Influence of Structural Stiffness Representation in Settlement Calculations and Practical Advice

Abstract

State-of-the-art geotechnical and structural analyses commonly rely on finite-element analysis, treating soil and structure models separately. Incorporating the overall stiffness of the structure into the geotechnical model enables more realistic settlement predictions and assessments of founding force distribution, ensuring optimized and cost-effective designs. Although stiffness increases with building height, the effective contribution to settlement control is limited to a finite number of floors, making the internal stiffness distribution more relevant than the construction method itself. Reliable predictions require modeling approaches that realistically represent structural stiffness. This study evaluates various equivalent-model techniques in PLAXIS 3D and SOFiSTiK for their practicability and accuracy and explains their application. While simplified methods can adequately capture total stiffness under uniform distribution, constructing a full 3D model may be faster or more practical than elaborate simplifications requiring extensive setup and post-processing. Moreover, the construction sequence and evolving stiffness during the build, as well as time-dependent effects like creep and limit-state-dependent stiffness changes within structural elements, significantly influence settlements. A 3D structural model allows these factors to be accounted for comprehensively. Therefore, whenever feasible, geotechnical settlement analyses should employ full 3D structural models.

1. Introduction

In practical geotechnical analysis, settlement troughs are often assessed by modeling only the foundation slab, while neglecting the stiffness of the superstructure. Although this approach reduces modeling effort, it can lead to substantial discrepancies in settlement predictions. A more accurate alternative is to include the structural system within the geotechnical model, thereby achieving a more realistic representation of soil–structure interaction [1,2]. Preliminary studies comparing two construction types, frame-type and wall-type buildings, have shown that neglecting the superstructure results in significantly larger deformations compared to a full 3D reference model. These differences are not merely quantitative, they also affect load redistribution within the structure, underscoring the need to represent superstructure stiffness adequately in geotechnical models.

Earlier research on this topic can be found in [3], while more recent studies have advanced the state of knowledge by applying modern finite-element workflows. These works have demonstrated how superstructure stiffness representation, load distribution, and soil parameter calibration influence settlement outcomes. Examples include reinforced concrete structures on footing foundations [4], three-dimensional SSI in layered soils [5], and the impact of superstructure loading schemes on mat foundations [6]. Further investigations addressed differential settlements of mats on sand [7] and high-rise systems with piled rafts [8], while field-informed and numerical case studies have clarified static SSI mechanisms and building responses [9,10]. Together, these contributions highlight the relevance and capabilities of modern FEM tools such as PLAXIS 3D [11], a widely used geotechnical software for soil–structure interaction, and SOFiSTiK [12], a finite-element package for structural engineering, in enabling detailed SSI analysis.

In everyday engineering practice, however, such advanced modeling is often considered too time-consuming and complex. To address this gap, a variety of simplified substitute models have been proposed in the classical literature, aiming to approximate the effect of superstructure stiffness with manageable modeling effort. The present study, building on preliminary results from [13], builds on these established approaches, comparing their accuracy and practicality in settlement analysis and providing guidance on their applicability in design-oriented workflows.

2. Method and Model Types

The structural models developed in this study were designed to represent a broad range of typical building configurations and associated stiffness distributions and have been extensively verified and published in previous works [14]. The models differ in both their geometric layout and structural design. Two fundamental structural systems are considered: a frame-type structure, characterized by low stiffness and relying on columns as vertical load-bearing elements, and a wall-type structure, which includes load-bearing interior walls and provides significantly higher overall stiffness.

Within each of these systems, two basement configurations are investigated: a “small basement,” whose dimensions correspond to those of the superstructure, and a “large basement,” where the foundation slab extends laterally beyond the building outline. In addition, two further variants are included in which an intermediate soft floor is introduced. This configuration, with an open ground floor and increased floor height, reflects a common design situation in practice, such as entrance foyers or open retail areas, which significantly reduce local shear and bending stiffness. An overview of all six model variants is shown in Figure 1.

All buildings are 15-floor structures with a constant floor height of 3.5 m. The floor plan dimensions range from 27 m × 7.5 m (small) to 34.5 m × 15 m (large) and are based on the assumption of double symmetry. Floor slabs have a uniform thickness of 28 cm, basement and exterior walls are 30 cm thick, and interior shear walls are 20 cm thick. The foundation slab is modeled with a thickness of 1.75 m, and the columns have a cross-section of 0.5 m × 0.5 m.

