Title:
Development of Fault-Detection ANNs for Structural Damage Prediction
Author(s):
Mohammad H. AlHamaydeh, Ahmed F. Mohamed, and Mahmoud I. Awad
Publication:
Symposium Paper
Volume:
350
Issue:
Appears on pages(s):
45-53
Keywords:
structural damage, profile monitoring, fault detection, artificial neural network
DOI:
10.14359/51734311
Date:
11/1/2021
Abstract:
Structural systems are critical components of modern societies, and as such, need constant condition assessments. Incurred damage and its accumulation due to various loading conditions may lead to complete structural failures with safety consequences. Thus, damage assessment and early fault detection are essential to decision-makers for strategic planning, performing resource-allocation, and obtaining the logistics of these systems. In this article, a data-driven methodology using fault detection of structural systems is proposed. This method utilizes Artificial Neural Networks (ANNs) to model damage due to earthquake loading using sophisticated nonlinear analyses of structures. As a proof-of-concept, the ANN approach for damage-detection is applied to a typical four-story reinforced concrete (RC) structure having varied concrete strengths. The approach is found to have a high potential for successful anomaly identification in RC structural systems.