Abstract theoretical mathematical concepts might not be your first port of call when considering BIM and structural models. But there are benefits to doing the inverse of BIM — or, in other words, modelling the connected space that lies between components, writes Martyn Day
We all know BIM models have a tendency to get big, fast. The more detail that gets added, the slower manipulating the model becomes. There’s an argument that, for certain analyses and deeper insight into the performance of a building design, using something simultaneously more lightweight and more intelligent might be a better way to proceed.
Enter Topologic, a free, open-source tool that breaks down buildings into an external envelope and subdivisions of the enclosed space. Creating separate spaces and zones by using zero-thickness internal surfaces produces a model that is optimised for better understanding of building performance.
Topologic’s story starts with Dr Robert Aish, the ‘father’ of Bentley Systems Generative Components (GC) and Autodesk DesignScript. He was writing a paper on the application of non-manifold topology as a lightweight form for architectural modelling as far back as 2013.
Meanwhile, Dr Wassim Jabi of Cardiff University, who was researching parametric design thinking and its role in building performance simulation, took Aish’s research on board. Jabi applied the concept of non-manifold topology to his own design and energy performance simulation research, publishing papers on his findings in 2014 and 2015.
In 2015, Dr Jabi and Dr Aish (who at this time was visiting professor at the Bartlett School of Architecture at University College London) decided to collaborate. The goal of their research project was to investigate non-manifold topology further, by building a platform independent software library to be used with Grasshopper, Dynamo and Blender.
Their proposal was funded by a threeyear grant of over £300,000, awarded in 2016 by the Leverhulme Trust. In 2019, the first alpha version of Topologic was released. Once the project was complete, Aish left. Today, the code’s use in AEC practice is being championed and developed by Jabi.
Topologic is now an open-source software development kit and plug-in for visual data flow applications. Jabi claims it will assist architects in understanding their buildings from a holistic perspective — both as a physical assembly of components and as a logical, spatial and hierarchical system.
Topologic enables connections with analysis and simulation engines, such as EnergyPlus and OpenStudio and can be used to analyse the thermal performance of a building without the need for a huge and detailed BIM model. It can also be used to plot paths such as mapping fire egress routes, identifying the least disruptive route for a new service pipe, or computing the most congested location in a city layout.
Pretty much all BIM models are produced using walls, doors and windows – or in other words, building components. The reality is that, in analysis, models do not need this level of detail. What they do need is more knowledge of connections and interfaces. And while you will find plenty of adjacencies of spaces in a typical BIM model, the way they are modelled tends to create lapping co-linear lines between spaces and edges that don’t meet. The result can create issues for confused analysis tools. Topologic tries to close those loops.
“Topologic is used to basically model spaces, rather than actual elements,” Jabi explains. “We are replacing detailed geometry with smart topology and information, thus reducing a very heavy geometric model into lightweight geometry. We add smart, rigorous topology to designs, such as how things are connected, and then imbue it with a lot of additional information.”
That information is not just attributes, he continues. It can also travel with geometric operations or topological information. This makes a Topologic model extremely lightweight and extremely powerful.
“Topologic is based on the idea of nonmanifold topology, which allows you to model spaces and create internal subdivisions, like cells. If you can imagine a cube and if every point on the surface of that cube were sentient, they would see the world divided into two sets – the outside and the inside of the cube,” he says.
“Now imagine that condition being violated, where a point on the cube can actually see more than two sets. The outside can see other sets of points inside the cube. That situation is called non-manifold. So basically, when you have a geometric engine that supports non-manifold topology, you can have extremely powerful representations.”
A building, for example, can be seen as an outside envelope with interior cells. These interior cells can encompass other interior cells. Hierarchical embedding is possible, too. “Then you can start to think of your building, your design, as a set of interconnected entities, usually defined as space,” says Jabi.
