Harmony Search Approach In The Strut And Tie Model To Optimise The Stress Distribution In A Concrete Box Girder

• Main Author: Lim, Alice Pei San
• Format: Final Year Project / Dissertation / Thesis
• Published: 2021
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utar.edu.my/4581/1/Alice_Lim_Pei_San.pdf; http://eprints.utar.edu.my/4581/
• Institution: Universiti Tunku Abdul Rahman

Stress evaluation for box girder structure has always been the popular research topic to understand the behaviour of box girder to ensur e capacity provided and good serviceability performance. methods introduced by rese archers for sufficient ductility Out of the many box girder’s behavioural study , s trut and tie model (STM) can effectively demonstrate the stress distribution using truss analogy. Good STM construction is important to avoid oversimplified and dense STM that may cause excessive crack width an construction using traditional trialandd structure failure. Good STM error method can be tedious and timeconsuming especially when dealing with complex structure. Optimisation using metaheuristic algorithm could provide an alternative for more efficien t STM construction. This study aims to develop a stress optimisation model using harmony search (HS) algorithm to control and limit cracks in the concrete. Firstly, stresses affecting parameter and critical area at the inner face of the box girder were ide ntified using stress analysis. Secondly, HS optimisation model was developed by constructing the objective function and optimisation procedure. Thirdly, optimisation model validation and efficiency evaluation were performed. Lastly, stress distribution at the inner face of the box girder was optimised. The critical area was located, and stresses recorded ranges from 25.79 MPa to 37.63 MPa. Optimal solution was found to converge results demonstrated a better stress distri at 269 th iteration. The optimisation bution with reduced total element stress from 9.21 MPa to 8.63 MPa (6.3%) and amount of reinforcement needed reduced by 6.26%. The optimisation results agreed well with the from FEM with percentage error of 1.2% approximated solution . Thus, the optimisatio n model was efficient in generating results with sufficient accuracy with lesser computational time and simpler computational procedure.