Deterministic design optimization for soil nail walls in a Brazilian highway construction

Abstract

Soil nail walls are widely used for slope stabilization in transportation infrastructure; however, their design still relies predominantly on manual, trial-and-error procedures or simplified Limit Equilibrium Method (LEM) analyses, which do not ensure optimality. Existing optimization studies are typically limited to idealized problems, lack integration with industry-standard Finite Element Method (FEM) tools, and rarely incorporate realistic construction stages or current design codes. This study addresses these limitations by proposing a fully automated Deterministic Design Optimization (DDO) framework that integrates the Sequential Least Squares Programming (SLSQP) algorithm with the commercial FEM software Plaxis 2D via Python scripting. The choice of Plaxis 2D enables rigorous simulation of soil–structure interaction, staged construction, and advanced constitutive behavior, while the scripting interface allows seamless automation, iterative model updating, and efficient exploration of the design space. The framework is applied to a real Brazilian highway case study, incorporating the requirements of the national standard ABNT NBR 16920-2. Optimization variables include nail length, face inclination, and total construction cost. Results demonstrate robust convergence and highlight the critical role of cost-based objective functions in achieving economically efficient designs. Validation against LEM analyses performed in Slide2 shows good agreement in factors of safety, with differences observed in predicted failure mechanisms. The proposed approach provides a practical, code-compliant, and scalable framework for automated geotechnical design, with significant potential for cost reduction in large infrastructure projects.