Topologic can be used to analyse the thermal performance of a building without the need for a huge and detailed BIM model. It can also be used to plot paths such as mapping fire egress routes, identifying the least disruptive route for a new service pipe, or computing the most congested location in a city layout
Explaining the fundamental database underpinning Topologic, Jabi says: “Behind all of this is the idea of a ‘graph’. This is unlike BIM systems, as we don’t have to use ad hoc methods to add topological connections. As an example, a door in Revit should know what two rooms it separates, but that only happens in Revit if you checkmark it at a certain point. If you don’t do that checkmark, that door doesn’t know what rooms it separates. In Topologic, that is built into the DNA of the software — everything knows what it is connected to, and it’s automatic and part of the data structure.”
Obviously one of the key times in a project to do analysis and to make important decisions is at the concept phase. Here, the industry has seen many exciting applications come to market. Most of these tools are based on the concept of spaces that lie adjacent to one another, which is Topologic’s core starting point, too.
Jabi explained that synergies between products have already kickstarted some collaborative development work. Topologic, for example, has worked with Testfit, because there are compatibilities between the models that the two products create.
“Testfit creates these simple, blocky models, where everything is interconnected. Once we understood how their file format was organised, we created a reader for it, and we imported Testfit models into Topologic,” says Jabi.
“This meant we could analyse the heck out of them, as we understood exactly the walls that are between two units, the walls between the unit and the corridor, the walls between the unit, and the elevator shaft etcetera,” he says.
The Topologic team has also created rules for generating graphic models, producing Revit models via Dynamo, ready for design development. “We could identify external walls, internal walls etcetera, so we were able to apply the right thickness and materials. So Topologic can be used as a driver to ‘thicken’ into a BIM model.”
While products like Testfit can generate hundreds of models very quickly, the software has no understanding of whether designs are energy efficient or if they constitute ‘good’ architecture. Even though Topologic can help to enable analysis and drive them into Revit, we wondered if the process might go in the opposite direction, from Revit into Topologic for analysis?
“We have started with a BIM model, rather than create one from Topologic,” says Jabi. “We took a model of a building that was filled with rooms, but they were not connected into apartments. So it was impossible to get an idea of the rentable area and have data for analysis.”
By running the model through Topologic, a graph was created where all the ‘graph islands’ (apartments) could be identified. The graph immediately went back to Revit, with apartments correctly assigned and colour coded, and schedules needed for analysis were created.
“So we can import an unstructured BIM model, run it through Topologic’s intelligence and make it a little bit more structured,” says Jabi. “But, you know, bad modelling is not something we can magically solve in Topologic. What we are advocating is to start modelling in Topologic first, and then move to BIM. Don’t model in BIM and move back to Topologic — if anything, that is the worst case scenario. While we have to deal with it, obviously, that’s not what we recommend. Start building even in SketchUp, or Blender, or wherever you have lightweight things, and then imbue them with intelligence, imbue them with information, and do lots of analysis on those lightweight models. Once you’re done, and you know what you’re doing, it’s just a click away to go to Revit models, so that it becomes an output not an input for us.”
With the rise of open source, Topologic plays well with popular tools such as Blender BIM. Bruno Postle, a colleague of Jabi and a member of the OSArch community, has used Blender, Topologic and Blender BIM to create BIM models, for example.
The process starts in standard Blender, where the user makes a simple cube structure and adds in slice planes of zero thickness to form spaces. Then, with one click, this is sent to Topologic, which behind the scenes starts to create a building with all the topology and information needed. Based on Topologic’s output, the data can be thickened with IFC information for building an energy and structural model. These full IFC files can be imported into Blender BIM. If the model needs changing, you can go back to the simple model, drag edges and so on, and convert it one more time.
In many ways, Topologic reminds me of the finite element analysis (FEA) packages used in the mechanical CAD world. While product designers are modelling every component and part in an engine assembly, the analysis teams are not using these explicit 3D models, because FEA tools need simplified geometry and lots of data about forces, materials and temperatures. The core output is performance information, to find design flaws and limits to the performance envelope.
Topologic is a spatial representation, a framework where questions can be asked before detailed modelling continues. And, as is the case with all these rapid conceptual tools, which predominantly report back financial information, Topologic can quickly identify any financial downsides of rapid design suggestions. The fact it’s free also makes it excellent value! Sometimes, it pays to think in levels of abstraction.
The source code can be downloaded here.