Quality requirements for forecasts in HVAC operation optimization
Keywords:building model,, machine learning,, thermal comfort, energy saving
Optimizing HVAC operation by taking into account predictions for presence, occupancy and inner loads, weather (mainly air temperature and solar irradiation) and thermal behaviour of the room or building can lead to significant energy savings while maintaining thermal comfort for the occupants. However, the quality of forecasts plays an important role for the success: High prediction qualities are essential for achieving the objectives in energy saving and thermal comfort. In the present paper, a simulation study is presented for the example of an office room with up to three occupants. Perfect and real (non-perfect) forecasts are applied for simulating predictive HVAC control in the course of one year. For evaluating the impact of forecast quality, the annual reduction of cooling energy demand and the decrease of thermal comfort are considered. Results show that there is a complex interaction between the different forecasts: The combined quality of all forecasts determines the benefit which can be reached from predictive control. If forecasts are not good enough, thermal comfort decreases significantly compared to perfect forecasts or the reference case without predictive control. Here, especially the forecast of room temperature development (thermal behaviour of the room) was found to be very important. If the forecasts are good, the annual cooling energy demand can be decreased by 19 % in the example while maintaining high thermal comfort.
How to Cite
This work is licensed under a Creative Commons Attribution 4.0 International License.