Abstract

Radiant slab heating systems are receiving considerable attention due to the multiple advantages they offer such as improved thermal comfort in buildings and suitability for other related applications in cold climates, particularly snow melting and de-icing of pavements and infrastructure. Hydronic radiant heating can utilize low temperature renewable energy heat sources. The operation of these systems can be optimized by applying predictive control and further the energy costs can be reduced by optimizing their interaction with smart grids by utilizing the flexibility in their demand profiles. However, compared to conventional air heating systems, radiant systems have several added complexities such as the slow transient heat conduction within the slab. Efficient design and operation of radiant slabs require several critical decisions on design and control variables to maintain comfortable thermal conditions in the space and achieve slab surface temperatures within the recommended range depending on the application.
This thesis presents a methodology for modelling of hydronic radiant slab heating systems with significant thermal mass (a concrete layer with embedded tubes) for predictive control to utilize the energy flexibility of the building and/or infrastructure in interaction with smart grids and dynamic pricing of electricity. The modelling approaches include low-order grey box models as well as frequency domain techniques. Each approach has its own specific advantages and unique information can be obtained from each one that complement each other when optimizing system operation and designing the control strategies.
Using the developed frequency domain model of the zone, key transfer functions are calculated for a case study. By means of transfer functions, the effect of different levels of thermal mass on the zone thermal response and quantification of the energy flexibility in response to grid signals is studied. The model is used to evaluate different design and operation options on a relative basis. It is shown how transfer function analysis provides insight into the building thermal dynamics without the need for simulations. A new transfer function that relates the radiant floor heat source at the bottom of the slab to the zone air temperature is introduced and its derivation is demonstrated. By means of the transfer function, the delay between the heat input of the radiant slab and zone air temperature is determined. Experiments with a full scale test room in an environmental chamber are used to validate the key design parameters obtained from the frequency domain transfer function regarding the operational strategies for energy flexibility of the thermal zone. Frequency domain techniques may also be utilized to establish the appropriate order for low-order RC models for different applications.
A low-order thermal network RC model for a case study, validated with experimental measurements, is utilized to study several predictive control strategies in response to changes in the grid price signal, including short term, more reactive changes of the order of 10-15 minutes notice. An index is utilized to quantify the energy flexibility with the focus on the peak demand reduction for specific periods of time when the electricity prices are higher than usual. It is shown that the developed predictive control strategies can aid greatly in minimizing the electricity cost of the building and up to 100% reduction in peak power demand and energy consumption is achieved during the high price periods.
Low-order thermal models are also utilized to study radiant slab heating systems of infrastructure that have much higher thermal mass than the radiant slabs in buildings due to structural reasons. Hydronic heating of roads and pavements surfaces to avoid ice formation has several advantages compared to traditional surface salting and other anti-skid methods which have a lot of limitations. However, the traditional method of designing such systems does not consider the thermal mass of the system and its potential for predictive control, energy flexibility and optimized performance. Finally, predictive control strategies are presented and studied to take advantage of the hydronic slab thermal mass and minimize the energy consumption and peak power demand and are validated in a full scale experiment in an environmental chamber.