Generative Design  

Transition of CAD design to BIM was a revolution in the AEC industry. BIM increased the speed of the design, the efficiency and the most important issue, contain the whole project information. Nonetheless, according to Hesselgren, there is a differentiation about the use of computers in architecture. All the CADs and BIMs that were used till now, didn’t do else but just recorded our design decisions. The computer didn’t talk to us. And record decisions surely is not the only thing that computation can offer to us. Therefore, Generative Design is the right solution. Generative design is not about designing a building, explain Hesselgreen, it is about designing the system that design a building. In other words, we define rules and parameters, and ask him to generate the solution for us.

According to Abrishami, Generative design refers to any design practice where the designer uses a system, such as a computer program, to produce the solution to the design problem with some level of autonomy.

Generative Design proofed to be a very effective solution in several complex projects such as: National Stadium and National Aquatic Center for Beijing 2008 Olympics. In both projects, the engineering and construction design were aided by Generative Design, otherwise would be impossible to be done. For example, instead of depending from a genius designer to design all the seats in National Stadium, with all the parameters as distances, view, height; it was used Microstation to design all the 50.000 seats for us, with all the 3.000 parameters and also generate the view that each seat will have.

Construction is becoming increasingly complex and to attend those requirements the AEC industry was followed by many technological advances. According to Abrishami, architecture is not anymore about aesthetical emphasis, but it is now oriented to performance based architecture. Nowadays the major projects are designed based on simulations, structural analysis, energy analyses and environment issue. Therefore, all those parameters, rules or algorithms define the building.

Concluding, according to the writer Jim Paul, the engineer profession gained its importance just when the Catapult was discovered, and for the first time was possible to control a force that is superior of human force. Similarly, the importance of Generative Design is increasing because    it is possible to design and control things that are superior to the human mind, and could never exist before. (Bentley AEC Magazine)


– Sepehr Abrishami. Integration of bim and generative design to exploit aec conceptual design innovation.

– Changing the face of Architecture. BE-Current, Bentley AEC Megazine.

Architecture and Coding

Today, most of the architectural design is developed virtually, so architects should learn how to efficiently use the digital sources (software). Even though, at the moment we have several very useful architecture software, nonetheless architects sometimes need tools that the software in use do not offer, therefore they should know how to speak with computers. So, what is the solution?

Among several things that an architect should know by now, code seems to be the next step for several reasons. First of all, the use of software consumes the major time of the design process, instead of that, architects should know how to manipulate those software, so they would spend less time with software drawings and more time with creative phases.

Secondly, each project has specific needs and solutions. None of the current softwares, offers all the tools to solve those problems. With coding, we can create our tools for each project need and also the customize them. Moreover, when we design and we have to do repetitive tasks, it is much better to write a code for this and let the computer work for us.

Design by Coding: Parametric Wave Pavilion using Python Script (source:
Design by Coding: Parametric Wave Pavilion using Python Script (source:

Another reason to learn code is the complexity of the project that we deal nowadays. Several programs such as Grasshopper, Rhino, Dynamo, BIM software are used to develop complex geometries, but all of them has shown their limitations. When it comes to design a complex structure, with thousand of connections, where each connection has a variation, drawing manually would be extremely difficult if not impossible. Instead of that all we need is think,  make the rules and write the code.


– Dimcic, Milos. Structural Optimization of Grid Shells. 2011, PHD thesis.

– Kilkelly, Michael. 5 Reasons Architects Should Learn to Code.


Dynamo was developed by Autodesk and it is a visual programming plug-in for Revit and Vasari. It offers a solution for the non-programmer designers which needs to customize the Revit tools to their specific needs without writing a single code.

Firstly, Dynamo used DesignScript in an attempt to make it more graphical language, which led to the problems with Aish mentioned previously. Now, Dynamo just as Grasshopper, works with nodes and wires, where each node is a logical piece and it contain lines, Revit elements or mathematical functions. With Dynamo it is possible to directly access the Revit API, through Dynamo’s Pathon node, giving the user more flexibility to work with different methods, as well as to combine them, to achieve the final goal.

Dynamo demo's using a Stadium dataset (source:
Dynamo demo’s using a Stadium dataset (source:

 Moreover, the capacity to visually script, define nodes and script using other textual programming languages make it a very useful tool. Once that Revit is oriented to the extraction of information (BIM), therefore it is not such a powerful solution as a modeling tool, but now with Dynamo the process has completely changed,  expanding the limits of what Revit (BIM) now offers to us.

