Use of thermo-active piling for slope stabilisation

  • Davide Vitali

Student thesis: Doctoral ThesisDoctor of Philosophy


Accidents caused by formation of ice on the asphalt layer of roads during the winter season and the failure of the earth structures on which roads are built often make the headlines due to the significant impact they can have on the transport of people and goods. A novel method, namely thermo-active pile row, was conceived to simultaneously increase the engineering performance of road embankments against slope failure and reduce the formation of ice.

In this thesis, the geotechnical implications of converting a row of traditional stabilising piles into a row of thermo-active piles for the harvesting of heat from solar energy – which is meant to be subsequently used for road de-icing purposes - were investigated. An extensive centrifuge testing programme was conducted to quantify the overall performance of unsaturated slopes reinforced by thermo-active piles under different thermal loading conditions. Load transferred to the piles, bending moment distribution along the pile, and overall shear resistance of the model were considered key indicators of the slope performance. An analytical model was also developed for better understanding the influence played by temperature changes in this heavily coupled soil-structure interaction problem. Outputs of the analytical model were fed into 2D finite element models whose objective was to back-analyse the centrifuge tests results and provide further insights on the geotechnical implication of injecting heat into unsaturated slopes. Four major contribution were made through this thesis.

Firstly, a small-scale model of thermo-actives pile suitable for centrifuge testing was design, manufactured, and tested. The peculiarity of this model pile relies in its capacity of simultaneously scaling both the mechanical and the thermal properties of a prototype thermo-active pile. The methodology here developed – which consisted in embedding silicon tubing in a reinforced plaster-based mortar thermally enhanced by the addition of copper powder – allows the proper replication of the quasi-brittle behaviour of prototype reinforced concrete geo-structures and their ability to exchange heat with the soil in which they are embedded. The developed modelling technique can be easily adapted to and adopted for the modelling of any other thermo-active geo-structures (e.g., diaphragm walls).

Secondly, two systems for centrifuge testing of model slopes reinforced by thermo-active piles were conceived, designed, and tested. The first device consists of a centrifuge-mounted heating system which allows for the in-flight transfer of heat energy to model geo-structures. Proof heating tests performed up to 50-g suggest that when an appropriate pipe configuration is designed, the heating system is capable of generating a turbulent flow regime within the water circulation pipes, hence maximising the convective heat transfer mechanism. The second device consists of a large direct shear box (LDSB) apparatus for the modelling of translation-type of soil failure under high g-conditions. Proof tests performed under different gravity conditions suggest the developed LDSB apparatus is suitable for centrifuge modelling the translational slips both in the case of unreinforced and reinforced slopes.

Thirdly, a total of nine centrifuge tests were designed at a scale of 1:24 and performed under a corresponding gravity level of 24-g. The influence of soil type (i.e. heavily compacted artificial silt and loosely compacted natural silt) and heating patterns on the performance of a slope reinforced by a row of traditional and heated thermo-active piles was investigated. A reduction in the peak strength of the tested reinforced models was observed in all the cases involving heated thermo-active piles. Conversely, the residual strength of heavily compacted models was not affected by changes in temperature associated with the use of heated thermo-active piles.

Finally, an analytical model for defining p-y curves in the case of a heated square thermo-active pile belonging to a row of stabilising piles embedded in unsaturated soils was developed. The model was obtained by coupling the p-y Duncan and Evans (1982) model for c-φ soils with that describing the influence of arching effect on the load transferred to a row of stabilising piles proposed by Ito and Matsui (1975). The resulting analytical model incorporates terms describing the dependency on matric suction of soil shear strength parameters, and the thermal dependency of both matric suction and soil confining stress (the latter due to pile thermal expansion). The sensitivity analysis carried out on the modified Ito and Matsui (1975) model suggested that the ultimate load transferred to a pile embedded in unsaturated soils of given matric suction increases with the increase of the soil air entry value (AEV). The analysis further showed how temperature changes predominantly affect the load transferred to stabilising piles due to the change in soil-pile contact pressure resulting from the thermal expansion of the soil-pile system. p-y curves obtained through the analytical model were hence utilised within a beam-on-non-linear-Winkler-foundation model implemented in ABAQUS. The ABAQUS finite element model was then used to back-analyse results obtained in centrifuge tests performed on heavily compacted soil models. Information gathered through the back-analysis procedure suggests that retro-fitting semi-empirical coefficients commonly adopted for defining p-y curves in the case of traditional piles might be temperature dependent and hence not applicable to the case of heated thermo-active piles.
Date of Award2020
Original languageEnglish
SponsorsTransport Scotland
SupervisorJonathan Knappett (Supervisor) & Anthony Leung (Supervisor)


  • Centrifuge modelling
  • Thermo-active piles
  • Energy foundations
  • Unsaturated soil
  • Thermo-Hydro-Mechanical soil-pile interaction
  • Slope stability
  • Analytical model
  • p-y curves

Cite this