Forecasting of Residential Hot Water Demand Profiles and Optimization by Dynamic Programming and A* algorithm
Masters in Engineering Student
Electricity consumption in South Africa is currently at a struggle with meeting the demand and as a result has caused numerous 'blackouts'. Demand Side Management (DSM) is the process of shifting electrical consumption of the grid during peak hours to off-peak hours of the day. Electric Water Heaters (EWH's) are the single biggest consumer of electricity in households and tends to have an increased interest for DSM due to flexibility and intelligent control.
The first goal of this research is to provide an improved method of generating synthetic hot water demand profiles. This will be achieved by using clustering techniques on measured water meter data to characterize different types of water usage events and statistically generating a water profile which accurately predicts the hot water consumption of a household.
The second goal is investigating different methods for optimizing the electrical energy of an EWH using the predicted schedule. Two promising optimization techniques were selected and are implemented in a Python environment to show how much electricity can be saved. The techniques chosen are Dynamic Programming and A* Search Algorithm Optimization.