Integrating occupant behaviour in the simulation of coupled electric and thermal systems in buildings

The presented work depicting the integrated modelling of probabilistic occupant behaviour in buildings consists of non-physical modelling with a physical multidomain impact. The human behaviour considering occupancy, the use of lighting and the use of electric appliances in dwellings has been implemented, but the same method can be used for other stochastic behaviour. The stochastic behaviour is used for simulation of coupled thermal and electrical systems in the building stock and is of high importance for the assessment of smart grids and distributed energy generation. Implementing stochastic occupant behaviour influences the internal heat gains which in turn influence the heat load of the building and the switch-on and -off moment of e.g. an electric heat pump. This, together with the power demand of the used electric appliances and possible on-site generation determine the load on the electric grid and possible instabilities. Here, the use of deterministic profiles for use at both the building and the building district scale no longer fits. Comparison between a determinsitic approach as proposed in ISO 13790 and the use stochastic profiles shows that the direct first order effect is on average rather small: the difference in total internal gains and its influence on the indoor temperature averages nearly zero and the standard deviations σ are small, however high peaks may occur. Also the difference in effect on the electric distribution grid voltage averages nearly zero, however here strong peaks occur which are of most importance for the grid stability. When taking in account the second order effect of heating by means of electricity, much larger differences are noticed: due to longer and more differentiated occupancy times, the average indoor temperature rises. Furthermore, the moment of heating differentiates compared to a determinsitic approach resulting in more but smaller peak demands towards the electricity grid.