Subsidence is a significant geohazard affecting many areas globally, including the western part of the Netherlands. Subsidence is the lowering of the ground relative to the surface level, and it occurs both on the scale of a single structure and larger areas. The shallow subsurface in these areas, characterized by compressible soil, contributes to subsidence, which may be initiated or intensified by additional drivers. Subsidence can cause damage to buildings, leading to increased costs and social concerns

In the current state of the art, less distinction is made regarding the difference in effects between different subsidence drivers. This study aims to provide insights into the relative influence of different drivers on the damage to existing buildings with a shallow foundation. More specifically, it aims to understand how various conditions and factors influence the damage parameters associated with soil deformation. This approach allows for analysing the interactions between the soil and the structure, ultimately contributing to a better understanding of the relative influence of the different drivers.

Numerical models in PLAXIS 2D have been carried out to compute settlements for various scenarios, involving different soil scenarios, building scenarios, and subsidence drivers. The primary emphasis of numerical modelling lies on the soil rather than on the structure itself. Various damage parameters have been established based on the numerically calculated settlements. The purpose of this analysis is to determine the impact of settlements on the building based on these damage parameters.

Several aspects related to settlement occurrence and its impact on buildings were investigated. The study provides multiple outcomes regarding the interaction between soil scenarios, different drivers of subsidence, and the presence of an existing building with and without a partial basement. This was achieved by considering the influence of each of these variables on the damage parameters, with emphasis placed on the relative influence of the different drivers. The approach includes a method that compares the influence of the situation with the existing building and the situation without the existing building (greenfield situation). This comparison shows the settlement behaviour caused by the presence of the building load and its interaction with the soil. Additionally, the combined effects of the soil scenarios, different drivers, and building scenarios on the resulting damage to an existing building are considered.

To conclude, for the soil scenarios considered in this study, the soil scenario with a weak spot (SS3) has the most unfavourable effect on an existing building. For the drivers considered in this study, when evaluating the relative influence of the different drivers, the global groundwater lowering (D2glo) has the most unfavourable effect. Considering the combined effect, the soil scenario exerts the greatest influence on the resulting damage parameters for the evaluated scenarios, followed by the type of driver and the building scenario.

Due to the Netherlands' topography and geology, many regions have weak soil resistance, causing numerous buildings to experience settlements from subsidence processes, with masonry structures being the most commonly affected. In these buildings, damage typically appears as cracks and deformations, indicating the foundation’s inability to support the structure. To enhance decision-making and the engineering of effective countermeasures in these scenarios, accurate and reliable building assessments are needed to predict the expected damage based on the amount of soil deformations. This thesis, therefore, aims to evaluate the capabilities of a set of state-of-the-art damage assessment methods. These methods have been applied to a case study, with their results benchmarked against evidence from a Building Foundation Assessment report, where recorded damage features were used to evaluate the accuracy and characteristics of each assessment method.
The study began with a Visual Assessment using a Decision Diagnostic Support Tool to analyze damage features and hypothesize the causes of the building's behavior. This was followed by an Empirical Assessment, applying empirical limits to relate expected damage to Subsidence-Related Intensity (SRI) parameters. Next, an Analytical Assessment used the Limit Tensile Strain Method (LTSM) to approximate building deformations, treating it as a linear-elastic isotropic masonry beam and correlating strain estimates to damage levels. Finally, a 2D Finite Element Analysis (FEA) using a continuum crack-modelling approach was conducted on the most damaged wall to more accurately reproduce the crack widths, crack locations and the behaviour of the wall.
The results show that while the building’s damage state can be approximated with reasonable accuracy, challenges remain in predicting specific damage features. The visual assessment successfully identified the building’s underlying mechanism. Empirical and analytical methods accurately predicted damage levels in 5 out of 6 walls, proving to be efficient assessment techniques. The 2D Finite Element Analysis (FEA) successfully simulated the crack pattern on Wall 2 with a Root Mean Square Error (RMSE) of +1 Ψ or +4.7mm against the maximum mean crack widths and reproduced 5 out of 7 cracks with similar characteristics. Additionally, FEA results showed that mesh sizes of 200, 100, and 50 mm made results deviate by σ = 0.33 Ψ and σCWmax = 2.3mm, with observable changes in crack shapes in EMM models.
To address the slight deviations in the less accurate analysis of the outer leaf, primarily driven by conservative crack width estimates, a Bayesian Optimization procedure was used on the outer leaf models to identify the optimal set of material parameters that minimized the discrepancy between the damage state of the results and the target damage level in the case study.
The implementation of the approach demonstrated sufficient efficiency in identifying the optimal set of parameters, despite the computational expense of the Finite Element models. The procedure’s effectiveness varied across models with it significantly reducing damage levels in the Engineering Masonry Model (EMM) variations but showed more limited improvements in the Total Strain Crack Model (TSCM). Additionally, the approach allowed for an investigation into the influence of material properties, revealing that Young's Modulus and tensile strength were the most influential parameters across both models. Furthermore, the results indicated that the influence of material parameters is highly non-linear, meaning changes in material properties do not always lead to predictable outcomes. Instead, specific combinations of parameters had a greater impact on reducing damage, demonstrating the complex interplay between material properties particularly in the EMM model variant.