Modeling the ultimate confined compressive strength and increase in strength of carbon-reinforced concrete column
Retrofitting concrete with carbon fiber reinforced polymer (CFRP) has been proven to be a method of increasing the ultimate confined compressive strength of concrete columns. The present study uses a hybrid of analytic hierarchy process (AHP) and artificial neural networks (ANN) that assesses the st...
Saved in:
Main Authors: | , , |
---|---|
Format: | text |
Language: | English |
Published: |
Animo Repository
2018
|
Subjects: | |
Online Access: | https://animorepository.dlsu.edu.ph/etd_bachelors/7080 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | De La Salle University |
Language: | English |
Summary: | Retrofitting concrete with carbon fiber reinforced polymer (CFRP) has been proven to be a method of increasing the ultimate confined compressive strength of concrete columns. The present study uses a hybrid of analytic hierarchy process (AHP) and artificial neural networks (ANN) that assesses the strength performance of circular, square and rectangular columns with CFRP confinement, steel reinforcement and a combination of both. Data on concrete columns were made by reviewing existing related studies after which a total of 935 data were collected for study.
The process of AHP was first used to determine the best set of parameters from the database. The researchers observed that for all the columns, the parameters with the highest weights/impacts were as follows: unconfined concrete strength (f'co), ultimate jacket strength (fCFRP), volumetric ration of CFRP (CFRP) and steel transversal strength (fs). Additional parameters such as the diameter (D) was found to have the highest impact for circular columns, and it was the corner radius for the square and rectangular columns.
Meanwhile, the MATLAB R2014a software was used to administer the self-organizing map (SOM) and back propagation (BP) ANN in the study. Using the parameters obtained from AHP, SOM was adopted to classify the data with similar characteristics by analyzing the input planes produced. In analyzing the SOM models, it was found that the behavior and classification of data in the first cluster of the circular columns was similar to the third cluster of the square and rectangular columns.
The feed-forward back propagation (BP) ANN was then utilized to attain the ultimate confined compressive strength (f'cc) and increase in compressive strength (f'cc/f'co) of the concrete columns. Four BP models were considered with varying hidden nodes with the highest regression R-value closest to one (1) were considered as the best performing models. For additional analyses on the BP models, a parametric study as well as comparison with existing models from previous studies were carried out. It was found that a significantly better model was achieved for the present study based on the obtained R-values. The actual and predicted confined compressive strength (f'cc) values of the generated model produce a higher correlation when compared to the previous models considered in the study. |
---|