Can AI Predict How a Model Building Shakes?
Project Description
Can artificial intelligence predict how a building will move during an earthquake? Students will help test this idea using small model buildings on a shake table—a platform that can reproduce earthquake motion in the laboratory. They will help assemble the models, attach small sensors, label cables and prepare test records. During each test, students will observe how the model moves and record any visible changes. They will then use supplied computer tools to turn the sensor readings into graphs and compare the measured motion with predictions from an existing AI model. Students do not need to build or train the AI system, and no previous experience in structural dynamics or programming is required. The project offers an accessible introduction to model building, earthquakes, sensors, data and AI through guided practical work.
Supervisor
PAN, Yuxin
Quota
2
Course type
UROP1100
UROP2100
UROP3100
UROP3200
Applicant's Roles
• Help measure, assemble and photograph small building models.
• Attach and label sensors and cables using a prepared layout.
• Complete simple checklists before each shake-table test.
• Observe the tests and record how the model moves or changes.
• Use supplied tools to create graphs from the sensor readings.
• Compare the experiment with the AI prediction and discuss where they agree or differ.
Applicant's Learning Objectives
• Discover how engineers use models and shake tables to study earthquakes.
• Gain hands-on experience with sensors and experimental data.
• Learn how graphs can describe the movement of a building.
• Understand why AI predictions must be checked against real experiments.
Complexity of the project
Moderate