The models were generated without considering construction stages and were subjected exclusively to self-weight and a constant live load of 5 kN/m². To reduce computational time, each building was modeled as a quarter-structure by exploiting geometrical symmetry. Boundary conditions included fixed horizontal displacements along the model edges and symmetry planes, as well as a fully fixed base. Groundwater was not considered, and all simulations were performed under drained conditions. All materials were assumed to behave linearly elastically (see

Table 1. Linear elastic material parameters for the structural and the geotechnical model.

Materials	Self Weight	Elastic Modulus	Poisons Ratio
	[kN/m3]	[MN/m2]	[-]
Concrete	25	30,000	0.2
Soil	18	16	0.2

). A zero-thickness interface was defined between the basement walls and the soil to enable realistic shear transfer. The soil was discretized with about 33,000 10-node tetrahedral elements, and the superstructure was represented by elastic plate and beam elements to capture its stiffness in a consistent but simplified manner.

In this study, the 3D structural model developed in SOFiSTiK serves as the reference for defining a reliable representation of global structural stiffness. To evaluate whether a corresponding structural model recreated within the geotechnical software PLAXIS 3D 2024 can capture the same stiffness characteristics, the building was reconstructed using equivalent structural elements. Since the nonlinear behavior of the soil cannot be adequately represented in SOFiSTiK 2024, a direct comparison between the two modeling environments requires both models to employ a linear-elastic soil approximation. Although such an approximation is generally not recommended for geotechnical analyses, as highlighted in [15], it is applied here solely for the purpose of validating the structural response.

Settlement predictions are highly sensitive to soil constitutive behavior and input parameters [16,17]. Key influences include the nonlinear stiffness evolution with stress [18] and strain [19] levels, the stress path [20], and the stress history [21], as well as plasticity effects in edge zones [22]. While simplified approaches often rely on a constant “representative” modulus, this neglects natural heterogeneity and uncertainty. Advanced constitutive models (e.g., hardening soil small [23]) and geostatistical treatment of parameter variability [24] offer more realistic descriptions. Furthermore, foundation and superstructure stiffness significantly interact with soil response [25]. In this study, a deliberately very deformable soil was adopted, not to represent a specific ground condition but to emphasize the relative influence of structural stiffness on settlement behavior. In practice, stiffer soils would reduce these differences, but the chosen parameters ensure controlled and stable settlement patterns that allow for a systematic evaluation of structural effects.

Extensive validation was carried out by comparing settlement results from the SOFiSTiK model and the PLAXIS model under identical boundary conditions and soil properties. As shown in Figure 2, the deviation in vertical displacements between the two simulations is less than 1%, indicating that the PLAXIS model captures the global structural stiffness with sufficient accuracy for the intended analyses.

3. Results

3.1. Influence of Floors on Structural Stiffness

The literature provides varying opinions and recommendations on how many floors contribute significantly to the overall system stiffness in the context of soil–structure interaction [26,27]. Especially in high-rise buildings, it can be advantageous to reduce modeling complexity by introducing a so-called “settlement-relevant number of floors” or contributing floors, which refers to the number of floors to be considered when estimating structural stiffness.

In general, basement levels should always be included, as they are typically constructed with stiff, continuous shear walls and contribute substantially to the overall bending stiffness of the structure. For buildings with inherently high stiffness, such as reinforced-concrete structures dominated by continuous load-bearing shear walls (referred to here as “massive shear wall systems”), it is common practice to consider two to five additional upper floors [28]. In contrast, for more flexible frame-type structures, the contribution of upper floors can often be neglected [29]. In these cases, it is important to assess whether vertical elements such as elevator cores or stairwells provide sufficient shear and bending stiffness to ensure composite action with the floor slabs throughout the building height.

To systematically determine how many floors contribute meaningfully to settlement-relevant stiffness, the number of active floors is gradually reduced in a parametric study. Specifically, floors are successively removed from a massless structural model, while the applied foundation loads remain constant. The objective is to identify the point at which differential settlements along the x-axis of symmetry begin to change noticeably, which marks the threshold beyond which additional floors no longer contribute significantly to the system stiffness.

The loads applied to the foundation correspond to the support reactions of a separately computed, fully rigid structural reference model developed in SOFiSTiK. These include the self-weight of the superstructure and floor slabs, as well as all live loads defined in the original design scenario (Figure 3). The analysis is based on six structural configurations that differ in global stiffness and basement geometry, as described in Section 2.

