Heating demand response in detached houses - comparing cost savings potential for consumers and the whole energy system

This thesis assessed the cost savings potential of load shifting demand response (DR) in the Finnish detached house building stock with direct electric heating. To model the interaction between buildings and the energy system, an archetype building model capable of simulating heating demand was integrated with a North European energy system model. Interior air and domestic hot water temperatures were allowed to vary within preset bounds so the buildings could shift electricity use towards economically beneficial hours. The annual operational costs of the integrated model were then minimised for two weather-years, the relatively cold 2012 and warm 2015, with the GAMS-based Backbone optimisation framework. Flexible buildings were modelled either with a grid-interactive or self-optimising consumption profile: the former interacted with the energy system to minimise total system costs, while the latter minimised only local heating/cooling costs based on fixed hourly electricity prices.

Grid-interactive DR provided system-level savings up to €50 million per year in total or €180 per flexible building. In contrast, large self-optimising DR penetration increased system costs and caused capacity shortages. The economic impacts were highly dependent on weather year: while 2015 exhibited frequent flexibility events and system benefits were gained with low penetration, in 2012 flexibility events were more infrequent and benefits unclear with penetration below 90%. In both years, most of the savings came from reduced natural gas use in energy production.

Both grid-interactive and self-optimising DR lowered heating and cooling costs relative to non-flexible buildings. As grid-interactive DR penetration increased, non-flexible buildings gained cost savings. Simultaneously, flexible buildings lost some of their relative benefit. For both years, increasing grid-interactive DR reduced the cost gap between flexible and non-flexible buildings. The economic conflicts of interest between self-optimising and grid-interactive buildings were very small. Further research is needed to assess how the electricity markets could better respond to self-optimising DR and prevent excess load shifting to inexpensive hours.