Agent-based Residential Electricity Demand Simulation

Residential electricity consumption constitutes12%of total power consumption inChina. Considering20%of electrical power goes to households sector in developedcountries and stably increasing power demand driven by GDP growth in China there existsenormous potential for energy efficiency and demand side management (DSM) techniques.Furthermore future smart grid with integrated information and communication infrastructureopens new scope for DSM. Thus developing powerful and efficient simulation tools forresidential power demand analysis is vital significant.In this thesis an agent-based residential power demand simulation packageSimulateurCdC by R is created. Common households appliances include television, plaque,dish washer, washing machine, oven, light, public light, refrigerator, freezer, personalcomputer, swim pool equipment, air conditioner and electric heating equipment are modeledbased on EU Energy Label system. Valuable electricity consumer behavioral knowledge isextracted from history power consumption record by data mining technique. The obtainedparameters are smoothly integrated into electric heating models. The package is capable ofgenerating household electricity demand daily schedules for the whole year for any timeresolution in a bottom-up approach. An iterative assessment method is proposed to facilitateimproving simulation accuracy with EvaluationCdC package. Evaluation of the simulationresult show quite similar shape with real power consumption observation curve. Anotheradvantage is the fast simulation speed. For1000houses it takes around130seconds to getfinal simulation results. High simulation efficiency is the main feature of the SimulateurCdCpackage.