The evaluation of the results presented in Figure 4a (small basement) and Figure 4b (large basement) illustrates the relationship between the number of active floors and the resulting differential settlements. The plateau-like regions in the curves clearly indicate that the lower floors contribute most significantly to the system stiffness relevant for settlement behavior. As a criterion for sufficient stiffness representation, a deviation of less than 0.5% from the full reference solution is used.

For models with a small basement footprint, it is evident that the lower three to four floors dominate the structural stiffness contribution, regardless of the construction type (frame or wall-type). However, further observations from the remaining models reveal that variations in stiffness across floors within a building can have a major impact on the number of contributing floors. For instance, in the wall-type model with a large basement, the stiff basement levels already contribute significantly to reducing excessive settlements. Yet these alone are not sufficient to reach the required accuracy. Only when up to eight additional stiffer floors above the basement are included does the deviation from the reference model fall below the 0.5% threshold.

Another notable effect can be observed in the wall-type model with a small basement and a soft intermediate floor. This configuration features a nearly constant stiffness distribution, except for a soft floor at ground level. As a result, a distinct kink appears in the settlement-difference curve, clearly marking the threshold of structural contribution. This abrupt drop in stiffness is caused by the soft floor, which, being composed entirely of columns, provides limited shear transfer to the adjacent floors.

Nevertheless, upper floors above this soft level must not be neglected. Their contribution depends on the relative stiffness distribution. In the WALL-LB-SF model, for example, a similar kink is visible, but the stiffer upper floors are still found to significantly influence the global behavior, an effect consistent with that observed in the WALL-LB configuration.

3.2. Substitute Model Approaches

3.2.1. Substitute Stiffness by Steiner

The DIN 4018 [30] outlines a method for calculating a substitute bending stiffness for structures using Steiner’s theorem. In this approach, the superstructure is divided into a set of cross-sections. For each section, the respective moment of inertia is calculated. The global substitute stiffness is then obtained by summing up the individual moments of inertia $I_{0,i}$ including their Steiner share (depending on cross-section $A_i$ and distance to the center $y_i^2$). The substitute bending stiffness is given by

$${EI}_{sub,mod}=\left({\sum }_{i=1}^{n_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}{E}_i{I}_{0,i}+\alpha {\sum }_{i=1}^{n^{\prime }}{E}_i{A}_i{y}_i^2\right)$$

(1)

Here, $n_{total}$ is the total number of floors, $n^{\prime }$ is the number of effective floors, and $\alpha$ is a reduction factor accounting for reduced connection stiffness and can be calibrated based on the floor height or connection details [31].

To apply this inherently two-dimensional approach to a three-dimensional foundation slab, the substitute stiffnesses determined from each cross-section is projected onto a participating slab width. This is carried out following the recommendations in [32], enabling a spatially resolved representation of the stiffness distribution. The result is a stiffness map over the foundation area (

Table 2. Projected substitute stiffness on the foundation slab, according to Steiner, obtained as substitute thickness $d_{\mathrm{s}\mathrm{u}\mathrm{b}}$ for individual cross-sections.

Frame-Type Structure with Small Basement	Structural Section	$d_{\mathrm{s}\mathrm{u}\mathrm{b}}$
	Foundation plate	1.75 m
	CS 1	8.07 m
	CS 2	8.98 m
	CS 3	7.50 m
	CS 4	8.19 m
	CS 5	9.19 m
	CS 6	7.5 m
Wall-Type Structure with Small Basement	Structural Section	$d_{\mathrm{s}\mathrm{u}\mathrm{b}}$
	Foundation plate	1.75 m
	CS 1	8.07 m
	CS 2	8.98 m
	CS 3	7.50 m
	CS 4	8.20 m

) that reflects the structural configuration and interaction effects realistically. The obtained substitute thickness $d_{sub}$ is subsequently implemented in the geotechnical model by utilizing plate elements with a constant stiffness of 30 MPa but varying thicknesses.

3.2.2. Substitute Stiffness by Meyerhof

In [33], the concept of substitute structural stiffness is based on the summation of projected moments of inertia of vertical structural elements, modified by a so-called dowel action factor. This factor accounts for the fact that the shear connection between walls and floor slabs is typically much stronger than that between columns and slabs.

The dowel action factor is greater than one and depends on the relative stiffness of the structural elements above and below the considered floor slab. The distribution of stiffness is therefore captured by comparing the respective moments of inertia of adjacent floors. Geometry is included by considering the slab dimensions relative to its field width and the total number of spans in the assumed cross-section.

