Engine test engineer getting a bit overwhelmed by vibration analysis 1

[xander18]

Hello everyone,

A little bit of background: I'm a recent grad (mechie BS) with a decent amount of engine experience for my age and position. Formula SAE powertrain lead, and did a combustion engine research project. I'm familiar with a lot of engine basics (I've read a lot of tech papers, textbooks like Blair and Heywood, mostly self taught). Now I'm out in industry and doing engine testing on a prototype engine. Unconventional design, it's an opposed cylinder/opposed piston 2 stroke diesel. We're a small company and I wear a lot of hats but essentially my boss and I are some of the only folks in the company with any real world experience (a lot of theoretical/simulation guys) so we kind of act as a voice of reason.

Since it's a prototype engine we experience a lot of failures. I'm competent at a lot of mechanical failure analysis; cycle fatigue, excessive loading, heat and friction, etc. But I'm feeling a bit lost when it comes to vibrational failures. I'm using an NI Zonicbook and eZ-TOMAS software to monitor accelerometers on each head and the block, I wrote a LabVIEW Program to perform torsional vibration monitoring on the crank encoder, and of course we have both a high speed combustion analysis system and a low speed acquisition system. But I just haven't quite been able to reconcile all that to perform the kind of conclusive analysis I would like to.

I've been reading some textbooks and looking for tech papers. I found Harris' Shock and Vibration to be a good text, from the small amount of it I've read so far. I just don't quite know how to really unite the theory with my data. How do I know that a component is going into resonance? How does resonance propagate through an engine? Say the front accessory mounting plate is vibrating at the modal failure frequency of my conrod, will that vibration necessarily propagate through the engine block to the conrod? More generally, how do I trace the sources of various frequencies of my vibration in the externally mounted accelerometers to specific engine components? If I see a higher order vibration at 2-3x the amplitude of 1st order (engine speed), what does that tell me about the source of that vibration frequency?

I know these are really vague questions, if anyone has any experience with this I'd really appreciate even pointing me in the right direction. Also if the only way to get familiar with this is to really read a textbook or two cover to cover and/or take a class then I would definitely look into that.

 

[GregLocock]

I found Harris' Shock and Vibration to be a good text, from the small amount of it I've read so far.

It is


How do I know that a component is going into resonance? How does resonance propagate through an engine?


Modal analysis

Say the front accessory mounting plate is vibrating at the modal failure frequency of my conrod, will that vibration necessarily propagate through the engine block to the conrod?

Yes, possibly although that particular scenario seems unlikely to cause problems.

More generally, how do I trace the sources of various frequencies of my vibration in the externally mounted accelerometers to specific engine components?

You can't, directly.

If I see a higher order vibration at 2-3x the amplitude of 1st order (engine speed), what does that tell me about the source of that vibration frequency?

1st order shouldn't be present unless it is a single cylinder engine, that means you haven't balanced the engine. You should see strong spikes at firing frequency and multiples, and any inertial forces.

So, what you need to do is wander along to ISVR at Southhamption univeristy and give them a contract or go on their courses.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Lou Scannon]

Along with measuring you might also make sure your design is sound from a vibration perspective, to begin with. This the text we used in my day at university: Kinematics and Dynamics of Machines, George Martin
There is a chapter dedicated to different crank layouts, and how to analyze their vibration.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

 

[xander18]

Greg,

I'm in the US so that course probably is unfeasible for me. But I'll look into other seminars in the area.

Hemi,

Unfortunately I'm not involved in design. If I were I wouldn't be nearly so worried about vibration, it's a very unique and experimental engine layout that hasn't lent itself to smooth operation. At least at this level of optimization.

 

[Tmoose]

" opposed cylinder/opposed piston 2 stroke diesel."

So both pistons do reach TDC at the same time and fire simultaneously?
If so that would be a nice surge of torque once per rev.
Maybe something like this -

http://gasturbinespower.asmedigitalcollection.asme.org/data/journals/jetpez/26807/s_030102j.5.jpeg

It would be load sensitive I'd think.

Do you have info about the component failures to share? They might provide clues if viBraTioN is behind the problem.

In the 50s Both Coventry Climax and Studebaker underestimated a designed-in geometrical stress concentration (in the crankshaft and con rods respectively) of new engines. The failures resulted during fairly normal engine operation.

 

[xander18]

Tmoose, I considered it but the company is pretty serious about IP and I play it safe. Also, the more important aspect (for me) is becoming a better engineer and expanding my skillset. Not looking for anyone to do my homework for me

Yeah you've got the right idea. 4 pistons, 2 cylinders. The cylinders don't fire simultaneously but the pistons do move essentially symmetrically (hit TDC at the same time, slightly different strokes).

