Knowledge Model for the Parametric Design of Concrete Viaducts

Abstract

The Architecture, Engineering & Construction (AEC) industry has been traditionally fragmented, with a lot of disciplines specialising in different aspects of a project and encapsulating different knowledge areas. This research project, entitled “Knowledge Model for the Parametric Design of Concrete Viaducts”, investigates the development of a knowledge model around common, prefabricated concrete viaducts and its translation into an integrated, parametric platform and communication medium between architects and structural engineers. This model can act as a basis of discussion between the architect, the engineer, as well as the client and support a fast decision making process especially in the first steps of design.
The main fields of interest for this project are Building Information Modelling (BIM) and parametric/associative design. BIM has revolutionised the industry by providing the practitioners with concepts and tools towards an integrated, object-oriented process. However, the current BIM technologies are often characterised by limited design flexibility. Parametric/associative design promotes the idea of designing structures, as systems of parameters which can generate numerous alternatives. This paradigm facilitates the development of complex geometries, enhances automation and supports optimisation algorithms. In this research, the advantages and limitations of the aforementioned fields, as well as a number of other concepts, such as Object-Oriented Programming, Extreme Programming and Model-Based Systems Engineering are investigated and the results are employed in the development of the model.
The main parameters of viaduct design are determined and structured in a knowledge model considering their limitations and typical values. Moreover, the generic geometry of the components and their interaction is considered. This knowledge model starts with a strict top-down UML class diagram and gradually evolves into a parametric platform for viaduct design on Dynamo. Due to the complexity of the parametric script, a Python-based User Interface is created to increase the usability of the platform. The parametric model is seamlessly connected to Robot, a structural analysis software, through a Python script, in order to derive instant structural analysis results. The geometry of the deck is transferred to Robot, the structural analysis software, along with the properties of the beams, the support conditions and the loads of Load Model 1. A linear structural analysis is automatically conducted and the engineer can directly assess the results in Robot. The link between the applications ensures that both disciplines are working on the same design and merges their different knowledge areas into one model.
The number of parameters considered in this research is 82, and they are divided in 3 categories according to the level to which the designer can influence them on the platform. As an outcome of the standardisation process and the restriction of the scope of the research, 51.2% of those parameters require input from the designer, while in the case of a symmetrical design (the same components are used on both sides of the viaduct) this percentage drops to 36.6%. This underlines the potential of such an integrated platform, since automation can significantly reduce the design time and both architects and engineers can focus more on the creative parts of the design and doing more design iterations. Important issues for the development of the platform are the validation of the results, in order to gain the trust of the user, as well as the limitations of the current parametric platforms. The Recommendations for further research include the immediate changes that can be implemented to the algorithm, such as the integration of cost analysis and the increase of the level of detail, as well as ways to unlock the full potential of this platform by considering cloud-based solutions and taking advantage of cutting edge technologies like machine learning.