MACHINE LEARNING BASED ACTIVE CONTROL SYSTEM OF STRUCTURE EXPERIENCING DYNAMIC LOADS
The design of structures capable to resist dynamic load is becoming one of the challenges in civil engineering. Implementing active control system is an innovative solution to the problem. Active control strategy can be in form of an active mass damper (AMD) which uses actuator to control its moveme...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/66821 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The design of structures capable to resist dynamic load is becoming one of the challenges in civil engineering. Implementing active control system is an innovative solution to the problem. Active control strategy can be in form of an active mass damper (AMD) which uses actuator to control its movement. This paper presented an analysis of a 3-storey experiment model. This structure is modelled numerically using finite element method and then, loaded it with some known earthquake excitations. Structure response is calculated using numerical method for a given load. The modelling is improved by implementing machine learning called artificial neural networks to active control systems using controlled responses of the structure with classical control algorithm (LQR) from random ground excitation as its training data. The result of structural response with and without control is analysed and compared. This paper will be a foundation to experimental research for writer’s master degree thesis.
|
|---|