We presented the forecasting method of economic and social indexes in the preceding paper. In this paper, we describe and evaluate the forecasting method of an region's electric power demand in terms of kWh per capita based on the results of the prefecture statistics. It can be seen that kWh per capita has strong correlation with private expenditure for light use and also with the third sector production amount for business use. Thus, the correlation depends on the type of electric services.
In the case of forecasting more detailed area kWh per capita such as within a region in the prefecture, the relation between the region's kWh and the average kWh per capita can be estimated from a multi-correlation equation with two variables. Those variables are the industrial structure designated with employee by industrial sector and the population gravity among the division. This method is especially useful for estimating light use and business use.
At the same time, time-series analysis and exponential smoothing method are used together and discussed here for the case of forecasting more detailed kWh in an area smaller than the region.
The method of forecasting kWh in each mesh and managerial area for distribution system was developed and evaluated.
Then, the maximum peak power, light use power and midnight power in August and December are estimated by summing up each load curve calculated from forecasted kWh in the respective month.