A Simulation Optimization Decision Framework for Composite Tunnel Support Design Using FEM, Particle Swarm Optimization, and Multi-Criteria Decision-Making
Abstract
Tunnel support design in rock masses requires controlling deformation while maintaining structural safety, constructability, and material efficiency under variable geotechnical conditions. This study presents a Simulation–Optimization–Decision Framework (SODF) for preliminary and comparative tunnel support design. The framework integrates FEM-based numerical modeling in PLAXIS 2D, Particle Swarm Optimization (PSO), and a Multi-Criteria Decision-Making (MCDM) model to evaluate composite support systems composed of shotcrete and steel sets. Deformation at the tunnel crown was used as the primary optimization response, while structural feasibility was verified through axial force–bending moment and axial force–shear force interaction envelopes with a minimum safety factor of FS≥1.5. Two case studies were analyzed using GSI values of 40, 45, 50, and 55. Case 1 considered Erm=500,000 kN/m2, with an additional condition of 250,000 kN/m2 at GSI=40, whereas Case 2 considered Erm=545,000 kN/m2, with an additional condition of 300,000 kN/m2 at GSI=45. Three PSO configurations were evaluated to compare the search intensity, the diversity of feasible solutions, and the computational effort. The results showed that feasible configurations were mainly concentrated between 10 and 14 cm of shotcrete thickness in Case 1 and between 10 and 12 cm in Case 2, with frequent selection of 254 mm and 305 mm steel sets. The MCDM analysis indicated that the lowest-deformation alternative does not necessarily coincide with the most suitable engineering alternative when material cost, shotcrete applicability, steel-set maneuverability, and local availability are considered.