Intelligent Monitoring System on Prediction of Building Damage Index using Neural-Network
Abstract: An earthquake
potentially destroys a tall building. The building damage can be indexed by
FEMA into three categories namely immediate occupancy (IO), life safety (LS),
and collapse prevention (CP). To determine the damage index, the building model
has been simulated into structure analysis software. Acceleration data has been
analyzed using non linear method in structure analysis program. The earthquake
load is time history at surface, PGA=0105g. This work proposes an intelligent
monitoring system utilizing artificial neural network to predict the building
damage index. The system also provides an alert system and notification to
inform the status of the damage. Data learning is trained on ANN utilizing feed
forward and back propagation algorithm. The alert system is designed to be able
to activate the alarm sound, view the alert bar or text, and send notification
via email to the security or management. The system is tested using sample data
represented in three conditions involving IO, LS, and CP. The results show that
the proposed intelligent monitoring system could provide prediction of up to
92% rate of accuracy and activate the alert. Implementation of the system in building
monitoring would allow for rapid, intelligent and accurate prediction of the
building damage index due to earthquake.
Keywords: artificial neural
network, building, damage index, intelligent, monitoring
Author: Mardiyono, Reni
Suryanita, Azlan Adnan
Journal Code: jptkomputergg120036
