15 years of ICE simulation at CD-adapco

Alex Read, CD-adapco London

With the addition of ECFM-3Z – the universal combustion model for Internal Combustion Engine (ICE) calculations– to the existing wealth of ICE functionality in STAR-CD and es-ice, engine designers have never had it so good. But where did it all begin?

We decided to take a trip down memory lane, to tell a story that began over 15 years ago in New York and London. It continues with CD-adapco building on its position as world-wide industry leader, by continuing to deliver the functionality that enables engine designers to model tomorrow’s engines (be they SI or CI engines, even when operating in HCCI or CAI mode) today.

Gerald Schmidt (CD-adapco’s Director of American Customer Support and long-time ICE champion) tells us how it all began, “It
was back in February 1990 when we did our first consultancy for Renault. Because of the highly complex nature of in-cylinder flow and the difficulty in obtaining and understanding experimental data, we’ve always had a high demand for running ICE calculations. The Renault work was a steady-state, cold flow, intake port simulation, but by the following year we ran our first transient, moving mesh simulation for DaimlerChrysler. In those days all the mesh motion and connectivity changes had to be defined by hand in newxyz [by user coding]. It took between 12 to 16 weeks to complete the model set-up and running.”

Even at this early stage, CD-adapco’s business model of close collaboration with end users and knowledge transfer through ‘expert system’ tools was already in action through the development of PORT, the first generation of CD-adapco’s ICE specific tools.

Steve MacDonald (President and founder of CD-adapco) takes over, “ Along side our consultancy activities, we were always looking to increase the level of automation and enable our clients to run these calculations for themselves. With these goals in mind we developed PORT, which allowed users to mesh cylinders for steady-state calculations from a predefined template. The first version came out in 1993, and was followed by NPORT in 1996. Of course the beauty of this is that you know exactly what the user wants and

needs because when you’re running a consultancy, you are the user, and if you take a wrong turn, the next time you meet with the OEM [Original Equipment Manufacturer] they set you straight.

”With NPORT, users could build complex engine configurations including the 4-valve topology template that many still use today, but they were limited to steady-state calculations: to run transient required considerable user intervention and know-how. The engine designers’ dream of being able to automatically generate the complete mesh motion for an in-cylinder calculation came true in May 1998 when the first version of ProICE was released.

“ ProICE took our model set-up and running time from 12 to 16 weeks down to 2 or 3. It was a real step-change in ICE simulations”, says Gerald.

Of course the development didn’t stop there. In July 2002, ProICE became es-ice, and year-after-year it continues to improve.
Today, with es-ice’s trimmed cell technique, the user simply defines the geometry and its motion, and es-ice does the rest. The model set-up time for a fully transient, moving mesh calculation has gone from 3 or 4 months, to 1 or 2 days, and we continue to make it faster. Users can increase the complexity of their model by ‘copying and pasting’ geometry to go from a single to multiple cylinders, simplifying it by automatically generating the mesh for a steady, cold flow simulation just by specifying a Crank Angle, automatically generating a sector mesh for faster turnaround, and much more.

 

Of course, all of this only focuses on mesh generation and setup. “ Engine simulations remain one of the most challenging areas in
CFD”, says Professor David Gosman (co-founder and Vice President of Technology at CD-adapco). “The challenge is not only to model the wide variety of physical phenomena (turbulence, compressibility, heat transfer, nozzle flow and atomization, droplet dispersal, evaporation and wall impingement, mixing, ignition, combustion, pollutant emissions), but to do so robustly, efficiently and accurately in the extreme conditions found in the combustion chamber, whilst the mesh shape and structure are dynamically changing.”

This is where STAR-CD strides ahead of the competition. Capabilities that are essential for ICE calculations, notably movingmesh
facilities and physics modeling, which have long been in STAR-CD, are only just being developed in some other codes. For
example, the highly optimized geometry movement method of cell layer addition and removal (available in STAR-CD since 1990),
removes the requirement for cumbersome and slow re-meshing methods. Spray models have been available in STAR-CD since
1992, and combustion models for SI and CI engines have been available since 1995. This means that while others are still working
on developing these basic capabilities, STAR-CD is able to focus on delivering the latest state-of-art combustion models, such as ECFM- 3Z, in a solver that has been and continues to be the tried and tested solution for engine designers all over the world.

In addition to the above, the benefits of CD-adapco’s policy of close collaboration with industrial and academic partners is apparent: the implementation of ECFM-3Z, in partnership with Renault, and the sponsorship of flamelet-based soot modeling at Lund University are the latest in a series of development initiatives to ensure that STARCD has the requisite tools-of-the-trade for ICE simulations.

In the highly competitive and changing world of engine design, one certainty is the continued presence of STAR-CD, helping to drive the engines of tomorrow.