From Digital Building Models to Predictive Maintenance: Integrating Image-Based Degradation Monitoring of Earthen Materials

The construction industry is currently facing an unavoidable paradigm shift: the transition from a linear consumption model to a regenerative circular economy. In this scenario, excavated earth, historically downgraded to special waste, is reinterpreted as a structural resource of profound value. This thesis investigates the long-term durability and structural health monitoring of Shot-Earth 771, an advanced earth-based composite placed via pneumatic projection. Although SE exhibits behavioral affinities with low-strength concrete, the thermal fatigue behavior of engineered earth materials remains largely unknown, as its large-scale adoption is hindered by an intrinsic vulnerability to environmental weathering, such as water saturation, which halves its dry strength. To bridge the gap between materials science and built environment management, this thesis investigates the durability of the SE 771 mixture subjected to freeze-thaw cycles, while simultaneously proposing an innovative digital workflow for low-cost optical structural monitoring in a BIM environment applicable to the safeguarding of "non-strategic" building heritage. An extensive experimental campaign was conducted, subjecting SE specimens to a non-linear protocol of thermal cycles up to 24 iterations. Mechanical degradation was quantified through destructive compressive strength tests and correlated with surface spalling using a Computer Vision algorithm developed in MATLAB, allowing for the quantification of the physically degraded area and demonstrating that the primary criticality is not dictated by the succession of cycles, but rather by the initial water saturation. By correlating the percentage of surface material detachment with performance decay, a predictive law was generated to estimate the residual mechanical capacity of the material, avoiding invasive on-site testing. Finally, this law was integrated into an OpenBIM environment using BlenderBIM (Bonsai). Through this workflow, the physical specimen is thus elevated to a reactive Digital Twin, capable of autonomously updating its structural health status by being color-coded according to the standardized visual scales of the Swiss KUBA guidelines for road inspections of civil structures and tunnels. This study analyzes SE as a resilient alternative for sustainable architecture and redefines the very concept of maintenance, laying the groundwork for rapid, scalable, and low-cost monitoring, transforming inert matter into a communicative and digitally manageable entity, and promoting the synergy between the use of eco-sustainable materials and the digitalization of Facility Management.