Strengthening Beams using FRCM Machine Learning Approach and Numerical Models - Master of Applied Science Thesis

Aug 4, 2026

Upload by Eden

Fiber Reinforced Cement Matrix (FRCM) strengthening offers improved performance compared with Fiber Reinforced Polymers (FRP), yet accurate prediction of failure and post-peak behavior remains difficult. This thesis develops a machine learning framework to evaluate the capacity of FRCM-strengthened beams using beam mechanical and geometric properties, steel reinforcement details, and FRCM layer characteristics including number, type, mortar compressive strength, and thickness. Three separate ML models are trained for pre-peak, failure, and post-peak stages, then validated and reinforced with verified finite element models for robustness on unseen data.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages105

Reading Time4 h 25 min

AccessFree to download

Abstract

Acknowledgments

Chapter 1: Introduction

General

Concrete structures and need for rehabilitation

Research objectives

Comprehensive Review

Numerical Analysis of Shear and Flexural Behavior

Development of a Machine Learning Model

Thesis outline

Chapter 2: A comprehensive review of using FRCM for strengthening concrete structure members

Introduction

Background

Flexural Strengthening with FRCM

Shear Strengthening with FRCM

Modes of Failure

Numerical Study

Detailed Approach

Simplified Approach

Artificial Intelligence (AI) Approach

A novel machine learning approach for predicting load-deflection curves of FRCM-strengthened beams

Introduction

Background

Experimental Studies

Numerical Studies

Strengthening Beams using FRCM Machine Learning Approach and Numerical Models by Kambiz Daneshvar A thesis submitted to the Faculty of Graduate and Postdoctoral Affairs in partial fulfillment of the requirements for the degree of Master of Applied Science in Civil Engineering Carleton University Ottawa, Ontario Copyright © 2023, Kambiz Daneshvar Abstract The use of Fiber Reinforced Cement Matrix (FRCM) over traditional Fiber Reinforced Polymers (FRP) systems has gained attention due to FRCM's superior performance in various conditions. Despite the existing studies, predicting failure and post-peak behavior of FRCM has remained a challenge. This research leverages Machine Learning (ML) techniques to evaluate the capacity of FRCM-strengthened beams. This innovative approach addresses the limitations associated with both experimental and numerical models by providing a comprehensive model capable of predicting the response of FRCM-strengthened beams. The model considers various factors, including mechanical and geometric properties of the beams, steel reinforcement, and characteristics of the FRCM layers such as number, type, compressive strength of mortar, and thickness. A novel approach to employ ML is adopted to utilize three separate ML models to assess the beam capacity in pre-peak, failure, and post-peak stages. Following the ML model validation, verified Finite Element (FE) models are used to ensure the proposed model's robustness in handling unseen data. Keywords: FRCM, Machine learning, Reinforced concrete beams; Numerical modeling Acknowledgments To begin, I am deeply thankful for the unwavering support and guidance provided by my supervisor, Professor Hamzeh Hajiloo. His encouragement and expert advice have been fundamental to my research and throughout my Master of Applied Science program at Carleton University. I am thankful for the trust placed in me, and I hope that this trust, established from the first day, has only grown stronger. I also extend my sincere thanks to my friend, Mohamad Javad Moradi, for his constant encouragement and support. Lastly, my heartfelt appreciation goes to my parents whose endless encouragement and support have been invaluable. Table of Contents: Abstract .......................................................................................................................................................... I Acknowledgments......................................................................................................................................... II 1 Chapter 1: Introduction......................................................................................................................... 1 1.1 General ......................................................................................................................................................1 1.2 Concrete structures and need for rehabilitation....................................................................................2 1.3 Research objectives...................................................................................................................................3 1.3.1 Comprehensive Review ........................................................................................................... 3 1.3.2 Numerical Analysis of Shear and Flexural Behavior: ............................................................... 4 1.3.3 Development of a Machine Learning Model:.......................................................................... 4 1.4 Thesis outline ............................................................................................................................................5 2 Chapter 2: A comprehensive review of using FRCM for strengthening concrete structure members. 6 2.1 Introduction ...............................................................................................................................

FAQs

Why is FRCM considered over traditional FRP for strengthening beams?

FRCM has gained attention because it shows superior performance under various conditions compared with traditional Fiber Reinforced Polymers (FRP) systems.

What makes predicting FRCM beam failure and post-peak response challenging?

Despite existing studies, predicting failure and post-peak behavior of FRCM-strengthened beams remains difficult for both experimental and numerical approaches.

How does the thesis design the machine learning model for beam capacity prediction?

It uses three separate machine learning models to assess beam capacity in pre-peak, failure, and post-peak stages, incorporating beam properties, steel reinforcement, and FRCM layer characteristics; then verified finite element models support robustness on unseen data.

Aug 4, 2026

Upload by Eden