主催: The Japanese Society for Artificial Intelligence
会議名: 2012年度人工知能学会全国大会(第26回)
回次: 26
開催地: 山口県山口市 山口県教育会館等
開催日: 2012/06/12 - 2012/06/15
This report identifies direction, speed and duration of a solar radiation distribution due to weather conditions, especially a cloud movement, from distributed observations of solar radiation. 25 observations around Osaka area are used to analyze spatial correlation patterns with delay, and the results are compared to whole-sky photographs. The correlation patterns differ according to cloud patterns and motions, yet when a motion of cloud is observed, the speed and direction of the cloud motion can be identified from the correlation pattern. Using the information above, a detailed map of solar radiation (spatial and temporal resolution of 500 meters and 5 minutes) could be drawn by spatio-temporal Kriging. The presented method parameterizes the direction and speed of a cloud cover ratio, thereby incorporating its directionality. Cross-validation with spatial Kriging shows the advantage of the spatio-tempral Kriging when a motion of cloud is ovserved.