Illustration 109440625 © Zishan Liu | Dreamstime.com

Analysis

How parametric design can benefit construction

14 August 2020

The construction industry needs a change of mindset to understand – and benefit from – parametric design, writes Paul Mullett

If you type the words ‘parametric design’ into Google, you will be presented with images of beautiful geometric patterns, ornate products and complex structures. These organic, and sometimes otherworldly, images align with and reinforce our perception of a branch of design that provides us with an ability to experiment with, shape and manipulate complex geometric shapes to meet modern architectural and aesthetic challenges.

The approach has of course been in use for decades, rooted deep in the art and science of applied computational geometry, and used typically to develop designs for canopies, facades, long-span roof structures and in other scenarios where geometry, architecture and structure work together in synergy, and where geometric complexities require a computational approach.

The outputs are impressive and grace the covers of our journals, feature in our newsfeeds, and are a regular feature at evening industry presentations. The promise of a parametric revolution has been unequivocally stated.

Meanwhile, the construction industry has started to awaken, opening its eyes to the potential for technology to drag it kicking and screaming into the 21st century and to realise a level of efficiency, quality and delivery certainty that other industries now take for granted. At the core of this transformation are data and its symbiotic partner, Artificial Intelligence (AI).

Building Information Modelling (BIM) is the recognised vehicle for transformation, providing a trusted repository to store and pass data throughout the project lifecycle. Furthermore, by using parametric design methods BIM can open the digital floodgates to so-called computational design tools, automating processes and enabling the application of AI (including its subset machine learning) to identify patterns, develop solutions and find additional value using so-called “generative design”.

And beyond this lies the holy grail of the digital twin, integrating real-time data and computational tools to develop and visualise predictions of performance or behaviour.

Why do we continue to see such inspiring, specialist applications of parametric design, and yet have an apparent inability to apply it at a broad, fundamental level?– Paul Mullett, Robert Bird Group

Delayed implementation

But most in the industry will profess that despite more than a decade of practical implementation, BIM workflows are still:

  • overly dependent on manual effort,
  • disconnected from other design-cycle processes,
  • frustrated by repeated changes, and
  • subject to human error.

Despite an increase in the availability of computational design tools and accessible AI algorithms, generative design remains a distant dream for many practitioners. BIM therefore remains largely a tool for production and coordination, and although the floodgates to a bigger, better digital world are open, what we actually experience is an occasional, unsatisfying trickle of 1s and 0s.

So why has the parametric revolution not occurred? Why do we continue to see such inspiring, specialist applications of parametric design, and yet have an apparent inability to apply it at a broad, fundamental level?

This juxtaposition is created by a basic difference in mindset. Faced with complexity, we rise to meet the challenge, pushing technology and ourselves to achieve what was not previously possible. We are however paralysed by the mundane. Faced with a problem the same as that solved numerous times before, we are inclined to solve it the same way or make just small, incremental refinements. “Optimisation is the enemy of innovation,” so the saying goes.

A different approach

Parametric design is just a different way of approaching a problem. Complex geometry or challenging engineering are not prerequisites. Indeed, if we stop to consider the definition of parametric as “relating to or expressed in terms of a parameter or parameters”, this statement of the obvious helps to take the mystique out of the term. Just like Dorothy in The Wizard of Oz, if we peek behind the velvety curtain, we will discover something quite ordinary and unthreatening.

At its heart parametric design is about creating data relationships. Parameters enable us to put data in boxes and connect the boxes together. So, in the case of geometry, rather than drawing geometry as points and lines, we define parameters and relationships that define the way the geometry is built. This way we can modify the parameters, and the geometry will change automatically.

We can also do this with non-geometric or abstract parameters, storing the data and passing it more effectively through connected workflows. We are then free to operate on these parameters in our preferred environment, to undertake and visualise computational design using analytical techniques or even using AI algorithms.

Parametric design doesn’t need to be all-encompassing, it doesn’t have to flow through the whole design cycle nor engage with AI.– Paul Mullett, Robert Bird Group

Simplicity is key

But if all this still sounds a bit too difficult, then let’s start small and let’s be specific. The key here is simplicity. Parametric design doesn’t need to be all-encompassing, it doesn’t have to flow through the whole design cycle nor engage with AI. Ignore the hyperbole surrounding parametric design’s intellectual genesis and don’t be in awe of the architectural wonders that have placed it on an apparently unreachable pedestal.

So, start by looking for a task or design element that is simple, repetitive or subject to change. Find the parameters at the heart of the problem and, using the range of industry tools available, explore how to approach the task or problem differently to achieve either a better outcome or the same outcome faster.

Think simple but see the big picture, acknowledging that delivering value and achieving returns often requires a perspective that extends beyond the immediate, single task in hand. Accessibility is no longer a barrier to implementation. Visual scripting has made parametric design accessible for all, and the tools come built into our everyday core software. Training is also readily available online, either for free or at very low-cost.

Whether its Rhino and Grasshopper, Revit and Dynamo or pure Python code, the mindset is the same. An evolution in thinking will quickly reap rewards, build technical capabilities and confidence and, when adopted more widely, will drive a step change in efficiency, creativity, quality and responsiveness.

Remember parametric design is the foundation on which all other data-based approaches are built. By parametricising our design – either physically or abstractly – we are digging a proverbial channel to turn the digital trickle into a free-flowing river, and one we can ride again and again as our design evolves.

And for those who are still not convinced, there is an important footnote to this monologue. With acceleration in the uptake in industrialised construction and a drive towards greater componentisation, parametric design will soon be an essential part of the digital design toolkit. The centre of gravity of parametric design is about to dramatically shift: from the poster boy of modern, aesthetic architecture to the workhorse of a functioning, connected project lifecycle.

The parametric revolution is coming, but it begins with an evolution of mindset.

Paul Mullett is group engineering and technology director at Robert Bird Group.

Illustration: 109440625 © Zishan Liu | Dreamstime.com