Description
Green building (GB) strategies are essential for mitigating energy wastage in thebuilding sector, which accounts for nearly 40% of global energy consumption. However,
due to the unpredictable nature of occupants’ behavior and inadequate energy
management, actual power consumption in GB can exceed intended values by up to 2.5
times. In the realm of curbing energy wastage, the amalgamation of cutting-edge
technologies like Building Information Modeling (BIM), Internet of Things (IoT), and
Artificial Intelligence (AI) into Building Automation Systems (BAS) shows significant
promise. However, data interoperability issues hinder the information exchange
between different digital platforms. Traditional BIM-IoT integration methods using
Dynamo scripts and Arduino IED have limitations in terms of interoperability,
scalability, and accessibility, which hinder information sharing among occupants. This
research aims to construct and validate a robust BIM, IoT, and AI-Enhanced BAS model,
combining these advanced data integration frameworks to facilitate information
exchange and elevate energy efficiency in green building environments.
This thesis starts by a systematic literature review to dissect the intricate
intersections between BAS and green building practices. This study illustrates the
applications of BAS throughout the green building lifecycle, emphasizing its pivotal
role in enhancing indoor human comfort and reducing the energy performance gap in
green buildings. This paper systematically illustrates 1) BAS applications throughout
the lifecycle of GB; 2) BAS applications in supporting GB indoor human comfort,
including thermal comfort, visual comfort, ventilation comfort, and acoustic comfort;
3) a research framework for reducing the energy performance gap in GB; 4) five BAS
and GB integration methods for improving energy efficiency; and 5) limitations and
challenges in the BAS-GB domain. The review reveals that current research in the BASGB domain is insufficient and predominantly concentrates on improving energy
efficiency and occupant comfort. There are four challenges to achieving comprehensive
integration of BAS and GB: uncertainties, long-term prediction and control, BAS-
II
supported sustainability goals, and privacy and security. These findings provide
essential guidance on BAS implementation for GB development. The five BAS-GB
integration approaches lay the groundwork for future research into achieving trade-off
objectives between energy efficiency and occupant comfort.
In addition, this research develops a novel AI-Enhanced BAS model that
incorporates BIM, IoT, cloud computing, and AI technologies to facilitate information
exchange between different digital platforms. This study validates the overall
effectiveness of the integrated BIM, IoT, and AI-enhanced BAS model through real
building case located in Suzhou, China. This research develops a novel IEQ monitoring
platform, named ‘LabMonitor’, which integrates BIM, IoT, and cloud computing
technologies. To verify the scalability and effectiveness of this approach, it was
implemented in a complex engineering laboratory setting that involved rigorous testing
and research activities. A workflow and an inverse Predicted Mean Vote (PMV)
algorithm are developed to achieve a balanced objective between the maintenance of
experimental materials and the thermal comfort of laboratory researchers. Additionally,
the ‘LabMonitor’ includes a cloud-based data acquisition system (DAS) that
comprehensively provides real-time indoor thermal, ventilation, visual, and acoustic comfort information for occupants. The BIM-IoT-cloud computing integration
approach shows superior performance compared to Dynamo-Arduino methods,
demonstrating potential for enhancing experimental efficiency, occupants’ comfort, and
promoting collaboration among researchers in green buildings.
The integration of LSTM and MPC models within the energy optimization
framework presents a significant advancement in enhancing energy efficiency while
maintaining indoor comfort levels. The LSTM model, trained on historical data sourced
from the ‘LabMonitor’ platform, demonstrated robust predictive capabilities for key
parameters like PMV, energy consumption, and indoor occupant behavior. The LSTM
model showcased commendable performance, particularly in accurately predicting
energy consumption levels. Subsequent application of MPC within the system yielded
substantial energy savings of approximately 19.07%, underscoring the practical impact
III
of these advanced modeling techniques in real-world energy management scenarios.
These findings emphasize the potential of leveraging sophisticated predictive and
optimization methodologies to drive sustainable energy practices without
compromising user comfort. The combined use of LSTM and MPC models not only
enhances energy efficiency but also offers a pathway towards more intelligent and
adaptive energy management strategies in BAS.
| Period | Mar 2022 → Dec 2024 |
|---|---|
| Degree of Recognition | International |