While Meyerhof’s formulation provides clear definitions for calculating substitute stiffness in frame-type structures (columns), it does not include an explicit procedure for systems with continuous, shear-connected walls. To address this limitation and enable the inclusion of such walls in the present study, each continuous wall is idealized as a column with an equivalent cross-section. This equivalent column is defined by a width b and a thickness d, which allows it to be incorporated into the substitute stiffness summation in line with Meyerhof’s method. This idealization enables a unified treatment of both frame-type and wall-type buildings within a common substitute stiffness framework while still acknowledging the enhanced shear transfer capacity of continuous wall systems:

$${EI}_{Meyerhof}={\sum }_{i=1}^{n_{total}}{E}_i{I}_{slab} \left(1+{\displaystyle \frac{\frac{I_a}{h_a}+\frac{I_b}{h_b}}{\frac{I_{slab}}{l_{field}}+\frac{I_a}{h_a}+\frac{I_b}{h_b}}}{n}_{field}^2\right)$$

(2)

where $n_{total}$ is the total number of floors;$I_{slab}$ is the moment of inertia of the slab;$I_a \mathrm{a}\mathrm{n}\mathrm{d} I_b$ are the moment of inertia above and below the slab;$h_a \mathrm{a}\mathrm{n}\mathrm{d} h_b$ are the height of the wall/column above and below the slab;$l_{field}$ is the length of the surrounding field $; \mathrm{a}\mathrm{n}\mathrm{d} n_{field}$ is the number of considered fields within the cross-section.

In contrast, if the wall is not rigidly connected to the floor slab, such as infill panels in frame-type structures, the DIN Technical Report 130 [27] provides a procedure for estimating the substitute bending stiffness. In this case, the total bending stiffness of the structural system is calculated as the sum of three components: the bending stiffness of the foundation system, the bending stiffness of the infill wall, and the bending stiffness of the skeletal frame, typically modeled as a single-floor frame.

To extend this inherently two-dimensional approach to a three-dimensional foundation slab, the substitute stiffness values determined for each cross-section are projected onto a participating slab width. This is carried out following the recommendations in [32], enabling a spatially resolved representation of the stiffness distribution. The result is a stiffness map across the foundation area (

Table 3. Projected substitute stiffness on the foundation slab, according to Meyerhof, obtained as substitute thickness $d_{sub}$ for individual cross-sections.

Frame-Type Structure with Small Basement	Structural Section	$d_{\mathrm{s}\mathrm{u}\mathrm{b}}$
	Foundation plate	1.75 m
	CS 1	3.72 m
	CS 2	2.15 m
	CS 3	2.02 m
	CS 4	7.63 m
	CS 5	2.20 m
	CS 6	2.78 m
Wall-Type Structure with Small Basement	Structural Section	$d_{\mathrm{s}\mathrm{u}\mathrm{b}}$
	Foundation plate	1.75 m
	CS 1	3.72 m
	CS 2	2.38 m
	CS 3	7.23 m

) that reflects the structural configuration and interaction effects. The obtained substitute thickness d_sub is then implemented in the geotechnical model by applying plate elements with a constant elastic modulus of 30 MPa but varying thicknesses. The model itself is massless and subjected to the support reactions obtained from the rigid reference model.

3.2.3. Beam-and-Plate Model

The so-called beam-and-plate model proposed by Werkle and Slongo [34] provides a fast and effective procedure for incorporating wall stiffness into settlement analysis. The core idea is to replace the load-bearing walls of a building, those rigidly connected to the foundation slab, with equivalent vertical beam elements (Figure 5). These beam elements are assigned the same cross-sectional and material properties as the original walls.

In practical implementation, the structural walls connected to the foundation slab are replaced by beam elements with an equivalent “substitute height.” This height is not derived directly from the actual floor height but is calculated based on the moment of inertia of the wall section. The reference axis is shifted to the level of the foundation slab, similar to the inclusion of the Steiner (parallel axis) component, which results in a substitute height exceeding the physical wall height. Separate substitute heights were determined for exterior and interior walls and for each construction type, frame-type, and wall-type structures. To account for openings in the exterior walls, the substitute height is further adjusted by applying an opening factor. The model itself is massless and subjected to the support reactions obtained from the rigid reference model. The calculation procedure and basic assumptions are summarized in

Table 4. Calculation and assumptions for the substitute beam height in the beam–plate model for the frame-type and wall-type model.

