Raw Earth Buildings and Industry 4.0: An Overview of the Technology and Innovation of the MUD-MADE Project

Research on digital production technologies for the building sector, although several decades behind other sectors, is beginning to become more and more systematic. The use of natural materials such as raw earth makes the sustainability of such processes even more pronounced than current building solutions. Despite this, many limitations still prevent the use of digital technologies employing raw earth for construction from becoming current. The article investigates the state of research on the topic, identifying the reasons for current limitations. It also describes the MUD-MADE research project that aims to overcome these limitations and make the use of digitally fabricated raw earth components for the building sector a reality. This project proposes a novel artificial intelligence-supported workflow for designing raw earth building components produced with digital manufacturing technology. The workflow can support the designer in a multi-objective optimization involving different performances (e.g., thermal, structural, acoustic) by saving material and maintaining feasibility. The workflow exploits parametric design to set a predefined visual script able to support the user. Indeed, the predefined script will allow the user to design a building component by selecting (or creating) different possible external shapes and infill geometries. The designer can include information about the local material and the available technology to digitally manufacture the component directly in the predefined code. In addition, the predefined script sets the boundary conditions and priorities for the expected performances. Moreover, performance priorities are defined by the user based on the requirements of the component to be achieved. Finally, artificial intelligence, exploiting the artificial neural network (ANN) will support the designer by automatically identifying the optimal configuration among the possible combinations of parameters and generative algorithms.