Quantifying building damage induced by underground construction at city scale
Doktorsavhandling, 2025
The resulting subsidence can increase the risk of flooding and damage to structures, causing significant costs to society and individuals. This is especially common in areas with deformation-sensitive clays, typical to Scandinavia. The pore pressure reduction caused by leakage to tunnels built in fractured rock, called underdrainage, is common in Scandinavia and may affect a large area of influence, requiring expensive mitigation measures. Planning for such mitigation measures to prevent significant damage involves a trade-off between costs and benefits of their implementation. We need an understanding of the processes involved in the large-scale subsidence due to underdrainage, and the potential damage. The problem is highly complex, uncertain, and coupled. Thus, a rigorous modelling framework was developed as part of this thesis.
A 2D hydro-mechanically coupled finite element (HMFE) model with an advanced rate-dependent constitutive model was used to assess the influence of key parameters affecting the time-dependent (greenfield) settlements caused by underdrainage. The time-dependent results from the HMFE model were then used along with hydro-stratigraphy as training and validation data for a metamodel, which produced settlements on a 3D hydro-stratigraphic domain. The results of the metamodel were used as input for a large-scale building damage model developed to calculate damage parameters based on various damage criteria. In urban settings, greenfield predictions potentially result in over-prediction of damage, as most buildings on soft deposits are built on piled foundations. Thus, the influence of floating piles was considered via axisymmetric finite element simulations, providing modification factors that can be applied to greenfield settlements for older piled buildings.
The developed models were applied to the case of the construction of an underground railway station in Gothenburg, Sweden, expected to affect a confined aquifer in an area with ongoing background creep settlements. Hypothetical scenarios of uniform pore pressure reduction under the soft clay deposit, and the subsequent consolidation and creep in the soft clay, were simulated to determine the magnitude of settlements at different time intervals. The results showed dependency on time and stratigraphy, where the transitional areas between the soil and rock are most prone to settlements in short term, influenced by the overconsolidation ratio (OCR). The predicted settlements, as well as the modified settlements due to the existence of floating piles were shown to be highly time-dependent, due to the slow consolidation associated with underdrainage in deep clay layers. Furthermore, the results also show that the estimated damage is highly dependent on the chosen damage criteria. Based on the results, only damage parameters derived from differential settlements should be used in early damage assessments.
underdrainage
creep
consolidation
building damage
groundwater lowering
subsidence
Författare
Pierre Wikby
Chalmers, Arkitektur och samhällsbyggnadsteknik, Geologi och geoteknik
The influence of parameter variability on subsidence
10th European Conference on Numerical Methods in Geotechnical Engineering,;(2023)
Paper i proceeding
A metamodel for estimating time-dependent groundwater-induced subsidence at large scales
Engineering Geology,;Vol. 341(2024)
Artikel i vetenskaplig tidskrift
A grid-based methodology for the assessment of time-dependent building damage at large scale
Tunnelling and Underground Space Technology,;Vol. 149(2024)
Artikel i vetenskaplig tidskrift
Modeling subsidence and building damage in central Gothenburg using machine learning
;Vol. 1(2025)p. 375-378
Övrigt konferensbidrag
This thesis develops a modelling framework to better understand and predict subsidence and building damage. Subsidence modelling, based on advanced 2D simulations combined with machine learning for extending the results to 3D, was used to predict subsidence at district scale. These results were linked to a building damage model that automatically evaluates damage for buildings in affected zones. Because many buildings are piled, additional simulations accounted for the resistance of their foundations to subsidence, refining damage estimates.
Applied to central Gothenburg, Sweden, the work shows subsidence to be strongly time- and space-dependent. Damage also depends heavily on the chosen assessment method. Additionally, piles that float in clay only resist damage after many years of tunnel leakage. By revealing how subsidence unfolds and impacts buildings, this research provides tools to predict risks, reduce costs, and support safer underground development in growing cities.
Digital Twin Cities Centre
VINNOVA (2019-00041), 2020-02-29 -- 2024-12-31.
Modellering av tidsberoende grundvattensänkning, markdeformationer och dess skaderisker
Trafikverket (TRV2020/54637), 2023-10-01 -- 2025-12-19.
Trafikverket (TRV2020/54637), 2020-09-01 -- 2023-08-31.
Styrkeområden
Transport
Ämneskategorier (SSIF 2025)
Geoteknik och teknisk geologi
DOI
10.63959/chalmers.dt/5809
ISBN
978-91-8103-352-6
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5809
Utgivare
Chalmers
SB-H1, Sven Hultins gata 6
Opponent: Mandy Korff, TU Delft, Nederländerna