Short and Long Term Optimal Operation and Robustness Analysis of a Hybrid Ground Coupled Heat Pump System with Model Predictive Control

In the strive against climate change the reduction of anthropogenic greenhouse gas emissions is a key factor. The building sector is the largest consumer of primary energy which comes mainly from burning fossil fuels. The main share of buildings energy demand is represented by space heating and cooling. For this reason, improving buildings heating and cooling systems efficiency is one of the most effective ways towards sustainability. Ground Coupled Heat Pump (GCHP) systems are the most efficient systems for heating and cooling of buildings. However, the high installation cost and the associated long payback period for such systems prevent them from being widely installed. Introducing Hybrid Ground Coupled Heat Pump (HyGCHP) systems in combination with Model Predictive Control (MPC) substantially reduces this drawback. The aim of this doctoral research is to find the optimal control strategy for a considered HyGCHP system with MPC and to guarantee its robustness to uncertainties. For the purpose of this work an integrated dynamic model of the HyGCHP system is composed, based on models of the system components—building, Borehole Heat Exchanger (BHE), primary heating and cooling devices (heat pump and ground coupled passive cooling heat exchanger) and supplementary devices (gas boiler and air coupled active chiller). The system operation is controlled by means of an optimization problem including short term objectives—achieving desired thermal comfort by minimized system operation cost. The optimization problem is solved for a time horizon of one year with imposing cyclic boundary conditions on the ground node temperatures to incorporate long term objectives—thermal balance of the ground for unbalanced heating and cooling loads—integrated with the short term objectives. The optimization problem is also solved for shorter time horizon and without cyclic boundary conditions with two intentions: (1) to see in what conditions the long term optimal operation profile can be reproduced with a more realistic short term control strategy; (2) to analyze whether in the long-term optimal solution the ground is exploited for Seasonal Underground Thermal Energy Storage (SUTES) or as a heat source/sink. Robustness analysis for state estimation uncertainty in the case of a HyGCHP system is performed for the most commonly used short term MPC strategy for heating and cooling systems in buildings (prediction horizon of one day, sampling time of one hour and control horizon of one time step). For that purpose an existing off-line method for robustness analysis is reproduced and clarified, then extended and applied to the HyGCHP system with MPC. This dissertation presents the integrated short-and-long-term optimization approach to analyze HyGCHP system operation, the short term MPC strategy to reproduce the long term optimal system operation profiles, and the extended robustness analysis method applied to the HyGCHP system with state estimation uncertainty in the MPC. The results show that the integrated short-and-long-term optimal operation profile with cyclic ground temperatures compensates the cooling dominated loads by annually mean ground node temperatures higher than the undisturbed ground temperature. The cooling dominated loads are mostly covered by passive cooling up to hitting the upper bound on the outlet BHE fluid temperature. The remaining peak cooling loads are covered by the chiller. On the short term a weekly optimal strategy reproduces the long term optimal annual profiles. For a single BHE this leads to the conclusion that because of high thermal dissipation the ground is not used as a thermal storage medium on the long term but as a heat source/sink on the short term. For the investigated borefield this conclusion changes. For realistically low ground thermal conductivity the optimal system operation includes using the borefield as a seasonal thermal storage medium on the annual term. The results for maximum allowed state estimation uncertainty computed with the robustness analysis method correspond to the performed HyGCHP system simulations with MPC. In conclusion, the method gives a reliable estimation of the maximum allowed state estimation uncertainty for guaranteed robustness. Improving HyGCHP systems with MPC can continue in two directions. First, towards investigating the SUTES ability of systems with other borefield drilling configurations. More compact borefields of more BHEs are expected to more efficiently operate as seasonal thermal storage mediums. This would enable SUTES in cases of higher ground conductivities, which is an opportunity of wider applications of HyGCHP systems. Second, the efficient implementation of optimal SUTES in short-term MPC strategies is an important research direction. The improved exposition of the robustness analysis method and the developed subsidiary method for application of the robustness analysis to the investigated HyGCHP system with MPC represent a fast offline computation to check and guarantee system robustness to state estimation uncertainty. The developed framework enables determining key boundary conditions for further system design and control: temperature measurement accuracy, model accuracy, MPC prediction horizon length. HyGCHP systems of the type presented in this dissertation which are characterized by bounded uncertainties can be controlled by conventional MPC provided that the level of the incorporated uncertainties is not higher than the guaranteed level computed using the presented method.nrpages: 156 + xxxviiistatus: publishe