Frame-Type Structure with Small Basement			Cellar Wall	Core Wall
	Moment of inertia	I0	8.6 m4	3617 m4
	Steiner part	Σ Ai y2(n′)	25 m4	10,852 m4
	Eq. moment of inertia	Ieq	34 m4	14,470 m4
	Eq. Structure height	heq	11.1 m	83.3 m
Wall-Type Structure with Small Basement			Exterior Wall	Interior Wall
	Moment of inertia	I0	3617 m4	2411 m4
	Steiner part	Σ Ai y2(n′)	10,852 m4	7235 m4
	Eq. moment of inertia	Ieq	14,470 m4	9646 m4
	Eq. Structure height	heq	83.3 m4	83.3 m4
	Opening factor	$A_{Wall}/A_{Window}$	0.35
	Eq. Structure height (mod)	heq,mod	54.5 m

.

3.2.4. Substitute Volume Stiffness

Another approach for approximating the structural stiffness in settlement calculations is the use of a homogenized substitute volume, adapted from [35] in the context of tunnel-induced settlements. Instead of representing the full structural geometry and stiffness, a simplified solid block with the same footprint and height as the basement is introduced into the geotechnical model. This volume is assigned with uniform material properties throughout and is intended to capture the mechanical response of the actual building in a computationally efficient manner. As with other substitute approaches, the model is loaded using the foundation reactions derived from the rigid structural reference model.

In practical application, the substitute volume is generated in parallel to the full 3D reference model. As described in Figure 6, a predefined settlement trough based on field data or synthetic input is then imposed on both models. The material parameters of the substitute volume are iteratively adjusted until the sum of the resulting support reactions matches that of the structural reference model. Once calibrated, the resulting material properties can be transferred to the geotechnical model and used for further settlement analysis.

To address local effects such as punching due to point loads, the location of the point loads was slightly shifted upward (half foundation height). This adjustment resulted in a more realistic load distribution across the substitute volume. The final iteration derives a substitute stiffness $E_{final}$ of 2910 MN/m² for the frame type building and 9690 MN/m² for the wall-type building, showing the clear difference between these two building types.

3.2.5. Knot-Dependent Substitute Stiffness (Force Method)

The approach introduced by Sommer [36] calculates the substitute bending stiffness by applying the force method. The technique enables the determination of a node-specific substitute stiffness by imposing a unit displacement at each support node. From this, an equivalent substitute thickness can be derived.

Initially, the foundation slab is idealized as a grid of continuous beams, supported by a defined number of nodes in both the x- and y-directions, as illustrated in Figure 7. In parallel, the full 3D structural model is rigidly supported at the same nodal pattern. A unit vertical settlement s is then imposed at each support node of the rigidly supported 3D model. The resulting reaction force $Q_k$ at each node reflects the local stiffness contribution of the superstructure—larger values are typically found beneath stiff structural cores, whereas smaller forces occur under isolated columns.

The force $Q_k$ obtained from the 3D model is then applied conceptually to the corresponding node in the beam grid system. Using the force method, the equivalent bending stiffness at that location can be obtained.

This calculation is performed successively for each support node, resulting in a distributed map of substitute bending stiffness values across the slab. These stiffnesses are subsequently converted into equivalent substitute slab thicknesses. An example of this thickness distribution for a frame-type structure is shown in Figure 8. The location of the structural cores is clearly identifiable by locally increased substitute thicknesses (e.g., 4.01–4.13 m), illustrating the spatial variation in structural stiffness.

4. Discussion

To evaluate the influence of different modeling approaches on the simulation of structural stiffness in geotechnical settlement analyses, several methods were tested and compared. Figure 9 provides an overview of the resulting differential settlements for both wall-type and frame-type buildings, using various simplifications and modeling strategies.

These include a detailed 3D structural model as reference, a foundation plate model without any superstructure, and multiple substitute stiffness approaches: the Steiner method, the Meyerhof method, a node-specific (force-based) calculation, as well as volume-averaged and beam–plate simplifications.

The objective of this comparison is to assess how well each approach captures both global and local settlement behavior and to evaluate their suitability depending on the structural system and modeling objective. The findings are discussed in detail below, and a qualitative assessment is provided in

Table 5. Qualitative evaluation of superstructure stiffness modeling methods. Recommendation of methods based on structural type, ability to represent local effects (e.g., core stiffness, edge flattening) where ✓✓ = very suitable, ✓ = suitable, and ✗ = not recommended.

