An IoT Solution for Optimal Energy Storage Operation

Greater availability and affordability of distributed energy generation and storage has made home energy management an increasingly complex task. To get the most value out of the system as a whole, forecasting of generation and demand (and sometimes price) is becoming more important. At the same time, cloud-based computational resources, data sources, and weather forecasts are now easier to access and apply to home energy management. This paper presents a testbed that integrates a number of different components within a single solution towards solving a single goal: minimising the cost of home energy consumption. Machine learningbased forecasts are fed into a dynamic programming approach to optimal energy storage scheduling, leading to increased savings for the customer.