Multiobjective optimization‐based design of stabilizing piles in earth slopes

Abstract

Summary Though the technology of using stabilizing piles to prevent landsliding is not new, the design of such piles with a meaningful optimization framework has been rarely reported. In this paper, a multiobjective optimization‐based framework for design of stabilizing piles is presented, in which both reinforcement effectiveness and cost efficiency could be explicitly considered. The design parameters considered in the proposed design framework are the pile parameters, including pile diameter, spacing, length, and position, and the design objectives considered are the reinforcement effectiveness and cost efficiency. The design of stabilizing piles is then implemented as a multiobjective optimization problem. In that the desire to maximize the reinforcement effectiveness and that to maximize the cost efficiency are two conflicting objectives, the output of this multiobjective optimization will be a Pareto front that depicts a trade‐off between these two design objectives. With the obtained Pareto front, an informed decision regarding the design of stabilizing piles is reached. The effectiveness and significance of the proposed multiobjective optimization‐based design framework for stabilizing piles are demonstrated through two illustrative examples: one is the design of stabilizing piles in a one‐layer earth slope and the other the design of stabilizing piles in a two‐layer earth slope. Further, parametric analyses are conducted to investigate the influences of the pile design parameters on the stability of reinforced slopes.

Physics‐informed neural networks for large deflection analysis of slender piles incorporating non‐differentiable soil‐structure interaction

“…Despite its robustness and accuracy, the computational procedure of the LFEM is tedious, given the element mesh must be dense to accurately capture the complex SSI variated along the pile. 7 Consequently, when costly repetitive simulations are required, such as for reliability assessment 9,10 and design optimization, [11][12][13] the computational cost of the LFEM may become prohibitively high. [14][15][16] F I G U R E 1 Large deflection analysis of slender piles considering nonlinear SSI.…”

“…All these studies highlight the potential of PINN in solving pile-related emerging concerns such as damage detection, 32,33 reliability assessment, 9,10 and design optimization. [11][12][13] Even though preliminary investigation of PINN for pile assessment has been conducted, 34 it does not comprehensively capture all the crucial factor for pile response, such as geometric nonlinearity, impeding its application. To fully realize the potential of PINN, the first step is to develop an accurate and reliable PINN framework for the large deflection analysis of piles considering nonlinear SSI.…”

“…utilized the PINN to effectively predict the water flow in unsaturated soils governed by the intricate Richardson‐Richards equation. All these studies highlight the potential of PINN in solving pile‐related emerging concerns such as damage detection, 32,33 reliability assessment, 9,10 and design optimization 11–13 . Even though preliminary investigation of PINN for pile assessment has been conducted, 34 it does not comprehensively capture all the crucial factor for pile response, such as geometric nonlinearity, impeding its application.…”

Efficient reliability‐based design of slope angles in spatially variable soils with field data

“…7,8 Various probabilistic methods which can take the spatial variability of soil properties into account have received increasing attention and been widely applied to the reliability-based design (RBD) in geotechnical engineering during the last decades. 5,[9][10][11][12][13][14] For example, Gong et al 15 presented a fuzzy set-based robust geotechnical design approach for the design of earth slopes wherein the uncertain soil parameters are represented as fuzzy sets. Peng et al 16 performed a reliability-based design of rock slope using Monte Carlo simulation (MCS).…”

“…Various probabilistic methods which can take the spatial variability of soil properties into account have received increasing attention and been widely applied to the reliability‐based design (RBD) in geotechnical engineering during the last decades 5,9–14 . For example, Gong et al 15 .…”