Torsional forced response analysis of 4S-4CYL engine

[jeyaselvan]

I was trying to estimate the torsional response of the drive train comprising of a compresor driven by a four stroke four cylinder diesel engine compressor. The model is based on eigen value / vector and modal summation approach.

I have obtained the torsional - crank angle informations (pressure + inertia) from the engine supplier, which was based on measurements of PCP of the engine.

I now have a problem. When I estimated the harmonic coefficients (cosine +sine) components and plotted it against for speed range, the amplitude of torisonal exciation for the dominant order (2.0 in this case) shows a very different trend. I have tried with both Fourier series and transform based approaches in MATLAB. I get very closer values.

With this trend, I was not able to predict correctly the torsional response at a DOF (sy Crank pulley). I already have a TVA report carried out by the engine supplier

Can anyone look into my excitations amplitude and comment?

 

[ivymike]

I didn't see the phase information for each of those orders. I assume that what you've got there is the torque "seen" by a crank throw due to pressure+inertia, and you're applying it in an appropriately-phased manner to each of four throws. You'd probably ought to verify that you have your firing order right.

 

[jeyaselvan]

Thanks ivymike

I have now included the phase information as well. I do see different behaviour of amplitude and phase for the second order. Could you please comment?

Jeyaselvan

 

[jeyaselvan]

I am attaching the file again.

 

[jeyaselvan]

Attachment again

 

[jeyaselvan]

Ivymike

To add further to your queries.
a)You are right that the torque "seen" by a crank throw (cylinder -1 in this case) due to pressure+inertia
b)The firing order is 1-3-4-2

What we see is the Fourier decomposition of the excitation(pressure+inertia) in cylinder -1. I understand that in this case, the orders other than 2,4,6,.. (in this case 4S, 4CYL, inline) will only remain, rest of them gets cancelled.

 

[GregLocock]

Frankly your data looks very odd. Sorry about that.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jeyaselvan]

Greg..

You are right. That is my concern. I have tried to decompose the Torque-crank angle information into half order harmonics. On reconstruction I get better correlation as seen in the attachement. I am using this to validate a code I am developing for forced response analysis, unfortunately i doubt whether I have the right input data.

The torque-crank angle information is given by the engine manufacturer. He has measured PCP and estimated the torque(press+inertia) and has given to me. I am not sure whether there is something wrong in their data itself,since I am not an expert to comment on this.

My observtaion is that the inertia torque in the data (last graph)is dominant has only the second order term ???

Jeyaselvan

 

[ivymike]

fix the red line in the last file... the bit where the slope suddenly inverts has to be wrong. I see the same thing in the other lines (which I assume are just phase-shifted copies of the first). It's as though you have a negative sign multiplied in there somewhere...

pressure torque and inertia torque don't look too crazy separately... (assuming they're for all cylinders combined) Inertia load is VERY high compared to cylinder pressure. Must be very light load - the engine couldn't be producing much power with the pressure torque shown.

There's a single cylinder torque trace shown on slide 10 of this presentation, for comparison

http://www.agen.ufl.edu/~tburks/Presentations/ABE4171/Off-Road Chapter 6.pdf

there's a combined torque trace for a four cylinder engine in there as well.

 

[jeyaselvan]

Thank ivkmike for your feedback

As you suspected, I believe there is a problem with the phasing of the pressure induced torque and inertia torque.

I have plotted this at a speed of 1600RPM for cylinder 1.
In the first case, I plotted the as is information given by the engine guys. In the next case,I inverted the phase of the inertia torque. Now it looks like I am running into another trouble.

I expect that the combined pres+inertia torque behaviour would be a combination of both the cases. .i.e.the behaviour of the pressure zone (0-180deg) from the first case and the inertia zone(180-720) from the second case.

It looks like there is a phase (shift) problem in the pressure and inertia torque information given by the engine guys. Is it the case?

Jeyaselvan

 

[ivymike]

That still doesn't seem right. The original phasing of the inertia torque was probably right, if 0deg is TDC for that cylinder. The crankshaft has to put work "into" the piston mass from 0-90 deg, takes it out from 90-180, puts it in from 180-270, and takes it out from 270-360. The combined torque traces don't usually look disjointed though, which gets back to my comment about the relative magnitudes of pressure and torque.

Cylinder pressure should require torque input (taking work from the crank) leading up to 0deg, then become positive from 0-180, opposite & shifted vs what's shown.

why don't you just set up a spreadsheet and calculate the following:
- piston position, velocity, acceleration vs crank angle in 0.5deg increments, for a crank-slider setup similar to what you have
- crankshaft torque due to piston+pin+small end of rod acceleration at the same increments
- apply a "close enough" cylinder pressure trace and calculate torque due to pressure at the same increments
- compare to what you've been provided

That'd be about an hour's work, most likely, and would shed more light than 25 more posts in this thread.

 

[jeyaselvan]

Thanks Ivymike.

I have only received the pressure and inertia induced torque from the engine guys.I was trying to use this as input for my torsional response and encountered with the problem thereon.

As suggested by you, I tried to estimate myself the pressure and inertia induced torque based on the engine information known to me and the measured PCPs. Even then, I have the calulated values (as in the attachment) different from what the engine guys have given by me. My calculation shows that the inertia torque are a lot lesser than the pressure induced torque.

I still could not figure out the phase influence of the inertia & pressure induced torques.

Mr.Ivymike, can you please comment?

 

[ivymike]

your calculated values look reasonable.

the engine data you've received has some kind of problem with it. Look at the cylinder pressure trace. Is it possible that you've multiplied the sin component of your fourier series by -1? (reversing the pressure trace?)

 

[jeyaselvan]

I have found the problem now. The data what the engine guys gave me is the summed up pressure torque and inertia torque. I have incorrectly interpreted them for that of the individual cylinders.

Ivymike, you observation and my estimation that the inertia torque was much higher was the point, where I was close to finding out the problem.

Thanks again Ivymike

 

[ivymike]

heheh... well, glad you solved it.