This website uses cookies primarily for visitor analytics. Certain pages will ask you to fill in contact details to receive additional information. On these pages you have the option of having the site log your details for future visits. Indicating you want the site to remember your details will place a cookie on your device. To view our full cookie policy, please click here. You can also view it at any time by going to our Contact Us page.

Pushing the boundaries of the digital domain

13 June 2016

The automotive industry is on the cusp of a digital revolution, with new ideas and improvements to existing techniques set to transform the way we develop cars, says Chris Pickering.

This digitisation is being driven from both ends of the product development process. Using virtual models has the potential to save huge amounts of time and money if it can reduce the number of physical prototypes required. 

At the other end of the spectrum, greater use of smart factories, additive manufacturing and the looming prospect of Industry 4.0 has seen the prototyping and production phases become increasingly digitised too. This is leading to a more integrated product development cycle, where designers and engineers are not just sharing the domain with each other, but with manufacturing specialists.

To a certain extent this is driven by technology. Virtual reality headsets and other advanced 3D visualisation techniques are now commonplace in design studios. By integrating this data with engineering models right from the start it opens up a dialogue that would traditionally be delayed until later in the process. This can even be extended to allow customer focus groups to ‘experience’ concepts in the virtual world.

For processor-hungry tasks such as CFD, the continuing rise of computing power is making it easier than ever before to run a simulation. Suddenly it has become practical to evaluate ideas that weren’t feasible to investigate in the past, as Rob Kaczmarek from CFD developer Convergent Science explains.

One European OEM told Convergent Science that its productivity had increased by ten times in the last two to three years since it started using the CONVERGE CFD code for combustion modelling. In the past, this company had a team of five engineers who could produce maybe a couple of hundred simulations a year. It now has two engineers dedicated to combustion modelling who can run several thousand cases a year, while their colleagues look into aftertreatment or other areas of the powertrain.

This is not just about number crunching ability, but working smarter. CONVERGE uses automatic mesh generation to remove the need for manual meshing. This has the potential to save hundreds of hours a year, Kaczmarek points out, as well as dramatically improving the consistency between meshes. 

With any form of simulation it’s all about finding a balance between accuracy and time-to-solution. Perhaps the biggest step forward that Convergent Science has made with CONVERGE is the development of Adaptive Mesh Refinement. Rather than using the base mesh throughout the simulation, it regenerates the grid at each individual time step – again governed by user-specified parameters that guarantee consistency. This means the deformation errors otherwise generated by stretching or skewing the base mesh can be eliminated. More to the point, it allows the mesh density to change in real time across the model – providing additional accuracy where it’s needed, such as turbulent zones or areas of intense chemical activity, but reducing the cell count (and hence the overall runtime) elsewhere.

By making a simulation faster, more accurate and more consistent it becomes practical to model a larger portion of the total system. It also becomes easier to link it with other tools, such as multi-physics models. On a more fundamental level, this also allows engineers to apply it in a more predictive role.

By making a simulation faster, more accurate and more consistent it becomes practical to model a larger portion of the total system

A lot of the time people use CFD as a forensic tool, but the objective is predictive CFD of full systems, comments Kaczmarek. You need a very consistent simulation that’s not too mesh-dependent to work in a truly predictive way. With CONVERGE the company is aiming to introduce a more predictive workflow – one where engineers have the confidence to take virtual prototyping much further before they invest in physical testing.

Virtual development also opens up new opportunities to look at human interaction. Driver-in-the-loop simulator testing, for instance, is increasingly regarded as a valid tool for vehicle dynamics development, precisely because it is has the capability to generate realistic responses from human beings in a safe, controlled environment.

Kia Cammaerts, founder of simulator specialist Ansible Motion, comments that one of the great benefits of simulator testing is repeatability. You know the exact grip of the virtual surface across every square millimetre of its area; you know exactly what the other vehicles will do. Getting rid of all the ‘noise’ makes it easier to examine small changes that might be swamped by variability in the real world.

As with CFD, the concept has been around for a long time but recent developments in hardware and software have led to a step change in realism. 

At Ansible Motion, much of the work centres on motion cuing and the engineers have developed a deep understanding of how the human vestibular system perceives movement. From a hardware perspective, it’s also about reducing mass and friction, so tiny cues can be fed to the motion platform with total accuracy. Combined with dozens of other factors, including ultra-low latency graphics, haptic feedback and high performance audio, this is helping to create a more realistic driving environment than ever before.

As well as traditional vehicle dynamics evaluations, simulator testing provides a tool to tackle some very modern challenges. For example, the ability to monitor a human driver in controlled conditions can be particularly useful when you’re looking at autonomous vehicle handover and ADAS systems.  

Foam cutouts and inflatable cars – both of which have been used for testing autonomous vehicles – only go so far. They may work with optical systems and perhaps proximity sensors, but microwave lasers require proper specular reflections in the right frequency range, so the sensor input has to be exactly the same as it would in the real world. 

As with all the other techniques that make up the digital revolution, the results can be fed right back into the development process, whether you’re looking at a revised ESC calibration or a new rear spoiler. And that in essence is what it’s all about – reliable, trackable, sharable data that allows engineers and designers to experiment with new ideas more freely. Ironically, it’s the digital world that’s putting human creativity firmly back in the driving seat.


Contact Details and Archive...

Print this page | E-mail this page