Since the last two decades, there has been a growing awareness about the climate change and
global warming that has instigated several Directorate initiatives from various administrations.
These initiatives mainly deal with controlling greenhouse gas emissions, use of non-conventional
energy resources and optimization of energy consumption in the existing systems. The European
Union has proposed numerous projects under FP7 framework to achieve the energy savings up
to 20% by the year 2020. Especially, stated by the Energy Efficiency Directive, buildings are
majorly responsible for 40% of energy resources in Europe and 36% of CO2 emission. Hence a
class of projects in the FP7 framework promotes the use of smart technology in the buildings
and the streamline existing rules. Energy IN TIME is one of the projects focused on developing
a Smart Energy Simulation Based Control method which will reduce the energy consumption
in the operational stage of existing non-residential buildings. Essentially, this thesis proposes
several novel solutions to fulfill the project objectives assigned to the University of Lorraine.
The developed solutions under this project should be validated on the demonstration sites
from various European locations. We design a general benchmark building framework to emulate
the behavior of demonstration sites. This benchmark building framework serves as a test
bench for the validation of proposed solutions given in this thesis work. Based on the design
of benchmark building layout, we present an economic control formulation using model predictive
control minimizing the energy consumption. This optimal control has maintenance-aware
control properties. Furthermore, as in buildings, fault occurrences may result in deteriorating
the energy efficiency as well as the thermal comfort for the occupants inside the buildings. To
address this issue, we design a fault diagnosis and fault adaptive control techniques based on
the model predictive control and demonstrate the simulation results on the benchmark building.
Moreover, the application of these proposed solutions may face great challenges in case of
large-scale buildings. Therefore, in the final part of this thesis, we concentrate on the economic
control of large-scale buildings by formulating a novel approach of distributed model predictive
control. This distributed control formulation holds numerous advantages such as fault propagation
mitigation, flexibility in the building maintenance and simplified plug-and-play control
strategies, etc… Finally, a particular attention is paid to the estimation problem under limited
measurements in large-scale buildings. The suggested advanced estimation techniques are based
on the moving horizon methodologies and are demonstrated on the benchmark building systems.