Abstract
The purpose of this research was improving the accuracy of the bottom-up residential energy end-use simulation model and estimating actual occupants’ behavior by optimizing the model using smart meter data. Using the simplex method, we optimized the input parameters of occupants’ behavior and the electricity consumption of appliances, and improved the accuracy of electricity consumption estimation for each household type divided by number of people and housing type. By optimizing, the error in daily electric consumption between the estimated value of the bottom-up model and the smart meter data could be improved to 9% or less. From the modified occupants’ behavior parameters, it was estimated that the difference between the bedtime data of the National time use survey, which was input data of the model, and the bedtime in reality is about 1 hour in the 5 person in apartment houses, whereas it was the 3hour difference in the 1-person families. It was presumed that there was a difference between the input data and the actual bedtime due to the attributes of the household members. From the analysis using the TV audience rating, it was revealed that the decrease in the activity rate at midnight can be accurately reproduced.
