determining distributed clamping pressure

[enginesrus]

I'm curious of the method used to measure and or check the distributed clamping pressure of something like head to block, or any other similar parts that are bolted together with or without a gasket that need to seal and maintain clamping pressure between the bolt locations? I'd just like to know the practice large auto manufactures use nowadays if any. I am familiar with the film.

 

[3DDave]

90% chance they use FEA, assuming they care. The reliable seals seem to use a thick enough gasket that the stiffness of the rest is essentially infinite and therefore the locations of the bolts doesn't matter. The unreliable (looking at you intake gaskets that were thin garbage that deformed and leaked) method is to just not care.

I often thought that the surfaces could be precision made with an inverse shape such that the contact pressure was exactly uniform. I made a straight edge like that; I formed it to the inverse of the shape of a beam under uniform loading supported at a single spot. This way I could push down on the middle and the paper underneath would see uniform pressure and not shift as I cut it with an X-acto. In comparison, a nominally flat straightedge pressed in the middle transfers only a point load and can easily pivot of the blade isn't perfectly parallel for the entire cut.

 

[Pete_K]

At a previous company we worked on a can filling machine. The product was supplied from a plenum. The plenum had a seal between it and the pistons that sucked it up (before depositing in the cans). The plenum was pushed up into the seal by multiple pneumatic cylinders. The existing machine leaked at the seal, so just like 3DDave said, we used FEA to generate a contact pressure plot. From this we could see an uneven distribution and eventually changed the locations of the cylinders to get better contact pressure distribution.

 

[richardl63]

I agree with the other posts, this is a perfect fit for FEA. I use it all the time for similar types of analysis.

Richard Lee, PE
Lee Engineering Inc

http://www.leeengineeringinc.com

 

[CWB1]

Standard practice to simulate a clamped joint is to use FEA to measure deflection and stress at various points. To test various joints there's all manner of ultrasonic, pressure, and other tests that've been commonly used longer than most of us have been alive.

Regarding head-block joints specifically, they are a critical joint still requiring quite a bit of engineering to efficiently get a solid seal despite repeated thermal, vibratory, and other loading.

 

[BUGGAR]

I have to go to the top of the head and point out Chevrolet's solution to the valve cover gasket with those little wings to spread out the load (point load, not pressure).

 

[robyengIT]

have a look here

https://www.eng-tips.com/viewthread.cfm?qid=442786

 

[enginesrus]

Just curious how well FEA calculations correspond with the real world when all is said and done? Are there follow ups to verify the FEA using old school physical methods of testing?

 

[SwinnyGG]

Are there follow ups to verify the FEA using physical methods of testing?

Yes.

I wouldn't necessarily call them 'old school'. Ultrasonic reflectometry is common now but wasn't 15 or 20 years ago.

 

[CWB1]

The accuracy of any computer simulation depends on the engineer. As FEA uses a 3d model to evaluate stress & deformation at many thousands of points simultaneously, its relatively easy to get a much more accurate analysis than simplifying classical equations as done in the roby's link above. The upside is that allows you to design much faster, much more efficiently while considering manufacturing and other variability. The downside is that as FEA is a bit more complicated than a classical equation, its simply more for the clueless or careless folks to screw up. FEA isnt a replacement for understanding classical mechanics/equations, its simply allows those of us who do to work faster, more accurately. That being said, FEA has basically displaced most simplified calcs/analysis as linked above except for noncritical parts.

Testing and validating FEA to test data is a judgement call like any other that depends on how critical the part is for function/safety and the confidence in the simulation. Companies today are doing less and less testing despite designs becoming ever more complicated simply bc testing is expensive vs running another iteration of a known, good simulation. Usually you'll see a lot of testing done to validate a new FEA simulation but over time as a company gains experience they can rely on the simulation more and test less.

 

[BrianPetersen]

Head gaskets that I've worked with have a number of features to compensate for slight surface imperfections and/or distortion.

- A "fire ring" stamped into the gasket material around the periphery of the cylinder. When clamped, the flattening action pushes the metal around a little, which should preload the gasket material against both surfaces a little more in that area, which is where the highest forces occur (due to cylinder pressure).

- The gaskets are coated with a soft material that is intended to fill the voids and surface imperfections (e.g. machining marks) in the mating surfaces.

- In some cases - Multi-layer gaskets accommodate for more distortion by doing both of the above several times over (in each layer of the gasket). If for some reason the head moves around differently from the block due to differing thermal expansion, the discrepancy is spread out over a few gasket layers instead of having to be fully accommodated over two surfaces (the top and bottom of the head gasket).

In addition, the features of the engine itself contribute.

I'm going by memory but I think the last motorcycle engine that I put together uses four 8mm studs around each cylinder which is 58mm bore. If the peak cylinder pressure is 100 bar (which is an educated guess) that's 26 kN of force (something over 5000 lbs, if you wish) trying to push the head up, and each bolt (assuming grade 10.9 - the recommended torque is about right for that) has a clamping load in that range - so it has 4 times more clamping load total between the 4 bolts that surround that cylinder. On that engine the studs extend well up through the head with nuts applied at the very top ... spreads out the clamping load better.

On those engines both the block and the head are aluminum ... same thermal expansion rate. Same coolant circulates through both. The holes in the head gasket are used to regulate the coolant flow distribution, and I'm sure someone has been playing with simulations in order to make sure the coolant flow distribution is right.

 

[dik]

They also have pressure sensitive films that change colour with pressure. Stuff can be clamped in place, and disassembled. Looking at the film, the pressure distribution if pretty well defined.

Dik