Bio-inspired burrowing mechanics

Cavity expansion theory, X-ray CT scanning, FEM, SPH, machine learning

Through two current projects on self-anchored burrowing robots, we are learning more about how animals adapt their burrowing strategies to the terrain. We seek to calculate the forces and moments engaged in the limbs of burrowing animals to design self-propelled probes that can excavate with minimum soil disturbance, to allow data collection on soil properties by embarked sensors. Machine learning algorithms are being developed to interpret the data collected and to optimize the design of underground structures.

Simulation of penetration and burrowing processes (FEM and SPH)

We are collaborating with Haley & Aldrich to understand why mole rats can move forward and backward with ease in the tunnels that they burrow. How can mole rats anchor themselves sufficiently in the soil that they previously excavated to move uphill and downhill? Answering this question will provide clues to develop a bidirectional propeller for self-anchored probe retrieval. The bidirectional propeller will facilitate collection of residual strength data, which will help assess the risks of ground loss or borehole instability in Horizontal Directional Drilling (HDD). We are developing a modeling approach that can be used to simulate anchor penetration and pull out at reasonable computational cost. The Smooth Particle Hydrodynamics (SPH) method is used to simulate the soil close to the penetration zone where large deformation of soil takes place, while the portions of the model where only small strain is expected (i.e., probe and far-field soil) are simulated by the Finite Element Method (FEM). FEM+SPH cone penetration simulations were benchmarked successfully against FEM results, which proved the concept of the proposed numerical approach.

Pictures: He and Arson, 2020

Influence of plant root growth on soil microstructure and strength

We study of the influence of plant root growth on the density of a granular soil. Our collaborators at L3SR established a protocol to obtain reproducible samples of sand with a sprouting seed. Maize and chickpea roots are grown in tubes filled by pluviation with Hostun sand. Root development and induced soil micro-structural changes are imaged every 24 hours by x-ray tomography for 7 days. The acquired images are processed and trinarized into separate phases (soil, root, voids). This allows measuring the evolution of soil porosity during root growth to be measured. Eventually, a sub-volume of soil around a root tip is defined after 7 days of growth and the evolution of porosity in that volume is characterized in reverse chronological order. The root does not affect the soil local porosity nor the topology of the root system at a distance larger than ten times the diameter of the main root. The evolution of local porosity is not fully understood at this stage, but it is reproducible.

Pictures: Anselmucci, Ando, Sibille, Lenoir, Viggiani, Peyroux, Arson, Bengough, 2018

Root-like cavity expansion at shallow depth

In plant roots, water uptake increases the turgor pressure inside the root cells, which pushes cell membranes against stiff cell walls. Roots thus expand radially, leading to a decrease in soil compressive stress ahead of the tip, allowing for root axial growth. Plant roots can be seen as pressurized vessels that adapt their shape to soil mechanical response, thus probing soil properties. In general, the deformation patterns and failure mechanisms of pressurised cavities at shallow depth are of relevance to many geotechnical applications, including tunneling and horizontal directional drilling. That is why we conducted an experimental study of a reduced-scale pressurised cavity under geostatic stress, in order to measure the effect of cavity length, vertical stress and soil density on soil deformation and failure. x-ray computed tomography was used to acquire images of the system at key stages of the cavity inflation process. A closed shaped failure region developed around the cavities, beyond which, shear planes of elliptic paraboloid shape formed, extending from the bottom of the cavities all the way to the free surface. The plane strain assumption did not hold beyond the central portion of the longest cavity tested (L = 6D). The volumetric strain and porosity changes inside the shear bands showed significant dilation in dense specimens, but contraction in loose specimens. The average orientation and the thickness of the shear bands were in agreement with those found in the literature for passive arching mechanisms (anchoring). The orientation of the principal strains around the cavity follows catenary shapes, similar to those displayed in active trapdoor mechanisms.

Pictures: Patino-Ramirez, Arson, Caicedo, Viggiani, Sibille, Ando, Lenoir, 2018

Machine learning - based estimation to blowout susceptibility around pressurized cavities

The need to solve complex problems that still do not have a closed form analytical solution is one of the challenges of current practice, including the susceptibility estimation of cavity blowout during horizontal directional drilling (HDD). Computational tools and high frequency data acquisition in conjunction with machine learning open up new opportunities to create tools that can aid in the understanding and design of these problems. We propose the use of a predictor function, calibrated with a pool of numerical simulations that can predict the susceptibility of blowout of a cavity, given a fixed geometrical and stress configuration based on the mechanical parameters of the soil the cavity is embedded in; results using k-means clustering (for classification) and support vector machines (SVM) to create the predictor function, showed and accuracy of about 87% in predicting the blowout susceptibility.

Pictures: Patino-Ramirez and Arson, 2020

Design of Horizontal Directional Drilling alignment by Ant Colony Optimization (ACO)

Horizontal Directional Drilling (HDD) is a trenchless method that consists in drilling an inclined and curved bore from an entry point to an exit point. In practice, HDD is designed iteratively by trial and error, to minimize the cost under geometric and mechanical constraints. Here, we optimize the drill path with continuous implementations of an Ant Colony Optimization (ACO) algorithm that sets the depth of the alignment and its entry and exit angles as the design parameters to optimize, to ensure minimal drill path length (cost), avoid collapse or instability (mechanical constraints) and remain in the construction domain (geometric constraint). We compare the ACO results to the drill paths designed in practice in two different scenarios: one in which the entry and exit points are fixed, and one in which the geometry of the central segment is constrained. Results show that ACO can be used to automate the otherwise time-consuming design process while minimizing the drill path length and the costs associated to it.

Pictures: Patino-Ramirez, Layhee, Arson, 2020