2014 年 80 巻 812 号 p. TEP0092
For the purpose of preparing for the further diffusion of wind power technology, the power reduction risk of wind power plants, i.e. a risk of frequent and large reduction in the output caused by time-varying wind speeds, needs to be decreased. Among measures against the risk, the geographical distribution of wind power plants needs to be examined first because additional facilities, such as backup power plants or electricity storages, are not required. This study quantitatively analyzes the risk reduction effect of geographically distributing wind power plants among 12 locations within the Hokkaido. The mean-CVaR (Conditional Value at Risk) model is adopted to estimate the optimum geographical distribution. We found that the share of time period with no power output can be reduced to mostly 0% by optimizing the geographical distribution while the share is more than 10% if all the wind power plants are intensively installed to the location with the highest average output. Another finding is that the geographical distribution can contribute to the risk reduction even among the locations in the Hokkaido that is a relatively smaller region compared with the regions surveyed by earlier studies. We also found that the CVaR is the better index of the power reduction risk of wind power plants than the standard deviation as far as focusing on the power reduction risk.