Design Optimization of Soil-Metal Composite Arch Bridges: Recent Swarm Intelligence Applications

Ali Derakhshani | Seyyed MohammadSina Fakhermand

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=169811
Date :  2022/09/20
Publish in :    Iranian Journal of Science and Technology - Transactions of Civil Engineering
DOI :  https://doi.org/10.1007/s40996-022-00956-0
Link :  http://dx.doi.org/10.1007/s40996-022-00956-0
Keywords :Soil-metal structure, Composite arch bridge, Optimization, Swarm intelligence

Abstract :

---

More economical construction of composite soil-metal structure, as an appropriate alternative to traditional concrete and steel bridges, is of utmost importance. This study aims at the optimization of the soil-steel composite bridge design based on Canadian highway bridge design code. The objective function accounts for the total costs of preparation and installation of the steel panels and, soil backfilling and compaction. The developed simulation–optimization model is created by connecting the structural design programs to optimization algorithms. The numerical simulation is conducted on three types of soil-steel bridges involving Part Arc, Low Profile Arc and High Profile Arc with four different span sizes for each bridge type. The construction cost is minimized using swarm intelligence techniques including the classical particle swarm optimization (PSO) algorithm as well as two novel approaches of whale optimization algorithm (WOA) and grasshopper optimization algorithm (GOA). According to the optimal results obtained by the developed model, the cost of construction significantly decreases compared to that of the design by the code of practice. In addition, the statistical description of the results of three algorithms reveals that in most cases, the GOA shows the best performance, followed by WOA and PSO. Moreover, the effects of changes in the costs of construction materials on the objective function are evaluated. It is inferred that the unit cost of steel is more effective on the cost function in contrast to that of the soil