Method	Frame-Type Structure	Wall-Type Structure	Core Behaviour	Edge Behaviour
Substitute Stiffness (Steiner)	✓✓	✓✓	✗	✓
Substitute Stiffness (Meyerhof)	✓	✓	✗	✗
Force Method	✓✓	✓✓	✓	✓
Substitute Volume Model	✗	✓✓	✗	✓✓
Beam–Plate Model	✗	✓✓	✗	✓✓

.

Among the tested approaches, the substitute stiffness derived using the Steiner method provides the best global agreement for both frame-type and wall-type structures. It reproduces absolute settlements well in central slab regions but overestimates stiffness near core areas, leading to inaccuracies in curvature representation.

The substitute stiffness derived using the Meyerhof method yields moderate results. While the general shape of the settlement trough is preserved, it tends to underestimate stiffness, particularly in frame-type structures. The method performs better for wall-type buildings but cannot accurately reproduce local deformation effects.

The force-based method, derived from a knot-specific stiffness response, offers good agreement for both structure types and enables spatially varying stiffness representation. By deriving stiffness from unit settlements and reaction forces, it captures local effects such as core stiffness more effectively. It is particularly suitable for automated evaluations and detailed interpretation of stiffness distribution, though refinement may be needed in zones of highly concentrated stiffness.

The substitute volume model is well suited for wall-type buildings. It effectively reproduces both the global settlement shape and the edge flattening resulting from vertically extended stiffness (frame action). However, in frame-type buildings with localized stiffness, such as cores, the homogeneous stiffness model is less accurate.

The beam–plate model performs very well at the slab edges and captures multi-floor frame action. It is ideal for wall-type structures but underestimates stiffness in column-dominated frame buildings.

In summary, the Steiner method and the force-based analysis perform well globally, the Meyerhof method is limited by core effects, and the substitute volume and beam–plate approaches best capture frame behavior and edge conditions. A qualitative evaluation of the methods is provided in Table 5.

Overall, these observations are not only consistent within the controlled framework of this study but also reflect the general trends reported in the broader context of soil–structure interaction research. This confirms that the comparative evaluation of substitute-stiffness methods presented here captures the essential mechanisms governing settlement behavior, while remaining focused on a systematic internal comparison under well-defined boundary conditions.

From a practical design perspective, the findings demonstrate that full 3D structural models should be preferred whenever possible, as they provide the most reliable predictions of settlements and bending moments. Simplified substitute methods may still be used for preliminary assessments, but their applicability is limited to approximate estimates.

5. Conclusions

This study addresses the bearing capacity and settlement behavior of buildings, focusing on different methods for incorporating superstructure stiffness in soil–structure interaction. Both simplified geometric models and analytical methods for calculating substitute stiffness were investigated.

The comparison between frame-type and wall-type structures shows that the distribution of stiffness within the building strongly influences the settlement pattern. An important result is that the number of floors contributing to settlement can vary. In buildings with uniform stiffness, the lower three to four floors usually have the greatest influence. However, if the basement levels are softer than the rest of the building, more floors become relevant for settlement.

A key finding is that models considering only the foundation slab without superstructure stiffness often lead to large discrepancies compared to full 3D models. The performance of each method depends on the structural system and the distribution of stiffness. For wall-type structures, the tested methods show very good agreement with the reference model. Local effects, such as curvature changes near slab edges, are captured well, except by the method using substitute thicknesses (force method), which is less accurate in these regions.

The situation is different for frame-type structures. Buildings with central stiff cores exhibit large local differences in settlement, which simplified models cannot capture accurately. The stiffness concentration in the core produces a distinct deformation pattern. A better match can be achieved by reducing the shear stiffness in the core or by including the basement structure in the model.

In general, the total stiffness of the structure can be represented reasonably well with simplified methods, provided that stiffness is evenly distributed. However, although developing a 3D model for geotechnical analysis requires additional effort, it may be more efficient and practical than using complex simplified methods, which often involve substantial preparation and post-processing.

Furthermore, the construction process has a strong influence: both the loading sequence and the gradual development of structural stiffness during construction must be considered. Time-dependent effects such as creep, as well as stiffness changes under service loads, are also important. A 3D model that explicitly includes the actual structural elements allows these effects to be represented appropriately.

All these considerations lead to a clear recommendation: whenever possible, a 3D structural model should be used in geotechnical settlement analysis.