Right-sizing Solar PV and Storage for Household Consumer Using Agent Based Modeling

The relatively high upfront cost of solar PV solutions has been the major issue in large scale adaption of solar technology, particularly in the developing world. One of the reasons for higher cost of installation is over provisioning of the solar PV and storage to cater for the variable generation of energy and the propensity of system designers to consider maximal peak load of the household for their calculations. This results in higher capital expense for the customer which in turn limits the population percentage which can participate in solar PV generation. To ameliorate this situation, we propose a solar PV and storage sizing mechanism which builds it's load profile based on an agent model of devices, consumers and events in the house. Instead of focusing on peak load we elicit from consumer, through an online system, their consumption habits and the usage profile for various devices during their daily routine. We construct an agent based model to simulate the usage of devices throughout the day and run Monte Carlo simulations to identify the most likely peak demand at different parts of the day. We integrate this information with solar generation potentials and use hill-climbing algorithm to identify the best solar PV-storage combination which will provide adequate service with the least system size. Our results show that for a typical household we can reduce the capital expenses by 60% without impacting their life style.