Structural parametric design for variant studies in the preliminary design phase

Abstract

An organized process is necessary to ensure that the building is built on time and within budget. This does not always go well however and four major difficulties can arise. One of the current solutions is performing variant studies. Designing variants is costly however, which limits the amount of analysed variants. The focus is set on improving this process, which will indirectly affect the four named issues. The advantages of structural parametric design (SPD) have the potential to increase the amount and complexity of variants. However, the disadvantage of large structures is the uncountable amount of possible calculations and parameters. Current SPD-tools have the same functionality in general, but differ in key ways which influences the choice for a tool. These mostly focus on some form of optimisation. This research however, focuses on generating options. The connection between a visual programming software (Dynamo) and FEA software (RFEM) named the DRC is chosen to achieve this. Attempting every possibility for the support structure is difficult. A method is therefore proposed called the "translation by component method" (TC-method), which gives the ability to more easily modify parametric models. This made it possible to divide the problem over several smaller scripts, called variants. Each variant is able to produce models, called options, which provide information about the critical load combination, the type of failure and the unity checks for most eurocode demands. The structural engineer can learn from each variant and apply this knowledge on subsequent variants. The engineer’s creativity is thus used to limit the amount of parameters and calculations. This process directs the engineer towards a solution. Options provide information for the architect as well, namely: the geometry of the support structure, the optimized cross sections and the costs of the option. The architect can use this information to weigh options against each other to fulfil the demands of the SoR as well as possible. This process is named optioneering. After the first variant was created, nine variants and 35 options were generated, analysed and reported within two weeks. Optimized options were obtained within 15 to 30 minutes. By learning from previous variants, it was possible to find a support structure which barely modifies the aesthetics of the structure. The options obtained during this process mostly have unique advantages and disadvantages which can be compared by the architect. A recommendation and summary is given for these results. The applied method for the DRC enhanced the process of the variant study by increasing the amount and complexity of variants. More information can be obtained and by learning from previous variants more informed decisions can be made for subsequent variants by the structural engineer. The architect has more freedom to choose the placement of structural elements and information of costs is earlier available. This can be used to better implement the demands of the SoR. The architect can take costs earlier into account as well which can be used to better estimate the costs-appreciation ratio of a certain choice.