The reliable infrastructure of building automation (BA) systems forms the foundation of smart environments and energy systems in our building towards increasing occupant comfort and safety while reducing the ecological footprint of buildings. This is achieved through the processing of data points collected from sensors and the control of installed actuators, and increasingly incorporates machine learning components. However, engineering of BA systems is intricately linked with the planning, installation, (pre-)commissioning, and operation of building services such as HVAC, and it requires an extensive amount of manual coordination which is often prone to errors, many of which are only detected late in the lifecycle and tends to lose transparency in data provenance. To address this, we propose the application of DevOps, a highly successful paradigm in the field of software engineering, to BA engineering process coordination. In addition, the possibility of using semantic data to develop artifacts such as requirements, construction, and devices of BA systems opens up the avenue of achieving continuous verification of the system as it is built and commissioned. Concretely, we propose a novel approach that integrates a semantic reasoner using the machine-understandable data of the building along with interactions facilitated by Web of Thing Thing Description to the DevOps workflow. The proposed approach is expected to ameliorate limitations of existing workflow management methods and thus provide transparency in the data provenance to gain trust for data-driven AI applications for BA.
Iori Mizutani, Ganesh Ramanathan, Simon Mayer
21 Nov 2021