Month: July 2021

In order to make informed decision about structural integrity or building operations, it is important to search, organize, and visualize vast amounts of vibration data effectively. We are using the highly-instrumented Goodwin Hall Smart Building as a platform for real-time data visualization to address some of the fundamental and practical challenges in this field.

Collaborators: Zhiwu Xie (Virginia Tech)

We are designing a framework for structural health monitoring that functions fully online, without any prior information about the structure. As part of this effort, we are exploring several key principles that could help to accomplish this goal:

1. Machine learning strategies that can handle high dimensional data, in order to avoid information loss due to dimensionality reduction.
2. The use of Generative Adversarial Networks to overcome data scarcity and data imbalance.
3. Alarm threshold tuning based on reliability analysis to reduce sensitivity to user-defined parameters.

Related Publications