[4M3-GS-13] AI application: Electric power

[4M3-GS-13-02] Regional Power Demand Estimation Based on Residential Location Choice Simulation

Tomofumi Tahara¹, 〇Hideaki Uchida Uchida¹, Hideki Fujii¹, Masahiko Kuroki², Shinobu Yoshimura¹ (1. The University of Tokyo, 2. JERA Co., Inc.)

Keywords:Power Demand Estimation, Multi-agent System, Distributed Power System, Urban Dynamics

In a distributed power system, residents who used to be energy consumers will become energy suppliers, and power can be exchanged complexly in their small region. To build a power system that meets the characteristics of each region, although it is necessary to predict the power consumption in the region, the conventional power prediction models cannot take into account the power demand structure in such a small area. In this paper, we build an urban model that reproduced the dynamics in Yokohama for the past 20 years, and the powe rconsumption of each region was predicted using this model. In the proposed model, we considered the house type, power generation, and electrical facilities that are not handled by conventional urban simulation. As a result, various scenarios could be analyzed for the power consumption associated with urban dynamics.