Piston Bowl Design

In this short blog post, we want to focus on the design of piston bowls for diesel engines — from a geo­met­ri­cal point of view … as we usually do as Upfront CAD enthu­si­asts. In our recent pow­er­train-related projects, we have been repeat­edly faced with the same design constraint:

Creation of an optimal bowl shape while the com­pres­sion volume is a user input, and there­fore fixed.

OK, this is not that com­pli­cated, but the idea is to keep this volume auto­mat­i­cally fixed during design studies for each new design. This makes sure that only feasible designs are gen­er­ated, and allows engi­neers to explore hundreds of suitable designs quickly. Piston bowls are not the only appli­ca­tion where the engineer needs to consider con­straints like this, of course. So the fol­low­ing example is a more general task where geo­met­ri­cal quan­ti­ties such as a (section) area, an area ratio, a volume, a centroid, center of buoyancy etc. are fixed right from the begin­ning of the project.

An Example

The fol­low­ing screen­shot shows a typical piston bowl design in CAESES®. The surface of rev­o­lu­tion is created by means of a 2D contour that is basi­cally a set of 3 curves (bspline, ellipse, line). The left window shows the contour design, while the larger window shows the surface plus the infor­ma­tion about the current piston bowl volume. 

As said before, the com­pres­sion volume should be a user input, and we need to find a way to consider this in our geometry setup. In this demo example, we have 3 design vari­ables (blue-colored arrows in the next picture) that are con­nected to the vertices of the bspline curve. The user can change these vertices along the x‑direction, as well as the upper two vertices along the z‑direction: 

There is another free variable that gets con­trolled by an opti­miza­tion engine, indi­cated by the red arrow in the above picture. The opti­miza­tion engine has access to the current volume of the bowl that is defined by the contour. This is a simple cal­cu­la­tion command in CAESES®. As soon as the user changes the control vertices, the opti­miza­tion is trig­gered auto­mat­i­cally and sets the free variable in a way such that the com­pres­sion volume is met. This works in a fully auto­mated way without any user inter­ac­tion, simply as a result of the internal update process in CAESES®.

Ani­ma­tion

As you might have noticed already during our recent blog posts, we really love ani­ma­tions because they much better illus­trate what is really going on. We have put together an ani­ma­tion that shows how designs get created where the target com­pres­sion volume is met for each variant. Just click on the image below to take a look at the animation:

Shape Opti­miza­tion

As soon as the setup is ready, the auto­mated creation of design can­di­dates within explo­ration studies or formal opti­miza­tions is a natural next step. Our users from the auto­mo­tive industry typ­i­cally run large studies of such cases where the sim­u­la­tion for each design is included as well. Addi­tional con­straints can also be con­sid­ered such as the distance of the bowl to the wall etc. The final screen­shot shows the design viewer in CAESES® where the geome­tries of design variants can be compared in a side-by-side viewer. Just click on it again to watch a very short video.

More Infor­ma­tion

There is a nice article with plenty of ani­ma­tion that shows piston bowl design and opti­miza­tion using CAESES and CONVERGE.

Follow Us

Do you work on similar tasks? What is your expe­ri­ence with it? Feel free to share your thoughts on this blog post. If you are inter­ested in stories like this and in CFD-driven design opti­miza­tion, then sign up for our newslet­ter. Don’t worry, we won’t bother you with too many emails. Of course, you can unsub­scribe at any time :-)