Wednesday, July 05, 2017

How Computational Engineering Supports Better Building Quality

News of Dynamo's achievements are starting to spread widely, but that's not to say it can't be helped further. I'm normally a bit more focused on the production side of computational architecture but this use case for computational architecture by ARUP and Populous really highlighted some of the benefits of using this technology. Populous is a small firm with a growing international presence but in this project's style and trends were not the only factors driving the early design phase. 

The linked case study explores how ARUP provided Populous with 12 detailed structural options in 12 weeks for Australia's proposed National football stadium. Without the computational support the article states they would have only produced 3 or 4 options in the same time period. From the perspective of the client, this increase in design options represents a better search of the total solution space. This translates for the building design manager as an increase in building quality. The source of this extra value is in the application of iterative design techniques. One of the characteristics of this application worth pointing out is the detail contained in the options. Dynamo was as much responsible for increasing the number of options analyzed but also, importantly, increasing the detail. In this data-intensive age where econometrics is replacing economics, the increased level of model detail drove more accurate cost estimates, certainly for materials, as ARUP's specialty is structure, but also I assume construction costs as well, which depends heavily on the complexity of the geometry. 

A last point to note is where ARUP stopped using Dynamo in the project: At the building performance analysis stage. ARUP owns a software company and therefore probably has little need for other 3rd-party software to analyze their projects to the quality they want. Furthermore, to the best of my knowledge, Dynamo is not optimized for the types of calculations seen in structural analysis, nor its interface, giving 3rd-party products an advantage. That is beginning to change as Dynamo matures as a programming language. A team in Europe has spent a great deal of effort putting together an Dynamo-based structural optimization package called DynaShape. Considering its aim, the software is a bit complex to use. And I can't speak to the quality of its output either, that being a whole other complicated subject. But the code is open-source so those curious enough can find out for themselves. There are some videos included in the link as well which give a clearer idea about the capabilities of the package, though I wish they were longer. 

No comments: