Automotive Topology optimization of an engine's cylinder block

[Dimitris2201]

I am currently working on my diplomatic work at university, with the topic: "Topology optimization of mechanical components in a vehicle's internal combustion engine using finite element analysis." More specifically, I am focusing on the aluminum engine block of a V6 engine.

I designed the block in SolidWorks, and after running a static analysis and collecting the relevant data, I proceeded to perform topology optimization.

However, I’m facing an issue during the optimization step: although I have defined the parameters correctly (load, constraints, preserved regions, etc.), when I run the simulation, the resulting optimized geometry includes holes through the block in physically unrealistic or structurally critical regions. These holes should not appear, as they compromise the engine block's functionality and integrity.

Could anyone help me identify what I might be doing wrong? Is there a recommended way to restrict or guide the optimization to avoid such problematic results? Any advice on which parameters I should review or adjust to prevent material removal in critical areas?

As for the parameters I used:I applied 15.1 MPa pressure on the cylinder surfaces (representing combustion pressure).I also applied 20 MPa pressure on the contact surfaces with the cylinder heads (representing bolt/clamping force.In the Goals and Constraints section, I selected "Best Stiffness to Weight Ratio" with a mass constraint of 20%.

For Preserved Regions, I selected:the contact surfaces with the cylinderheads,the contact surface with the oil pan,and the internal surfaces at the front and rear ends of the block, since these were the areas where the largest undesired holes were appearing.

Under Thickness Control, I set a minimum thickness of 10 mm. For Symmetry Control, I selected Half Symmetry using the Right Plane. Finally, for the mesh, I chose a medium mesh density (roughly in the middle of the scale).Based on the parameters I mentioned above, I’m not sure which of them might be incorrect or misconfigured. I've tried several adjustments — for example, changing the minimum thickness by a few millimeters, as well as modifying the mass constraint percentage — but the resulting topology is always very similar and still contains undesired holes in critical areas.Is there anything in my setup that stands out as a possible cause of this behavior?I would really appreciate any insight on what I might need to change or refine in order to get a more realistic and functional result.

Thank you very much in advance!

 

[berkeman]

Welcome to PF.

Can you upload some images of the results of your simulations? Use the "Attach files" link below the Edit window to upload PDF or JPEG files with the images.

Also, are you sure that the holes would be detrimental? If you have specified some acceptable level of strain for the simulation, the hole regions may still be meeting those levels. If it seems hard to manufacture them, remember that some very strange and innovative engines have been fabricated using 3D printing of metal materials...

 

[berkeman]

If it seems hard to manufacture them, remember that some very strange and innovative engines have been fabricated using 3D printing of metal materials...

More info: https://www.prattwhitney.com/en/newsroom/news/2024/07/24/additive-manufacturing

 

[Baluncore]

These holes should not appear, as they compromise the engine block's functionality and integrity.

If a hole appears in a wall, then presumably it is not load bearing, and so could be covered by a thin sheet cover material with a gasket. That will reduce the weight and the cost.

Maybe you are trying to optimise too many parameters at one time. Getting the fundamentals right is more important than perfection.

After all, the world is a strange and interesting place. No two engines will be subjected to the same external influences. There must be a margin of safety in every parameter, which will greatly reduce the need for perfection in the optimisation.

If less than one in ten thousand engines do not fail in normal operation, you are making them all too strong and heavy.

 

[Dimitris2201]

If a hole appears in a wall, then presumably it is not load bearing, and so could be covered by a thin sheet cover material with a gasket. That will reduce the weight and the cost.

Maybe you are trying to optimise too many parameters at one time. Getting the fundamentals right is more important than perfection.

After all, the world is a strange and interesting place. No two engines will be subjected to the same external influences. There must be a margin of safety in every parameter, which will greatly reduce the need for perfection in the optimisation.

If less than one in ten thousand engines do not fail in normal operation, you are making them all too strong and heavy.

So it could be that I'm putting too many parameters together.?

 

[Baluncore]

So it could be that I'm putting too many parameters together.?

Yes, it could be.
As we don't see your model, we can only guess and consider the process in the most general terms.

 

[Dimitris2201]

Yes, it could be.
As we don't see your model, we can only guess and consider the process in the most general terms.

 

[Dimitris2201]

Welcome to PF.

Can you upload some images of the results of your simulations? Use the "Attach files" link below the Edit window to upload PDF or JPEG files with the images.

