Putting the “Smart” into the Dar Smart Bridge

What if the bridges of tomorrow were sentient? What if they interacted with their users and learned from their own experience to improve the next generation. What if the bridges of tomorrow were able to build themselves using past experience and knowledge of their surroundings. What if a robot could arrive at a river and print its way across.

At Autodesk Research, we explore new ways of designing, making, and building things. We often work with our customers to prove out these new workflows. With Dar, an international consulting organization for the Architecture, Engineering, and Construction industry, we set out to explore a new workflow for creating civil infrastructure, using technologies more familiar to the manufacturing industry.

In this three-part series, I will share the design process, the materials and sensors we used, and the manufacturing process for successfully 3D printing our 5m smart bridge.

Part 2: Materials and Sensors

The Dar 5m bridge was printed from fiber-reinforced polymers. For large-scale printing, these polymers are supplied in granulate form, with chopped lengths of fiber captured within. This granulate material is typically dried in an industrial dryer, before being blown to an extruder where it is melted for extrusion.

Choosing the right material for this project was a difficult process. As an emerging technology, new suppliers and material grades are appearing regularly. Among many other criteria, we agreed it was critical to find a material that was recycled (and recyclable at the end of the bridge’s life) while strong enough to be used seriously in a piece of civil infrastructure. We posit that if recycled materials can be used for such high-performance applications, they can certainly be used for any other less-critical components.

We tested a variety of materials, and considered many parameters for each, before deciding to work with Mitsubishi Chemical Performance Polymers (MCPP), who were able to provide materials that balanced these competing factors.

MCPP helped us to select the ideal grade of material which would meet our challenging strength requirements. The material is composed of 70% PETG taken from post-industrial waste, and 30% glass fibers by weight.

As is the case with all materials we looked at, some data we need isn’t included in the Technical Data Sheet as standard. These polymer granulates are traditionally used in injection molding, which produces parts with relatively isotropic material properties, unlike 3D printed parts which are printed in layers. Given that ASTM and ISO standards do not yet exist for additive manufacturing, it follows that Technical Data Sheets are still tailored towards the more established technology of injection molding.

In the image below, a dog-bone specimen cut at 0° to the layer will be significantly stronger than one cut at 90° to the layer.

And so, we conducted a range of tests ourselves. This included tests for tensile, compressive, and shear strength in these directions, and a study on the impact of interlayer temperature on the tensile strength.  We know 3D printed parts almost always have lower tensile strength perpendicular to the layers, but this can be compounded by poor interlayer adhesion. Due to the print direction, our bridge relies heavily on these interlayer bonds, which is why it was imperative to understand the impact of interlayer temperature. We conducted a range of test prints where we measured the temperature of the previous layer immediately before the next layer was deposited, cut out dog-bone coupons, tested them in a lab, and used the results to infer the minimum and maximum temperature range the print must stay within to meet our strength requirements.

These tests would be futile if we weren’t able to keep the part within the temperature window we defined. We used a thermal imaging camera and sensors to monitor the temperature of every layer of the print. From this data, we can verify that each layer is within the temperature region we required, to give confidence the bridge will be as strong we intended. The intent is to formally publish the results of these tests soon.

Making a ‘Smart’ Bridge

As the owner of a piece of civil infrastructure, what information would you want to know? Perhaps you want to know how much foot traffic it sees in a day. Or perhaps you want to know the stress and wear that pedestrians and weather put on the bridge over the years. By embedding strain, temperature, and humidity sensors into the bridge, we are able to collect a wide range of data to give the bridge life, allowing it to report back on its health and usage statistics.

To integrate the sensors seamlessly into the design, we ran cables through the bridge branches. These cables terminated at predetermined sensor points where we installed strain gauges.

Meanwhile, we cut channels into the top deck, and embedded Fibre Bragg Grating sensors. These sensors look like plain glass fibres, but with grating etched at defined locations along their length. The technology is pretty interesting – you can find out more on the sensors we used here.

With all this data, how do we present it in a way that is meaningful and engaging to a user? That’s where the work of another of our Research teams comes in. Project Dasher is a visualisation tool that contextualises data from sensors monitoring real world performance into a virtual 3D environment. For the Dar bridges, the concept was demonstrated by overlaying strain values in real time on the CAD model.

More to come soon on the manufacturing process!

Peter Storey is a Research Engineer at Autodesk.