Piston Bowl Design and Optimization

A piston bowl is a recess in the piston crown pri­mar­ily used in (direct injec­tion) Diesel engines, where it basi­cally forms the com­bus­tion chamber. The design and shape of the piston bowl influ­ences the movement of air and fuel during the com­pres­sion stroke, thereby affect­ing the air/​fuel mixture. A good mixture leads to a more effi­cient com­bus­tion, result­ing in more power or better fuel economy. Addi­tion­ally, using an effec­tive piston bowl shape can reduce the in-cylinder emis­sions (like NOx and soot) and the cost for after-treatment.

CAESES’ Piston Bowl Design Capabilities

CAESES® has been effec­tively used to design state-of-the-art piston bowls and brings along several key capa­bil­i­ties for this specific task. Here is a short summary:

• Arbi­trary profile para­me­ter­i­za­tions can be used. The design of your specific piston bowl is not limited to certain pre-defined tem­plates, but you are free to model your own custom para­me­ter­i­za­tion (or we set it up for you, fol­low­ing your spec­i­fi­ca­tions). You want to use linear and circular segments, or rather splines? It doesn’t matter.
• The bowl profile can be varied in cir­cum­fer­en­tial direc­tion, allowing for, e.g., ​“wavy” bowl shapes.
• Robust vari­a­tion of the bowl geometry is possible with no failed variants. As for other geome­tries, one of the most impor­tant targets of our software is 100% robust geometry vari­a­tion, obtained by smart para­me­ter­i­za­tion and depen­dency-based models.
• The com­pres­sion ratio can be auto­mat­i­cally adjusted for each geometry variant. This is crucial for making sure that every gen­er­ated variant has the same com­pres­sion ratio and not wasting time on infea­si­ble designs. It is done with an internal opti­miza­tion loop, where the vari­ables for the adjust­ment can be chosen freely. It is even possible to define an order of prece­dence, so that the auto­mated adjust­ment first tries to match the com­pres­sion ratio with the first given variable. If that doesn’t suffice, the next variable is added, and so on.
• Other auto­mated adjust­ments can be carried out as well, like adjust­ment of the spray angle in relation to the changing bowl shape.
• The geometry can be exported in several dif­fer­ent formats suitable for your CFD/​meshing tools. Many of the formats support patch naming, so that the down­stream tool can cor­rectly identify surface patches for the assign­ment of indi­vid­ual mesh settings or boundary conditions.
• A design study on the piston bowl geometry can be combined with an inves­ti­ga­tion of the injec­tion strategy, or other process para­me­ters (such as fuel com­po­si­tion, EGR amount, etc.). In our Software Con­nec­tor inter­face, we can para­me­ter­ize and modify any value that goes into the input files or scripts for the CFD solver.

Example Case: Piston Bowl Opti­miza­tion with CONVERGE

This is a simple opti­miza­tion study that was carried out to demon­strate the workflow using CAESES® and CONVERGE. The bowl profile con­sisted of two circular segments, two spline and several linear segments. Four para­me­ters were selected for the optimization:

Lip Entrain­ment 

This para­me­ter controls the size of the lip at the rim of the bowl and thereby the inner diameter of the bowl. Note that as an option, the area (i.e. volume for a 3D bowl) can be adjusted auto­mat­i­cally to meet a given value. This is shown below to give you an example of the depen­den­cies. For the visu­al­iza­tion of the remain­ing para­me­ters of this article, this option is not active so that the single para­me­ter effect is more obvious. 

Bowl Radius 

This para­me­ter controls the radius of the bottom of the bowl. 

Diameter Scaling

This para­me­ter controls the overall diameter of the bowl, by scaling the complete profile in outward direction. 

Spray Angle 

This is not a geometry para­me­ter, but rather a value that goes into the CONVERGE control files. 

CONVERGE was coupled to CAESES® using the Software Con­nec­tor. A sector of the geometry was cut accord­ing to the number of nozzles and exported in a special format for Converge CFD (“surface.dat”), includ­ing indi­vid­ual IDs for the iden­ti­fi­ca­tion of the dif­fer­ent patches. The opti­miza­tion was set up with a multi-objec­tive genetic algo­rithm that is part of CAESES®. Two con­cur­rent objec­tives were con­sid­ered: NOx and soot pro­duc­tion. The opti­miza­tion run com­prised about 50 designs and there­fore CFD simulations.

All design vari­ables had a sig­nif­i­cant effect on the objec­tives, but the two objec­tives were – not very sur­pris­ingly – anti-cor­re­lated. Three designs were chosen from the result­ing, still rather sparse, Pareto frontier: the one with the lowest soot, the one with the lowest NOx and a com­pro­mise from the middle of the frontier. The baseline was already pretty good in terms of soot, so, while big improve­ments were possible in terms of NOx pro­duc­tion, only the design with the lowest soot could improve upon the baseline.

For­tu­nately, that design also had lower NOx emis­sions so that an overall improve­ment would have been be possible by select­ing this design. Possibly, further improve­ments could have been found by running an addi­tional opti­miza­tion step, refining the search in the area around the best design so far.

Download Tech Brief

A short summary of the CAESES® piston bowl design capa­bil­i­ties can be found in this tech brief (PDF). 

More Infor­ma­tion

If you want to hear this story in more detail, you can check out the record­ing of the webinar we ran jointly with Con­ver­gent Science. If you want to get a first impres­sion of a para­met­ric piston bowl created in CAESES®, try the piston bowl web app. There is another related article about piston bowl design which shows a basic para­me­ter­i­za­tion, and a short case study from VW. If you would like to discuss your piston bowl appli­ca­tions with us, do not hesitate to get in touch! Finally, check out the auto­mo­tive industry section for similar applications.