In the course of the development of practice, the terms of Fault Detection and Diagnosis (FDD) are increasingly used to isolate and diagnose building automation problems. Operators use data and warnings to identify problems that affect building performance. Many solutions focus on identifying problems that make it easier to maintain occupant comfort, reduce energy costs, and improve maintenance outcomes.

Commercial HVAC industry has been using fault detection and diagnostics for decades. Many FDD algorithms have been developed and deployed to detect and diagnose faults using sensors commonly installed for control purposes. These algorithms intend to provide operating personnel with actionable information. However, traditional rule-based FDD algorithms have limitations that their effectiveness highly relies on building experts. Each building has its uniqueness. It is impractical to have experts implementing proper FDD algorithms in every building. Therefore, we believe that it is good timing to utilize artificial intelligence (AI) with conventional FDD algorithms. We also believe that best rules stem from a combination of human-machine interaction, machine learning, expert domain knowledge, and AI.

AI is not magic, to make it work, sufficient data in good quality are necessary. Domain knowledge is also the most critical to link AI’s findings with the physical world. We might still be years away from a fully smart AI which can interpret all data collected and help human-beings better detect and diagnose faults and prevent disaster. But now, AI-driven BAS/IoT systems that use deep learning and reasoning can already provide 24/7 energy management and optimization based on data, including measures of occupancy, indoor and outdoor temperatures, humidity, air quality, and more.

We hope soon that AI-driven FDD for building management systems should take place where faults in equipment and operations are detected and then isolated to the actual fault cause for its human counterparts to act on (or, it is even better the AI can fix faulty systems on its own!).

At Entellio, we are interested in helping our clients collect data from the building and to identify faults to prevent downtime and maintain system integrity. As part of the services offered, our extensive diagnostic library can be adapted to a specific system for specific sequences in your demand.