A message-driven architecture to integrate isolated building data into digital twins

Abstract

This thesis aims at developing and implementing a framework for the Digital Twin (DT) of buildings that their systems are inside a network-wised isolated environment, hence face data integration challenges. There are several studies focusing on conceptual models and implementation of DT in Architecture, Engineering, Construction, Operation and Facility Management (AECO-FM). DTs can be applied to facility management, preventive maintenance, monitoring and anomaly detection, and enhance the efficiency of constructions projects and workforce safety. However, due to certain features of this industry, this sector has been slow in adoption of DTs. Moreover, despite increasing number of DT implementations, less attention has been given to the challenges of implementing them. Data integration issues including data access, collection, storage and exchange are among the main challenges identified by scholars. Specifically, the majority of DT studies have assumed that data are readily accessible for implementing DTs, while it is rarely the case in practice. This research focuses on the data integration problem in implementing a DT in buildings where accessing sensor data is restricted for various reasons such as data accessibility and privacy issues, or due to deployed legacy BMS/BAS systems. A methodological framework for implementing a DT in such scenarios is presented in which file-based offline sensory data is transformed to message-based online data. In addition, a distributed message-driven architecture is designed to implement this methodological framework. A case study of a community center is implemented to assess the applicability of the proposed methodological framework. This thesis contributes to the field, by relaxing the widespread assumption of easy data accessibility for DT implementation in the literature. The proposed methodological framework and its cloud-native software architecture can be applied to similar scenarios in which sensory data is not readily accessible. The proposed framework is reusable, scalable and flexible in the data types that it can handle. Moreover, Data is critical to collaborations within AECO-FM projects. Therefore, this study is beneficial to practitioners as well allowing them to use the potentials of untapped data which is being constantly collected in data silos of legacy systems of various stakeholders but kept isolated.

Date	6 Dec 2022
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Ali Motamedi (Supervisor)