For the first time since the 1970s, concrete plans are in motion to bring humans back to the Moon through the Artemis programme. The dawning era of space exploration aims to set the groundwork for long-term off-Earth settlement, with sights on bringing the first humans to Mars. In order to provide a safe means of inhabiting extraterrestrial (ET) environments, a self-supporting habitat shield concept has been developed by the Specialist Modelling Group at Foster + Partners, which is based on the use of mechanically interlocking masonry manufactured from in-situ resources. Significant progress must still be made in the structural design and validation of such systems, however, before they can be safely implemented. Due to the novelty of the proposed structure, existing research does not provide numerical evidence of its stability under self-weight, or under seismic loading. Research in this direction is valuable, as it explores alternatives to popular design concepts based on monolithic additive manufacturing and addresses research gaps related to the analysis of mechanically interlocking structures.

The aim of this project is to provide novel insight into the stability of self-supporting dry-stone vaults under the influence of microgravity and ground motions. This is achieved through a three-part analysis approach, which starts with an investigation into the fundamental structural behaviour of dry-stone, Nubian-type vaults based on a series of parametric studies. Key geometric parameters of the vaults are varied, and the resulting distinct element models are subjected to quasi-static pushover-type analysis in 3DEC. The parametric studies indicate what configuration of the vault geometry can maximise seismic capacity. This optimal vault geometry is modelled in Rhino using geometric constraint solving, which is used to produce a model which can be assembled from mechanically interlocking components. Through pushover-type testing and dynamic time history analysis, the performance of the resulting vault model is assessed with respect to possible moonquake loading. The final part of the design process accounts for uncertainties in material properties by conducting sensitivity studies, which indicate what degree of variation in performance is possible.

After completing the three mains sets of analysis needed to design a moonquake-resistant vault, several comparative analyses are carried out to provide context to the structural performance of the Nubian-type vault. The first of these additional investigations models an equivalent monolithic vault geometry in DIANA and conducts finite element analysis to determine its capacity for lateral acceleration. The second comparative study determines how covering the vault in loose regolith may influence its structural performance. The third study investigates how the variation of gravity affects the stability and seismic performance of the vault.

Several key findings are obtained in this project, leading to an understanding of the structural behaviour of Nubian-type vaults and how they may be a possible solution to shielding demands in ET contexts. The final vault model was shown to remain stable under a uniform lateral acceleration equivalent to the PGA for a possible moonquake with a 475 year return period. Dynamic analysis conducted using an artificial ground motion record obtained for the same return period produced less favourable results, resulting in partial collapse of the outermost six courses of the vault. Computational constraints necessitated the implementation of an undamped analysis, however, so these results are considered to be conservative. Further analysis is needed, but these results suggest that with minor adjustments, the vault may be able to safely resist moonquake loading. The comparative monolithic vault analysis demonstrates that an additively manufactured vault may provide slightly better resistance to static lateral acceleration than a component-based vault. When considering the structural redundancy and energy dissipation necessary for safe seismic design, however, the discrete vault is expected to perform better. Results show that covering the vault in loose regolith is expected to reduce its overall stiffness. Conclusions about changes in stability cannot be made with confidence due to modelling issues at the interface between the regolith and the structure. By studying the influence of gravitational variation on the vault, results indicate that it may perform favourably on the Earth or on Mars, though the possibility of material damage must be studied further.

Designing structures that are more sustainable is a relevant topic within the construction industry. By choosing materials that have a low embodied energy value and optimizing the structures that can be constructed by these materials, one could potentially minimize the economical and environmental footprint of a structural design.

As burnt clay bricks have a relatively low embodied energy value, are relatively cheap as a construction material and are relatively durable, it is interesting to investigate the optimization of masonry structures. To achieve this, use is made of topology optimization. To accurately optimize the topology of masonry structures however, this optimization must be performed based on the results of a discrete element analysis. This thesis presents several methods to set up such a model based on masonry structures of arbitrary size and lay-out, departing from the smallest scale: the individual brick.

