The output of the wind power depends on natural phenomena, so the influence on the connection for power system is of particular concern. Therefore, if wind power generation can be predicted, effective use of wind energy can be expected. This study proposes the method for wind speed prediction with neural network, NN, and VAR model using the time-series wind speed data of multiple neighboring sites. Also the study evaluates the wind speed prediction performance with wind data of multiple sites including meteorological towers in Tohoku district.
A gyroscopic power generator is developed which generates 1.8 W by using a rotor of 100 mm diameter spinning at 500 rpm. In the conventional vibration generators which use simple vibration, the power was less than several tens of mW in the wearable size. The gyroscopic generator increases the inertia force by spinning the pendulum in high speed and causing the gyro effect. Also a mathematical model is developed which calculates generated power considering the gyro effect, electromechanical transformation and the gear loss. The calculated powers closely coincided with experimental ones in the range from 1 mW to 1.8 W. Finally the performance of generators which have the same device sizes and rotor spinning speeds as those of 2.5" and 3.5" HDD are calculated. They generate 0.72 W and 1.8 W respectively which show a possibility to be used not only for IoT's but also for motor driving sources used in, for example, wearable air conditioners.