Transient thermal comfort constraints for model predictive heating control
Aachen / E.ON Energy Research Center (2018, 2019) [Book, Dissertation / PhD Thesis]
Page(s): 1 Online-Ressource (xxii, 163 Seiten) : Illustrationen, Diagramme
Changes in the electricity supply system induce the challenge of matching the highly fluctuating and unpredictable renewable energy generation with the yet inflexible electricity demand. This leads to an increasing need for energy storage capacities and demand-flexibility. Space heating of residential buildings is accountable for up to 18% of Germany’s final energy consumption. Thus, electrification of residential heating systems in combination with advanced controls utilizing the structural thermal mass (STM) of buildings as thermal storage could provide some demand-flexibility, at least in the winter season. However, the comfort constraints for resulting temperature drifts have not been conclusively elaborated yet. Existing findings on transient thermal comfort are inconsistent and often based on experimental setups not applicable to conditions resulting from load shifting (LS) in residential buildings. Furthermore, prevailing sophisticated concepts for STM activation have limited applicability to the residential sector, since simple and inexpensive control, easily adaptable to a large diversity of dwellings, is required. In this work the transient thermal comfort during different non-linear temperature drifts within the range of 18-26 °C is evaluated. Realistic thermal conditions as expected to occur in STM activation events are emulated in an advanced climate chamber. 320 subjects following different activity scenarios and having free choice of clothing are surveyed about their perception of these conditions. It is found, that thermal comfort under transient conditions cannot be ensured by defining a universal range of comfort temperatures and a fixed limitation of the rate of temperature change. Moreover, conventional thermal acceptability estimations are unreliable under transient conditions, since effects of adaption and alliesthesia decouple the actual perception of a thermal environment from specific thermal conditions. This work defines resilient dynamic comfort constraints for residential temperature drifts. Temperature fluctuations within the range of 19-25 °C and rates of temperature change between 1 and 3K/h are identified as acceptable, as long as subjects’ current activity level and the direction of the temperature drift are taken into account. The average preferred thermal sensation for all scenarios was slightly warm, resulting in an unsymmetrical comfort range allowing temperature increases of up to 2 K and decreases of 1 K.To evaluate the implications of these findings upon residential STM activation, a model predictive control (MPC) algorithm is developed and used to control dynamic heating operations as a measure of residential LS. The self-learning algorithm does not require extensive measurement data or expert knowledge for parametrization. It optimizes heating operations required for LS according to a dynamic primary energy factor signal, while observing transient thermal comfort constraints. The implemented black-box model predicts thermal conditions within the observed thermal zone with sufficient accuracy to support MPC. Based on that model, the control algorithm is capable to perform beneficial STM activations according to the given primary energy (PE) oriented utility function. For the observed exemplary system, the PE demand can be reduced by 3-7% while maintaining or even improving the thermal comfort. The low computational effort of the developed MPC enables an implementation on simple and affordable hardware, as the Raspberry Pi 3.
Wolisz, Henryk Stefan