REHVA Journal published and article about a Reinforcement Learning control framework that helps to optimise the control of the building energy system, taking into consideration occupants´ behaviour and PV power production. Both occupants´ behaviour and renewable energy production are very hard to predict: the first because occupants´ behaviour can be very unique, and the second because the production of renewable energy can be volatile.

The research explored in the article aims at developing a self-learning control framework that considers the hot water use behaviour of occupants, and variating solar power production, and learns how to optimise the operation of the system to minimize energy usage.

Read the article here.