Development of an optimum operation algorithm for smart house with storage battery control based on demonstration tests

Abstract

In recent years, a smart house combining renewable energy systems has been rapidly developed. In our laboratory, we develop an algorithm to decide the operation plan of the storage battery introduced in the smart house. In the previous research, we simulated the operation of storage batteries for representative day aimed at leveling power load. From the simulation results on the representative day of the January, April and July, the power load is leveled in each month. Also, the average value of the power demand after power load leveling in one day is defined as the power load leveling line. However, because the power load leveling line differs from each day, it is necessary to investigate the change of the power load leveling line in several days. Also, in predicting power demand for one house, because how to use electricity is left to customers, predicting power demand is difficult. When planning the operation of storage batteries for several days with multiple houses, because the demand power is smoothed by the smoothing effect, it is expected that the power load leveling line will almost unchanged on each days. However, it has not been investigated how much the power load leveling line will change. Therefore, in this paper, we simulate the operation of storage batteries in one house or 20 houses for 3 days and investigate the change of the power load leveling line. Furthermore, this paper shows how the fluctuation width of the power load leveling line by 20 houses can be suppressed more than in one house. The smart house system consists of a storage battery, bidirectional inverter, photovoltaic power generation, and electric demand load, which connected to the commercial power grid. Operation of storage batteries is planned by Genetic Algorithm (GA) analysis method. The simulation period is from 14th to 16th in Kitami City, Hokkaido, in January (winter), April (middle term), July (summer). The data using for simulation is solar radiation data obtained from NEDO, and power demand data obtained from Hokkaido Electric Power. The simulation result shows that the fluctuation width of the power load leveling line for 3 days in 20 houses is smaller than one house, and further the fluctuation width of January, April and July in 20 houses for 3 days can be reduced by 6.6%, 0.2% and 27.3% than one house. We shows that power management is better to control at multiple houses.