 

[GregLocock]

Have you built a dynamic model of the masses etc to estimate the no load vibration? This could be a pen and paper exercise.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[EHudson]

'Mechanical Vibrations' by Den Hartog has a pretty good piston engine section. The only test that comes to mind with any reference to balancing opposed pistons that I can think of is 'Engine Dynamics and Crankshaft Design' by Angle. Memory says that may have been a much different mechanism (lots of parallel rods) - the text is quite old.

After you have read a bit more about vibration you may wish to check out the opposed piston engines in L.J.K. Setright's 'Some Unusual Engines' to see if there is something to be learned from them.

As far as a study of torsional vibration goes 'Torsional Vibration' by the old B.I.C.E.R.A. group was recently re-printed (Cambridge U.)

 

[pontiacjack]

I'm needing to learn representative values of cyclic torsional accelerations (and jerk) that an inline four cylinder engine's crankshaft experiences, for a crankshaft modification I'm considering. Scanning a list of SAE papers showed quite a number of papers which should be helpful [but I've yet to obtain and read any of them]. Have you looked there?

 

[turbomotor]

When I have had to deal with possible torsional vibration issues, I contact Bill Brogdon at bmepinc.com

Bill can both lend insight to your analysis and educate you further. He is the type of engineer that will provide a highly beneficial interaction with a sharp young engineer interested in learning. Get your company to sign an NDA and get his help.

(By the way, Bill does not know that I am writing this response, and I assure you that I get no kick back from recommending him. But I have had a very good learning experience working with him.)

 

[turbomotor]

And I also agree with EHudson regarding Den Hartog.

 

[xander18]

Hello everyone, thanks for all the suggestions. We actually made some really good progress on the issue last week. I pulled a conrod, hung it, put accelerometers on it, and hit it with a hammer. Great fun and the simulation guy I was working with loves validating his sims with real world data. Very rewarding. Slowly making progress on this and I'm enjoying the engineering design process.

Ehudson, I'm amazed that you've been able to come up with this many relevant texts to an unusual engine design. I'll look into all of them.

Pontiacjack, I have but unfortunately my company doesn't have an SAE subscription. I will likely ask them to buy a handful of the more relevant ones for me. I lived for SAE papers while I was doing FSAE though, it's always my first stop on a project like this.

Turbomotor, thanks for the referral. I'll look into him and see if my company wants to reach out.

Thanks again!

 

[tbuelna]

Maybe you can clarify your situation a bit. In one post you stated this engine was an inline 4, but in another post you stated the engine was a 2 cyl, 4 piston configuration.

If the engine is an OP 2T with 4 pistons, that would mean it has a pair of 2 throw cranks. But if it is a 4 cyl OP engine, then it would likely use a pair of 4 throw cranks. The number of throws on each crank, their angular spacing, the combustion pressure forces produced in the engine, the structural properties of the crankshaft, etc. can all have a huge effect on the torsional vibration situation in the crankshaft. A short, stiff 2 throw crank would usually present less torsional vibration problems than a longer, less stiff 4 throw crank. However, a 2T OP diesel engine presents some unique problems due to the gear drive typically used to synchronize the opposing crankshafts.

 

[xander18]

Wait, I didn't say it was an inline four, did I? Not that I can in my posts. It's a 2 cylinder opposed piston engine, you got that right. But it has a single crankshaft and longg rods in tension to get to the outer pistons. So the piston is technically 3 throws but 2 of them are identical as 2 rods go to each outer piston. Those rods are the source of much of our troubles because the length and the small cross section lower the failure mode frequencies significantly compared to a traditional rod.

Sorry, I'm trying to walk a line between not divulging company IP and being clear.

 

[tbuelna]

xander18- Apologies to you. The post about a 4 cyl engine was not yours.

But as you note, the long slender tension rods do present a unique situation with regards to structural vibration modes.

Out of curiosity, have you found that these long, slender rods are always loaded in tension throughout a full cycle, and at every operating condition?

 

[xander18]

tbuelna you're a clever man because you've hit on a key concern. Theoretically the rods are always in tension, at least under normal operating conditions. But given a misfire or loss of combustion the forces go into compression. One of those pesky real world concerns that are easy and tempting to ignore for the sake of design elegance.

I have to say, despite the learning curve problems like this really get me going. I love a new challenge like this and after working through it I just feel that much more competent as an engineer. Another 'tool' in my mental toolbox, as it were.