Experimental and mathematical investigations of soil-conduit interaction within a loaded soil slope
Date of Award
Doctor of Philosophy
School of Engineering
Sanjay Kumar Shukla
Buried conduits are an essential feature of modern underground infrastructure. They are the primary source of utility conveyance around the world because of their economic and safety benefits. The understanding of soil-conduit interaction is vital to ensure the stability of a soilconduit system. Starting in the early 1900s, numerous studies have been conducted to investigate various aspects of the soil-conduit interaction. The researchers have suggested that soil-conduit interaction generally depends on soil type, conduit material, diameter and burial depth of the conduit, applied loading, and soil movement around the buried conduit. However, a vast majority of these studies have analysed the soil-conduit interaction under a level ground surface. However, in reality, conduits not only travel across plain areas but also pass through hilly terrains to reach the end consumers. Keeping this practical fact in view, limited research studies have been carried out in the past, especially investigating the effect of changing landscape on the soil-conduit interaction. However, significant research gaps still remain, which require further detailed study to enhance the understanding of the soil-conduit interaction in sloping terrains.
This research aims to analyse the soil-conduit interaction in a loaded soil slope through experimental and numerical methods. For this purpose, extensive laboratory experimentation, finite element modelling, analytical formulation, and intelligent modelling have been conducted. The experimental study investigated the following aspects: (a) the load-settlement response and bearing capacity of a footing located over a conduit buried within a soil slope, and (b) the stress distribution around a conduit buried within a soil slope. A finite element model was developed to study the structural response of a conduit buried within a soil slope to the applied surface pressure. An attempt was also made to analytically formulate an expression to calculate the vertical load on a conduit buried under a sloping terrain. Finally, this research also focused on building executable finite element modelling-artificial intelligence-based models and converting them into simple mathematical equations for estimating the following: (a) width of Marston’s soil prism for the reinforced concrete and corrugated steel conduits, and (b) settlement of a footing located over a conduit buried within a soil slope.
In the experimental phase, laboratory model tests have been conducted on a strip footing located on top of conduits buried within a soil slope, under static loading condition. The design of the model test cell and the conduit installation technique helped in reducing the scale influence. Using this test setup, two aspects of soil-conduit interaction have been studied, including the effect of the buried conduit on load-settlement response and bearing capacity of the overlying footing, and the stress distribution around the buried conduit due to the applied surface pressure. The experimental results have shown that the conduit buried within a soil slope can have both negative and positive effects on the load-settlement response and bearing capacity of the overlying footing. The negative impact generally happens when the buried conduit intersects with the shear failure planes of the loaded footing and may result in a 252% higher footing settlement and a 40% reduction in its bearing capacity. To avoid any adverse effect on the load-bearing ability of the surface footing, the conduit should be buried at a depth of least three times the footing width. Further, the stress distribution around a conduit buried within a soil slope is highly dependent on the distance of the buried conduit from the yielding soil mass and the resulting soil-conduit interaction. The vertical stress on the crown of a conduit buried within a soil slope can be 84% higher than that of the conduit buried under the level ground.
In the numerical phase, a commercial finite element modelling software has been employed to investigate the structural response of a conduit buried within a sandy soil slope to the applied surface pressure. The developed numerical model has been validated by using results from the aforementioned experimental phase. The comprehensive study included the effect of the proximity of the free slope surface on the conduit deflections, shape deformations, and developed bending moments. The movements of soil particles in relation to the buried conduits have also been discussed in detail. The results show that due to the proximity to the free slope surface, the conduit undergoes unrestricted deformation on the slope side of the conduit, which causes a conduit deflection that is 360% higher than the conduit buried under the level ground. Further, the slope side shoulder of the conduit also experiences high bending moment, which can be 590% higher than the conduit buried under the level ground. To avoid the effect of the slope surface on the conduit, it needs to be buried at a crest distance of more than five times its diameter or at a burial depth that exceeds three times its diameter.
An expression has been developed in the analytical phase for calculating the vertical load on a conduit buried under a sloping ground surface. Using the arching phenomenon, the soil over the buried conduit (inner soil prism) was assumed to yield relative to the adjacent stationary soil mass. Thereafter, the concept of stress transformation has been used to calculate the stresses on a soil element located under a sloping surface, which have been defined along inclined principal planes. The results obtained from the developed analytical expression show that a 10-degree increment in slope angle increases the vertical load on a conduit by about 19.87%. The results of the derived analytical expression have also been compared with numerically obtained results for illustrating the accuracy of the developed expression.
For data analytic methods, extensive finite element modelling (FEM) has been conducted to generate a large-scale dataset. The FEM allows incorporating a wide range of input parameters used to define different types of soils and buried conduits, and their intricate relationships, on the model outputs. The generated FEM-based dataset has been used to build various artificial intelligent (AI) models. The predictive strength of the models has been checked through crossvalidation approach, rigorous statistical testing, and multi-criteria approach. This approach has been utilized for estimating the width of Marston’s soil prism for the reinforced concrete and corrugated steel conduits, and the settlement of a footing located over a conduit buried within a soil slope. For the estimation of the Marston’s soil prism, Artificial Neural Network (ANN) model appeared to be the most accurate for reinforced concrete ( R2 = 0.983, MSE = 0.1268, RMSE = 0.356, and MAE = 0.247), and corrugated steel ( R2 = 0.959, MSE = 0.0166, RMSE = 0.128, and MAE = 0.045) conduits. Whereas, for estimation of the settlement of a footing located over a conduit buried within a soil slope, Multi-layer perceptron (MLP) has been the most accurate with r, RMSE, NSE, SI, and RPD values of 0.974, 0.323, 0.928, 0.44, and 3.75, respectively, and highest ranking (total score = 48). The aforementioned accurate models have been employed to develop simple mathematical equations for the convenient use of practicing engineers.
Access to this thesis is embargoed until 21 October 2022.
Khan, M. U. (2021). Experimental and mathematical investigations of soil-conduit interaction within a loaded soil slope. https://ro.ecu.edu.au/theses/2458