One of the main feature of Dynamo is the possibility to automate processes and to customize computational design. According to William Wang, frequently we find ourselves developing simple logical systems that link together space, geometry and function. For example, when we have a project where the size of the office determines the number of furniture loaded into the room, but also the type of door that is used. If this task is done manually, it will be a tedious work and prone to human error, because we should count manually the desks and also change the doors manually for each area. With Dynamo this task can be automated. Dynamo would count automatically the furniture loaded, measure the areas and determine the parameters of the door.


 – William Wang. Dynamo: More Than Grasshopper Lite. Jan 26, 2015. (source:

– Autodesk. Dynamo: VIsual Programming for Design


DesignScript is a design programming tool created by Robert Aish. Since Bentley Generative Components (GC) and McNeel Associates Rhino -Grasshopper had already their popularity, Autodesk needed a new product to compete with them, so they hired Robert Aish, which had before developed GC, and would best identify the market needs. The first objective of this tool was to create a textual language, but Autodesk insisted in a graphical language to be able to compete with Grasshopper.

DesignScript is a tool that proceeded with the improvements of associative and parametric modeling of other software. According to Aish, as an associative language, DesignScript maintains a graph of dependencies between variables which can be values or geometry. When DesignScript is running and we change a value using this variable graphic, it works as a propagate mechanism of changes.

As a design process, DesignScript is intended to be used by the designers that want to explore more alternatives and  performances to reach the best solutions.

According to Aish, DesignScript as a pedagogical tool is designed around the concept of a learning curve and supports a very gradual approach to learning programming. While in other software, with direct manipulation, associative or parametric modeling, the designer can obtain results after a little effort, in programming or scripting, a considerable time and effort spent can still result without much evidence of success.

Design script in order to be more flexible, aims to be:

– focused on the end-user; introduce concepts that facilitate to the users that are no accustomed to design with programming. This is achieved by permitting the user to use its logical framework in order to produce the design models.

-multi-paradigm; introduce different programming paradigms in a single language.

-host-independent; geometry models are generated  in different CAD application and also can access their different geometries, simulations library and create a correspondence between them.

– extensible; permits the users to add new tools and classes.

To conclude, now DesignScript is part of Dynamo and it is available both as graphical nodes and textual language. With this integration became possible to run project in Revit (BIM) or run standalone. It is very accessible to the new users that want to explore programming skills. For the non-programming users it offers a direct approach with graph node diagramming similar to grasshopper, which is simple and require no understanding of programming concepts. Therefore, the fact that DesignScript use a graphical node diagramming similar to Grasshopper bring again the same problem of complexity and impossibility to came back and change our product.


– AISH, Robert. DesignScript: origins, explanation, illustration.

– AECMEGAZINE. DesignScript. Published: 08.04.2012



Water Cube (source:
Water Cube (source:

The Beijing National Aquatic Center or as it is often referred as the ‘Water Cube’, was built for the 2008 Olympic Games. The project came as a result of an international competition in 2003 and the winning firms were PTW architect and the OVE Arup Consulting engineer in partnership with the China State Construction and Engineering Corporation (CSCEC).

The concept of the building is associated with water in its bubbly state and the square as the primal shape of the house in Chinese tradition. Once that in Beijing, the water source is scarce, a building that symbolize water has a very strong presence for the inhabitants. Moreover, the building works in duality with the complexity of structural elements and the simplicity of  the whole building form.

The Water Cube plan is a square with 177 m on each side, 31m height from the street-level and it has 17.000 seats (11.000 temporary and 6.000 permanent). It contains five pools and an organic shape restaurant. The building after the game served as Beijing’s premier diversion centers.

ETFE facade (source:
ETFE facade (source:

The bubble shape covers the building complex structure with 100.000 m2 of Ethylene-Tetrafluoroethylene (ETFE), a transparent envelope that weighs only 1% of glass weight. The result was 3000 irregular shapes of ETFE that rely on a complex structure of 22.000 stainless steel and 12.000 spherical steel nodes.

Sustainability issues (source:
Sustainability issues (source:

The building design was also concerned with the environmental issues. ETFE was chosen as the cover material because it permits the penetration of natural light and heat inside the building. Moreover, the solar rays that hit the building are 90% absorbed by the ETFE bubble and reused to heat the pools and interior areas. Another important issue is the fact that the roof of the Water Cube catches 80% of the water, which is reused and recycled for the building needs.

