firing frequency = numebr cylinder-1 ??? Is this really firing? 1

[Sofistioelevib]

Good morning to everyone,
some strange spectra today during analisys.

DIESEL TCG 2020 V16 (caterpillar) connected directly on MARELLI generator
Speed: 1500 RPM

I found about 87Hz as exactly 7 times firing frequency (1/2 RPM) till 10 mm/s as below



Lot of time i found some problem with firing but frequency was related exactly to cylinders number but in this case this frequency is like CYLINDERS -1 frequency and this is no sense for me.
Anyone had some experience like this? What suggestions from you in order to inspect better this component and find the origin?

Regards

 

[GregLocock]

You get 2 answers because I can't find out if it is a 2 stroke or a 4 stroke!

4stroke

Nope, firing frequency for a V16 4 stroke at 1500 rpm is 200 hz, which doesn't solve your problem admittedly.

So you've got 3.5 order. Wow, that is odd. Also odd is the absence of ANY firing order vibration (and multiples).


2 stroke

Firing frequency for a V16 2 stroke at 1500 rpm is 100 hz.

So you've got 7th order. Wow, that is odd. Also odd is the absence of much firing order vibration (and multiples).

Discussion

You do seem to have a lot of integer multiples of 1/2 order possibly caused by cylinder to cylinder variation. I always find it hard to figure these things out from single speed data, we always sweep the engine speed.

So I'm going to suggest you have a strong 87 Hz resonance - maybe crank bending, torsional or bending of the block +generator, excited by cylinder to cylinder variation. Maybe the harmonic damper failed.

have you got equivalent data for a good engine? Is this no load or full load?

 

[Sofistioelevib]

Thank you Greglocock,
you are right... information about stroke is important.
This is a 4 stroke engine with 16 cylinder and machine was full load.
About resonance/bending is exactly what i suppose but i'd like (and i have a chance to do because i can stop machine at request ) to measure in some way (please suggest every approach you are going to use to measure this value).

Furthermore probably next week i'm going to take measurement on exactly same machine near this one so i can compare data.
Update this post with the new measurement.

Anyway every experiences with this trouble will be helpfull

 

[GregLocock]

Welcome to the world of experimental modal analysis! Simplest way with what you have already is to mount two accelerometers. Move one around and plot the ratio of the amplitude at 87 Hz to the reference fixed accelerometer. Hopefully you'll get a ) shape along the length of the unit. That isn't true modal anllysis but it will do as a start.

It might not be a vertical bending mode, it could be lateral or torsional.

 

[Sofistioelevib]

Update

Performed measurement on identical machine.
Same spectrum, same time domain, same amplitude this is an indication that 87Hz (7X firing, 3.5X RPM) is not really a faulty component but it is related to a work condition of the machine.
Anyway for me still to be unknown phenomenous and according with @GregLocock probably it is related to some resonant frequency (hor, vert or torsional bent).
Because this is a reciprocating machine i suppose torsional problems it is likely.
I know torsional vibrations is not simple to evaluate or monitor do you know techniques or some stuff to use in order to evaluate?

But most important... It is a usual calculation provided from motor manufacturer? In case 87Hz it is forcing frequency already calculated, simply i'm going to monitor 87Hz amplitude over time.

Furthermore, decreasing the speed the order (ORDER) frequency still present but with different Hz and this is not in line with resonant frequencies that still at same frequency.

 

[GregLocock]

As i said we do an rpm sweep, which makes it obvious if it is resonance or a forced response.

You can measure TVs with a TV analyser (surprise) - a toothed wheel on the nose of the crank, a pickup, and an FM demodulator, which turns your pulse train into a usable signal. An F to V box would work as well.

Here's a nice example of an rpm sweep, the horizontal lines are resonances. the x axis is seconds, it's a run down test

 

[Sofistioelevib]

Thank you @GregLocock