71
- 9. Data from northern sites 2010-2012 120 30 100 20 Data availability(%) Temperature (属C) 80 10 60 0 40 -10 20 -20 0 -30 Januari Februari Mars April May June July August September October November December Met mast DA 97,76 98,57 100 100 100 99,3 99,99 99,98 99,99 98,13 97,32 97,89 Met mast DA icing 83,29 77,75 95,55 98,08 100 99,3 99,99 99,98 99,99 96,73 84,78 78,43 SODAR DA 83,63 76,95 90,07 93,01 93,21 96,87 97,88 95,77 96,08 97,62 97,42 92,99 Temperature -7,53 -11,16 -1,88 4,91 8,15 13,24 16,43 12,98 9,57 4,42 0,21 -6,19
- 10. Data from southern sites 2010-2012 120 30 100 20 Data availability(%) Temperature (属C) 80 10 60 0 40 -10 20 -20 0 -30 Januari Februari Mars April May June July August September October November December Met mast DA 99,9 99,09 99,93 99,63 99,81 100 100 100 99,99 100 98,96 100 Met mast DA icing 87,93 92,64 95,25 98,74 99,81 100 100 100 99,99 99,48 94,56 89,37 SODAR DA 92,1 93,63 98,29 98,11 98,52 97,45 97,94 95,81 96,14 94,68 96,15 94,94 Temperature -4,79 -6,07 -0,35 4,61 8,94 13,46 16,83 14,54 10,93 6,13 1,89 -3
- 11. Summary of results: Data availability Data source Sites Cold months Warm Warm-Cold (Nov Feb) months (difference) Met mast DA North 97.89 % 99.84 % 1.79 Met mast DA icing North 81.06 % 99.03 % 17.64 SODAR DA North 87.75 % 96.38 % 7.32 Met mast DA South 99.49 % 99.92 % 0.43 Met mast DA icing South 91.13 % 99.16 % 8.03 SODAR DA South 94.21 % 99.92 % 2.91 Met mast DA All 99.08 % 96.38 % 0.76 Met mast DA icing All 88.49 % 99.03 % 10.55 SODAR DA All 91.61 % 99.84 % 4.77
- 12. Summary of results 100% data availability is defined as 100% available wind speed data at 100m (inhomogeneous data set) Met mast data availability icing is the data availability with invalid/iced met mast data removed A simple icing filter: Any time interval of at least 4h during which the value of average wind speed changes less than 0.25 and the average temperature is less than 2属 C Using several icing filters in combination will detect even more data as affected by icing SODAR data is intrinsically filtered using a quality filter within the SODAR (invalid data values are marked or discarded) If icing is considered, data availability is lower during cold months (November - February) If icing is considered, data availability for the northern sites is less than that of the southern sites SODAR has slightly higher data availability during cold months Ignoring cold climate effects will negatively affect wind measurement equipment negatively affect measurement data Increases risk in bankable data Less reliable investment decisions
- 13. SODAR cold climate problems SODAR - Sound Detection And Ranging. Uses sound waves to measure wind. Operating range approximately 20 200m Less reliable at low and high heights SODAR cold climate problems Sound waves are affected by snow, sleet and rain (both ongoing and accumulated) Snow and ice may introduce problems in antennas if it accumulates in the antenna compartment (disrupts sound waves) Solar panels less efficient during winter time (less sunlight), more dependent on other generators (diesel). Generators are themselves prone to cold climate effects Power supply errors caused by heating failure (SODAR operating temperatures approximately -30 属C to +50 属C) Affected by general cold climate problems (wear and tear, engines and moving parts are less reliable if temperatures are freezing)
- 14. SODAR case study Battery voltage Wind speed over time Battery voltage over time Case: Battery discharged in 24 hours. More power required during cold time periods. Generator not starting due to low temperature (less than -20属 C)
- 15. SODAR case study Snow/ice Wind speed from 3 heights over time 3 days of affected data Snow and ice affect data availability (gaps in the graph) More than 50% of the data is invalid or missing Data availability decreases when the measurement height increases (e.g. less data available for higher heights)
- 16. Met mast cold climate problems Met mats are constructed in different ways and have different sensors Some configurations can be more error prone than others Met mast cold climate problems Sensors are affected by snow, ice, sleet and rain (both ongoing and accumulated) Snow and ice may introduce problems in the sensors, they can freeze stuck or slow down Equipment and heated sensors require more energy in cold climate General wear and tear increases during winter time
- 17. Met mast case study Frozen sensor Wind speed over time More than one week of affected data Non-heated anemometer frozen (flat line at the bottom) Heated anemometer is still working properly
- 18. Met mast case study Wind vane icing Wind direction over time 3 days of affected data Wind vane completely frozen (flat line) Exactly the same wind direction is reported for a long period of time
- 19. Met mast case study Iced sensors Wind speed over time More than one week of affected data Anemometers are frozen or slowed down due to snow/ice Problematic case, hard to determine which data is valid / correct (if any)
- 20. Key Takeaways (1/2) It might not be feasible to fully prevent all cold climate effects on measurement data. However, measures can be taken to prevent major loss of data if and when such effects occur. Utilize an external heating system (diesel generator). This will reduce wear and tear and help to keep equipment within operating temperatures. Closely maintain equipment and monitor measurement data, more so during cold climate periods Use proper monitoring software to monitor equipment and data. It can alert you of problems before they occur or get out of hand. If possible, try to have a contact person with physical access to the equipment who does regular checkups. Increase the number of checkups during cold climate periods . Use both a met mast and a SODAR / LIDAR in combination to decrease risk of loosing data. They are affected differently by cold climate. If one stops working, the other might still function well. For met masts, use both heated and unheated sensors and use several different types of anemometers (cup, aerovane, sonic).
- 21. Key Takeaways (2/2) Measurement data affected by cold climate should be post processed or the results based on calculations from such data will be inaccurate. No concensus about the best method. It depends on the data source, the purpose of the measurement, the target audience and the availabile options for this specific measurement. Two common methods for post processing data 1. Discard data - If only data from a short period of time is affected, simply discard the affected data and adjust future calculations 2. Gap fill data - Replace affected data with valid data from one or several other data sources (data transformation might be necessary depending on the situation) Data from another sensor on the same or close height Data from another SODAR or met mast close by Data from metrological and statistical models
- 22. Sponsors Several organizations have contributed data to this study. They wish to remain anonymous. Without their help this study would not have been possible. Thank you! (You know who you are)
- 23. Thank you for listening! Oscar Winter CTO, Greenbyte AB Kristinelundsgatan 16, 411 37 Gothenburg, Sweden Do you have any questions about the study? Please come and see us in the Greenbyte monter after the presentation and we will gladly answer them!