The project was developed in three different phases: the design competition, the design development and the preparation of tender documents.

In the beginning, in the phase of the competition, the model structure was built with scripts in Microstation VBA. Then, it was exported to another CAD platform for the purpose of visualization and rapid prototyping. For visualization, ARUP developed a virtual reality, which demanded to export files from Microstation to Rhino (via IGES file) and then from Rhino into 3dsmax, to generate AVI files, but this didn’t work well because it generates enormous files with 1.2 GB. Moreover, for the prototyping, these files were exported from 3d Microstation to STL file, and then to DXF for structural analysis (Strand 7.0).

Data transfers at the competition stage (source: BIM handbook)
Data transfers at the competition stage (source: BIM handbook)

In the phase of the design development, the Microstation model was exported to Strand 7.0 for structural analyzes and optimizations. Then, this model through VBA scripting was exported to AutoCAD (dwg) and Microstation Triforma (DGN) drawings, but also in Microsoft Excel (XLS) database.

Data transfers at the design development stage.(source: BIM Handbook 1)
Data transfers at the design development stage.(source: BIM Handbook 1)

In the phase of Tendering, the structural dimensions were achieved by selecting sections of the Wireframe analysis model, and importing them to Microstation. Then, with the VBA scripting were attributed section dimensions, reference number, and property information, for each structural element. Then, those elements were shaped by a Triforma structural elements, to achieve the correct model.

The extraction of information and structural data was facilitated by Bentley Structural, a BIM software. Bentley Structural created an enormous library of the steel elements with their sections and their correct dimensions. The extraction of data, with all the information related to dimension, materials, and other info, was effectively generated by Bentley Structural. There were needed 112 sections, well detailed to document all the structural information. The main benefits of Bentley Structural use were that it reduce the human errors in the project and also it has the capability to automatically update in other documents such as plans, sections and schedules, all the changes made in the project.

The fabrication of structural elements was made by on-site welding, with 3000 workers and over 100 welders. This is a unique practice for ARUP because usually it is made with pre-fabrication to avoid time loss and expensive cost, but that was not accepted by the client in China. As a result, it was needed to generate 15.000 drawings for this non-standard structure to be welded.

The Strand7 Finite Element model of the Water Cube. (source:
The Strand7 Finite Element model of the Water Cube. (source:

The biggest challenges of the Water Cube were structural design and fabrication. First of all, it was needed to optimize the overall dimensions of steel structure, but also needed to fulfill the design requirement for the seismic situation of Beijing. This process of optimization is followed by continuous changes to achieve the maximum benefits. According to Abdelmohsen, this optimization was constituted of 22.000 beams, which should meet the 13 Chinese steel codes strength equations at 5 points on each beam for 190 load combinations, those variables if we multiply will be 271.7 million design constraints to be optimized. For that reason, Arup used the Visual Basic 6.0 to write a software from the beginning, which would control all the optimization process. If this process would be done manually, with simple software, it would be extremely laborious and would demand months (if not years) to achieve the best structural optimization.

Then, for documentation purposes was needed a conversion program, from the analytical software to a CAD software, and it was done by ARUP, also via scripting. This scripting method was extremely useful for several reasons: 1. Generated the full documentation of the project (plans, sections, schedules, 3d Model), 2. Increased the modeling speed (the whole model was built in 25 minutes), and 3. Improved visualization by exporting the models to 3D AutoCAD (for the client and contractors).

One of the most important observations on this project is the use of scripting. According to Arup Senior 3D Modeler Stuart Bull, the ability to use VBA scripts to create geometry, which links the analysis and engineering model to the working 3d CAD model, was irreplaceable, because then with Bentley structural was possible to extrude all the 2D documentation and 3D model, for contractors and fabricators.

According to Eastman, the positive impacts of BIM (Bentley structural) are related to structural optimization, information exchange software, and interoperability. With BIM was possible to rebuild the model daily or weekly, based on the structural analyses and optimizations, to keep updated tender documents of the project, avoid human errors, but also resolve other issues as sustainability, fire protection, building performance.

To conclude, this complex and unique project would be impossible without the strong commencement of the team, constant accurate collaboration and a very important issue, that was the use of sophisticated digital tools.

- BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers and Contractors
- AEC MAGAZINE, source:

History of Building Information Modelling

Nowadays, we are used to hear that BIM is a new paradigm of designing, modeling and building, but this concept is not recent at all. In the earliest years of computing, in 1962, Douglas C. Englebart published the paper Augmenting Human Intellect, with his vision about the future of architecture and some strange ideas and concepts appeared, such as:

‘’the architect next begins to enter a series of specifications and data–a six-inch slab floor, twelve-inch concrete walls eight feet high within the excavation, and so on. When he has finished, the revised scene appears on the screen. A structure is taking shape. He examines it, adjusts it… These lists grow into an evermore-detailed, interlinked structure, which represents the maturing thought behind the actual design’’ (Douglas C. Englebart)

Then in 1975, a closest concept of BIM was first documented by a working prototype called ‘’Building Description System’’, and it was published at AIA Journal by Charles M.  Eastman, where also several concepts of BIM were mentioned, such as:

“. . . interactively defining elements . . . deriv[ing] sections, plans, isometricsor perspectives from the same description of elements . . . Any change of arrangement would have to be made only once for all future drawings to be updated. All drawings derived from the same arrangement of elements would automatically be consistent . . . any type of quantitative analysis could be coupled directly to the description . . . cost estimating or material quantities could be easily generated . . . providing a single integrated database for visual and quantitative analyses . . . automated building code checking in city hall or the architect’s offi ce. Contractors of large projects may fi nd this representation advantageous for scheduling and materials ordering.” (Eastman 1975)

Meanwhile, in the early 1980s, other parallel researches about BIM were conducted in Europe and in the USA. While in Europe, this concept was named as ‘’Product Information Models’’, in the USA was described as ‘’Building Product Models’’, therefore, those two nomenclatures merged later into ‘’Building Information Model’’.

The first use of the term ‘’Building Modeling’’, in the sense of BIM was in a paper by Robert Aish in 1986. This paper consisted of a case study where he applied Building Modeling System, illustrating arguments and concepts of the BIM that we know today. Therefore, it was a short leap to the term ‘’Building Information Model’’, which was first introduced in the paper ‘’Modelling Multiple Views on Buildings’’ by G.A. van Nederveenand F. Tolman, in 1992.

In parallel with the development of the concept and nomenclature, several softwares (Brics, US-based Bausch & Lomb modeling system, Rucaps) tended to introduce those changes into their functions, and most of them are dimly remembered today.

But the first true BIM software was Radar CH, that later became ArchiCAD and it was developed by Gabor Bojar in Hungary. Even though ArchiCAD was the world’s first BIM software, it was not so successful until the recent years, due to the computing limitations of the time and also the inconvenient business situation.

After the ArchiCAD release, the Parametric Technology Corporation (PTC) was established and their first constraint based parametric software, PRO/ENGINEER was released. Thereon, two co-workers of PTC, Irwin Jungreis and Leonid Raiz, which already owned the know-how of PRO/ENGINEER, split from PTC and established their own software company (Charles River Software in Cambridge).

Their goal was to create software that can deal with bigger and complex projects, than the ArchiCAD does. So, in 2000 they developed a program called Revit, which utilized a parametric change engine. In 2002, the company was sold to Autodesk, so the Revit began to be more promoted and more sophisticated.

Revit brought many innovations that revolutionized the market, such as: parametric families, construction phase control, schedules and visual programming environment. Moreover, Autodesk invested also in collaborative design, to increase the collaboration between large teams of architects, engineers and contractors. So by 2004, in Revit 6 they released the Revit structural and Revit mechanical to help the communication between engineers and architects.

Since we have different BIM software in the market, it was impossible to exchange data from different BIM platforms. So, to combat this problem, in 1995 was developed the International Foundation Class (IFC), which still continue adapting and improving the model exchange.

After some years of BIM development, in 2005, was hold the first industry-academic Conference on BIM (Laiserin, 2005), where a broad range of software-designers and vendors, as well as successful users where presented showing their achievements and results.

Reference: – Quirk, Vanessa. “A Brief History of BIM / Michael S. Bergin” 07 Dec 2012. ArchDaily. Accessed 10 Mar 2015. <;

-THE BIM HANDBOOK, Chuck Eastman, Paul Teicholz, Rafael Sacks, Kathleen Liston

– AUGMENTING HUMAN INTELLECT : A Conceptual Framework. October 1962. By D. C. Engelbart –