Railway gravity retaining wall design using the flower pollination algorithm

Abstract

An optimized cross-sectional area of a retaining wall assists in reducing material costs. Nevertheless, the design solution obtained from the traditional methods is not always the optimal one. Therefore, this study uses the flower pollination algorithm (FPA) to optimize the configuration of a gravity retaining wall, attempting to find the smallest cross-sectional area that meets design requirements. A key novelty is the use of a metaheuristic for gravity retaining wall design considering railway loading. First, the FPA is described in detail along with the general design methods for retaining walls and the definition of the optimisation problem. Next, a gravity retaining wall case study is performed and the FPA method is compared against alternative approaches, such as genetic algorithm (GA) and particle swarm optimization (PSO). Lastly, a parametric analysis is performed to assess the impact of the design parameters on the optimization outcomes. Results indicate the efficacy of FPA as an optimization method for the design of gravity retaining walls. Moreover, the reliability and speed of the algorithm's convergence are attractive compared to alternative algorithms. The data further suggests that the integration of a landward-leaning wall back with a tilt angle of 14° can expedite the realization of the optimal minimum cross-sectional area. The wall height emerges as the primary influence over the minimum cross-sectional area, which displays marginal sensitivity to ground bearing capacities above 300 kPa. The algorithm can adeptly detect the thresholds in design requirements where retaining walls are impractical (e.g. when ground bearing capacity is a mere 100 kPa and the wall height exceeds 7 m), thereby indicating the need for alternative design exploration.