Also, are you sure that the holes would be detrimental? If you have specified some acceptable level of strain for the simulation, the hole regions may still be meeting those levels. If it seems hard to manufacture them, remember that some very strange and innovative engines have been fabricated using 3D printing of metal materials...

 

[Dimitris2201]

The points where it opens holes are not acceptable because I will have leaks.

 

[Baluncore]

The points where it opens holes are not acceptable because I will have leaks.

My mind reading ability is low today. Are the blue areas the holes?

If the initial block is cast, then to avoid cracks and pores in the metal, all parts of the casting must cool at a similar rate. That will decide the initial wall thickness. You must also allow for the cooling galleries around the piston sleeves. If you are considering stress alone, you will need to remove the extra material using a milling process, before line boring and planing the block surfaces.

You have allowed for peak cylinder pressure, and the pressure of the head against the block. But where are the cylinder head bolts threaded into the block? Tension in the head-bolts is countered by compressive forces in the block and head material, surrounding the head-bolt holes. Those head-bolt threads will be pulling up on the crank main-bearing caps through the bolts that retain the caps. There is that belt of volume, deep in the block, where all the threads meet, from above and below, to oppose the torque-generating, compressive forces in the connecting rods.

Consider an air-cooled engine, where the piston sleeve, with cooling fins, is all that opposes the tension in the surrounding external head bolts. There, the outer surface of the block has completely disappeared, along with the coolant galleries.

Engines are designed by the work experience of many engineers, who know what has worked, and fear what has failed in the past. It is unlikely that one's experience will correctly advise one, on where the emphasis must next be placed, in the trajectory of evolution.

I expect the cyclic optimisation aim, should be to reduce the mass of each component in turn. That will require a definition of all the forces in each part. The engine block will need to be subdivided for the purpose of analysis, then optimised in isolation.

 

[Dimitris2201]

My mind reading ability is low today. Are the blue areas the holes?

If the initial block is cast, then to avoid cracks and pores in the metal, all parts of the casting must cool at a similar rate. That will decide the initial wall thickness. You must also allow for the cooling galleries around the piston sleeves. If you are considering stress alone, you will need to remove the extra material using a milling process, before line boring and planing the block surfaces.

You have allowed for peak cylinder pressure, and the pressure of the head against the block. But where are the cylinder head bolts threaded into the block? Tension in the head-bolts is countered by compressive forces in the block and head material, surrounding the head-bolt holes. Those head-bolt threads will be pulling up on the crank main-bearing caps through the bolts that retain the caps. There is that belt of volume, deep in the block, where all the threads meet, from above and below, to oppose the torque-generating, compressive forces in the connecting rods.

Consider an air-cooled engine, where the piston sleeve, with cooling fins, is all that opposes the tension in the surrounding external head bolts. There, the outer surface of the block has completely disappeared, along with the coolant galleries.

Engines are designed by the work experience of many engineers, who know what has worked, and fear what has failed in the past. It is unlikely that one's experience will correctly advise one, on where the emphasis must next be placed, in the trajectory of evolution.

I expect the cyclic optimisation aim, should be to reduce the mass of each component in turn. That will require a definition of all the forces in each part. The engine block will need to be subdivided for the purpose of analysis, then optimised in isolation.

However, I don't understand what could be wrong and holes are opening. Is it easy to explain to me, please?

 

[Baluncore]

However, I don't understand what could be wrong and holes are opening. Is it easy to explain to me, please?

I must assume that your optimisation removes material when that material plays only a small part in distributing the forces, or storing elastic energy.

You need to set a minimum wall thickness constraint in your optimisation model.

The alternative is to allow it to become a skeleton, showing you the energy flow paths between any attachment points.

 

[Dimitris2201]

I must assume that your optimisation removes material when that material plays only a small part in distributing the forces, or storing elastic energy.

You need to set a minimum wall thickness constraint in your optimisation model.

The alternative is to allow it to become a skeleton, showing you the energy flow paths between any attachment points.

I have set the thickness control to 10mm. And my block is 19mm thick. What you said about the energy flow, how can I see it in Solidworks?

 

[Baluncore]

If you have set the minimum thickness to 10 mm, then why are holes still appearing in the wall?
Check the thickness control is being applied to the walls that are developing holes.

I do not use Solidworks.

 

[Dimitris2201]

If you have set the minimum thickness to 10 mm, then why are holes still appearing in the wall?
Check the thickness control is being applied to the walls that are developing holes.

I do not use Solidworks.

I understand, maybe that's the problem. Thank you very much.