Polymorphic uncertainty modeling ofa 3D printing process of concrete

点击翻译按钮获取中文摘要

Over the past few decades, the digitization in the construction industry has been progressed considerably. It has been integrated into processes ranging from design and planning to lifetime monitoring and maintenance of structures. The increased automation and rationalization save costs and time, while improving the workspace safety, additionally. In comparison to conventional, i.e. mostly subtractive or forming, production methods, additive manufacturing techniques enable a design-driven manufacturing process so that structures with a high degree of complexity, design freedom and functionality can be produced. Current research focus aims at developing 3D printing techniques, optimizing them and applying them in daily practice in order to obtain structures with well-defined properties and reliable behavior under different loads. For most materials and printing techniques it is common, that uncertainties dominate the processes. For instance, during 3D extrusion printing of concrete, the viscosity of the pumped fresh concrete may vary, affecting the geometry, the hardening and the final properties of the structure. It should be accepted that parameters during a process vary randomly, where however auto-correlations during the process are present. These correlations are not fully known, motivating the application of fuzzy-random variables to cover aleatoric and epistemic uncertainties simultaneously. Furthermore, some input parameters of a model representing a 3D printing process may exhibit cross-correlations with other parameters, which should be accounted for. Again, for the input parameters with vaguely known cross-correlation, fuzzy-random variables should be considered. Taking this together, the project results in a polymorphic uncertainty description of a transient finite element-based 3D printing process of concrete, where cross- and auto-correlated random processes are converted into cross-correlated random fields, namely the spatially and randomly varying material properties of the resulting structures. Having a precise uncertainty description included in the model, this can be further taken for reliability analysis of the process, where different failure modes are to be considered (strength-based failure, geometric requirements, buckling and layer interface strength).Finally, the model can be used to steer a 3D printing process optimally under a rigorous consideration of all types of uncertainties.