In this seminar, an overview on the current status of data-driven smart building technologies, especially in the areas of Automated Fault Detection and Diagnosis (AFDD) and building-to-grid integration, will be presented. Malfunctioning control, operation, and building equipment, such as unstable control loop, biased sensor, and stuck outdoor air damper, are considered as the top cause for “deficient” building systems, which strongly affect a building’s performance. Meanwhile, extensive research has shown that buildings and building equipment can provide flexible electrical loads to improve grid resilience and overall efficiency and reduce peak demand. Yet significant challenges still exist to allow easy-to-use and cost-efficient AFDD and building-to-grid solutions to be adopted in the field. Data-driven methods, especially those that use machine learning and artificial intelligence strategies, have shown great promises to overcome many of the barricades. This seminar will discuss the needs, gaps, and promising data-driven methods in the areas of building system AFDD and building-to-grid integration. Issues with data-driven methods and future research directions will also be presented.