First, a method is developed to create arbitrary shapes for bricks. An algorithm is developed to parametrically create structures for two distinct shapes. A procedure to abstract these structures and translate their geometrical representation to a simplified numerical model is then presented. Several
methods for structural analyses are detailed and their results are evaluated and compared. The results of these analyses are used to optimize the topology of the initial structure by means of the Method of Moving Asymptotes. The resulting structures are then verified using 3DEC. Finally, some applications of the developed method are presented along with future visions.

The extraction of natural gas in the northern part of the Netherlands, from the region of Groningen, has been causing human-induced seismic activities for the past several decades. This is a problem since the existing building stock in this region, which consists of mainly unreinforced masonry buildings and historical structures, are not designed to withstand seismic events due to the lack of empirical earthquake-resistant design features. Further, the combination of a soft topsoil and the gas extraction, is responsible for ground settlements which may compromise the capacity of the existing buildings.

The bed-joint reinforcement technique is a strengthening method which consists of cutting a slot in the bed-joints and installing steel bars embedded in a high-strength repair mortar. Although this strengthening method is commonly applied in the Netherlands to counteract settlement damage, limited investigations on the performance against seismic loading are available in the literature. Therefore, an experimental campaign (Licciardello et al., 2021) was conducted at Delft University of Technology in which a quasi-static cyclic in-plane test on a full scale wall was performed to characterize the performance of the bed-joint reinforcement technique. The wall featured artificially introduced cracks (pre-damage), achieved by the inclusion of plastic sheets between bricks and mortar, to account for the settlement-induced damage. Compared to the un-strengthened walls, tested in a previous experimental campaign (Korswagen et al., 2019) under similar conditions, it is observed that the bed-joint reinforcement technique can provide a significant increment in terms of displacement capacity and ductility of the wall but not in terms of the force capacity.

In this thesis, numerical simulations of both un-strengthened and strengthened walls from the experiments were performed using 2D-models and the nonlinear static analyses (monotonic and cyclic) were carried out in the finite element software DIANA. The objective of this research was to compare different numerical modeling approaches and material models to find the best suited one for simulating the in-plane seismic response of both un-strengthened and strengthened masonry walls. Moreover, the objective was also to extrapolate the experimental results to other wall configurations, which are not experimentally tested, to investigate the combined effect of the bed-joint reinforcement technique and the change in size and location of the window opening on the in-plane response of the wall (parametric study).

In the scope of this thesis, three numerical modeling approaches were investigated (Figure i). The bricks and mortar joints are modeled as one homogeneous continuum in the macro-model. On the other hand, the bricks and mortar joints are modeled separately for the continuous and detailed micro-model where interface elements are included at the brick-mortar bonds for the latter one. The discrete (simplified) micro-model was not investigated because the reinforcement bars cannot be connected to the mortar joints since they are substituted by zero-thickness interface elements. Moreover, the Discrete modeling approach of the reinforcement was used in order to simulate the pull out behavior of the bars. Cracks were modeled using the discrete cracking approach and the smeared cracking approach where the former one was used at the brick-mortar interfaces and the latter one was used for cracking in the mortar joints (micro-models) and in the masonry composite (macro-model)…

Cracks in masonry structures are a cause for concern as they signal a potential lack of functionality and/or aesthetics. It thus becomes important to identify the cause of damage in order to mitigate it and to prevent its occurrence in the future. Similarities in crack patterns may correlate to similarities in the damage cause. Currently, the assessment of similarities in crack patterns and their corresponding damage causes is done by masonry experts and structural engineers. This process is often expensive and subjective. The use of a Convolutional Neural Network (CNN) may offer an alternate robust and dependable means to automate the assessment of masonry crack patterns by processing their images.

The main research goal of this MSc thesis is to answer how accurately can the CNN -- fitted to data generated from finite element models -- estimate masonry crack pattern similarities. To develop a neural network that can perform such an automated assessment of masonry crack patterns with a high degree of accuracy, a large number of crack patterns with similarity ratings given by human experts are required. This data is collected in increasing complexity, first from a statistics-based approach by generating synthetic crack patterns from Markov walks. This is followed by a computational physics-based approach, such as the Finite Element Method (FEM), that generates crack patterns on 2D masonry façades subjected to differential settlements and out-of-plane loads. Finally, real-world data is also collected. This data is used to fit and test a convolutional neural network developed by Kleijn (Kleijn, 2022). Continuing along the previous line of research done at TNO (where 12 crack patterns were chosen and developed using the statistics-based approach), this thesis focuses on developing parametric finite element models of 8 out of these 12 Pattern IDs. Additionally, real-world images are also collected from Gouda in The Netherlands. This data is then used to form crack pattern image pairs that can be assessed for their similarities by 28 raters using three similarity label categories: crack pattern similarity label, damage severity label, and the overall similarity label. Using these labels, the raters assessed 2587 image pairs generated from the statistics-based approach, 500 image pairs from the computational physics-based approach, and 50 image pairs from the combination of images from the statistics-based approach, computational physics-based approach, and the real-world cases.

An inter-rater agreement analysis is performed on the similarity assessments using Krippendorff’s alpha measure. Additionally, the agreement of each rater with a chosen standard rater is studied using Lin’s Concordance Correlation Coefficient (CCC). Using Lin’s CCC, the intra-rater agreement is also assessed for the standard rater to see how consistent a rater is with their own annotations. These labelled image pairs are then used to fit and test the regression neural network to evaluate its accuracy in predicting the similarity labels. The neural network is also fitted to and tested with various combinations of labelled data to study its generalisability.

It is found that in all three sets of data, Krippendorff’s alpha is less than 0.80 for all the labels, which indicates an insufficient agreement among the raters. It is also seen that, in general, agreement among the raters increases with their experience level, i.e. the descending order of agreement within the rater group is: industry experts, PhD students, and MSc students. Studying the Lin’s CCC of each rater’s performance compared to that of the standard rater helps to choose the raters who can be considered as reliable as the standard rater. Additionally, the intra-rater agreement analysis of the chosen standard rater shows that the highest self-consistency (agreement) is achieved for the crack pattern similarity label, followed by the overall similarity label and finally the damage severity label, with corresponding Lin's CCC values of 0.96, 0.86 and 0.72, respectively.

The neural network is tasked to predict the similarity level in each similarity rating for each image pair in the test sample. The ground truth of this neural network is established by averaging the similarity ratings given to each image pair by multiple raters. It is found that the neural network is able to achieve a sufficiently high degree of accuracy when fitted to and tested with all the image pairs generated from the computational physics-based approach. The crack pattern similarity label, the damage severity label, and the overall similarity label achieve an accuracy of 87%, 82%, and 69%, respectively. However, the generalisability experiments on the neural network that consist of predicting the similarity of a type of crack pattern image pair that is not included in the fitting data set, show very poor performance with respect to the prediction accuracy of the similarity labels. When the neural network attempts to predict the similarity of Pattern ID or a façade geometry that it did not see in the fitting procedure, it predicts all three labels with an accuracy that varies from 40% to 50%. Additionally, the neural network is also fitted to images generated from the computational physics-based approach and then tested with a pool of image pairs generated from the statistics-based approach, computational physics-based approach, and real-world images. The average accuracy with which the three similarity labels are predicted is even lower, lying between 25% and 40%.

This MSc thesis concludes that the neural network fitted to data generated from the computational physics-based approach and assessed by all the raters is able to predict the crack pattern similarity label, the damage severity label and the overall similarity label with sufficiently high degrees of accuracy. However, the generalisability experiments on the neural network show very poor results. This indicates that in order to achieve a greater prediction accuracy, the neural network may need to be fitted to a considerably larger sample of crack patterns that covers all of the relevant situations. Furthermore, the substantial inter-rater variability in the labelling of crack pattern image pairs suggests that even an ideal neural network architecture may not be able to overcome the inconsistencies in the fitting data.