hi,
I have performed vibration experiment using FFT analyzer (make BnK) on thin cylinder which is a evaporator of refrigeration system (length = 2387mm, OD=505mm, ID=489 mm)to check it vibration response since screw compressor in directly mounted on it. speed of compressor is 2950 rpm (50Hz), it has male screw with 5 lobe connected to main motor (LMF=5 x 50 =250Hz), and female screw with 6 lobe (LMF = 5/6 x 250 = 208 Hz).
now the calculated freq of thin cylinder using eulers-bernoullis beam equation, it was observed that second freq of cyliner matches with motor freq. but in experimental graph it has been observed that the peak at 50Hz is less than peak 250Hz (max peak), but since it is near to resonance condition, max peak should be at 50Hz.
i have taken reading at three position in horizontal direction and 3 position in vertical direction on cylinder.(equally spaced but center is offseted by 150mm)
can any one tell me reason for such results

vibration_result.doc (862 Kb, 23 downloads)

Dear Vivek,

Namaskar.

I was unable to open the attachment. I would like to present this logic to you. Let us treat your equipment like a pump with a 5 vane impeller. It is common to see peak at vane pass frequency. Which is 5 times 50 Hz in your case.

It is my understanding that there seems to be nothing abnormal when your 1X is smaller than the VPF.

hi, riyaz
i m again attaching new pdf file.
actually my query is regarding calaculated natural frequency for cylinder whose second natural freq (48.52) matches with motor freq (2950 rpm/60=50Hz).
it means, near resoance condition. thus max peak must be at 50 Hz, but it is still at 250hz
why such result came?.

vibration_result.pdf (307 Kb, 16 downloads) vibration result

hi ron
u suggested to Try with a bump test on the cylinder to see where the natural frequencies are.actually we dont have impact testing equipment , is there any alternative to find natural frequency.

do you think, the actual natural freq of cylinder differs from calcualted value from eulers-bernoullis beam equation.
is it because of compressor weight. but one think i would like to tell you, FEA results gives same values for both conditions (free undamped as well with compressor load).
now, second frequency (48.5 Hz) from calculation matches with motor freq (50 HZ). but still it shows max peak at 250 Hz (1st Screw LMF), is it because of position of pick up i have choosen. what do you think.

Vivek,

This is my understanding of things, correct me if I am wrong.

Do not get confused between the electrical frequency and the frequency at which shaft turns. Electrical frequency in Mumbai is 50 Hz and in the US it is 60 Hz. When we use variable frequency drive we call it drive frequency. During equipment operation you record the RPM and that is what you see on the spectrum. Spectrum has the RPM / CPM / Hz on the X axis and amplitude on the Y axis. I am sure this will clear the confusion you have.

Let's talk little bit about natural frequency. Calculated natural frequency would change after equipment installation.

I do my natural frequency test using single channel method. I am sure you too can do it. i mount the sensor on the bearing frame and use ordinary a 4 X 4 for excitation. I look for the peak and it's corresponding frequency which would be the natural frequency.

I am sure this will give you some understanding.

hi, riyaz

actually 50Hz freq i am talking about is of motor freq, since motor rated speed is 2950 rpm with 2 pole, that gives 2950/60=49.16Hz(~50 Hz).
it is not related to electrical frequency.

i m very new to this field, so could you please explain about 4x4 excitation. is it related to impact testing.
we dont have impact testing equipment with us right now, is there any alternative to check nat. frequency.
only my concerned is that calculated value of natural freq. for thin cylider from beam (fix -fix type)formula as mention in attachment is concide with the motor running frequency. which say resonance.
u r saying after assmebly frequency get changed (i.e. increased ) thats why it is not case of resonance and peak at 50 Hz is smaller than peak at 250 Hz for male screw means safe condition?..

waiting for reply

thanks
vivek

It is a common term to use in the US when someone says hit it with a 4 X 4. It means use a piece of wood having a cross section of 4" X 4". This wood is used to get excitation because you do the natural frequency test when the equipment is not in operation. For structures shakers are used.

You record spectrum when your equipment is not in operation by attaching sensor on one side and hitting with 4 X 4 on the other side. Most of the spectrum would have peaks listed at the bottom. Look for the highest peak. That is your natural frequency.

Only natural frequency would tell you if you are in resonance condition.

If you have for example the calculated natural frequency of a pump at 3000 CPM. If this pump is connected to suction and discharge piping there is a possibility that the natural frequency can go higher.

Experts on this forum may correct me if I am wrong.

I am also new having done schooling in Mumbai in the year 1979!!!

Your predominant peak is going to be at the lobe pass frequency due to pressure pulsation and it's not unusual. You expect to have 1x the highest as your theoretical fn is close to motor running speed. If this is an evaporator I suspect simplification to thin wall cylinder could create some error plus you mentioned the compressor is mounted on and as such the system mass will be higher and fn of the system will go down. What about any piping connected to this evaporator? I think your model is oversimplified.

hi riyaz ,
hi alex,

riyaz i m looking forward to do 4 X 4 test, to get actual fn. And i will take 4 - 5 readings to make assure it give peak at same frequency, if it works then problem may solve.

alex, yes u r correct due to compressor weight the frequency as to reduced. but if this is the case we might got the peak some where less that 50 HZ,since while compressor motor approach from 0-50hz it will meet the actual one. but i have observed in all loading case readong it give same trend. no clear pitch between 0-50 Hz, then peak at 50 Hz, 100 Hz and max at 250 Hz further in harmonic with 250 Hz.

yes one thing is important to tell, it take 3 sec for compressor motor to reach rated speed that it remain constant (as it is constant speed variable flow type)
means there is no enough time to FFT analyser to collect samples thus no peak at concerned freq.

i thing ,we have to take new reading during starting of the compressor to see the change in behaviour.
in present readings we have choosen 6400 lines and 15 average setting to collect signals.

i m attaching jpeg of system to get an idea.
pipe line seen is a suction line bet. evaporator and comp

thanks riyaz and alex for nice suggetions.

riyaz happy to hear that u studied from mumbai.
which college

vivek

hi riyaz ,
hi alex,

riyaz i m looking forward to do 4 X 4 test, to get actual fn. And i will take 4 - 5 readings to make assure it give peak at same frequency, if it works then problem may solve.

alex, yes u r correct due to compressor weight the frequency as to reduced. but if this is the case we might got the peak some where less that 50 HZ,since while compressor motor approach from 0-50hz it will meet the actual one. but i have observed in all loading case readong it give same trend. no clear pitch between 0-50 Hz, then peak at 50 Hz, 100 Hz and max at 250 Hz further in harmonic with 250 Hz.

yes one thing is important to tell, it take 3 sec for compressor motor to reach rated speed that it remain constant (as it is constant speed variable flow type)
means there is no enough time to FFT analyser to collect samples thus no peak at concerned freq.

i thing ,we have to take new reading during starting of the compressor to see the change in behaviour.
in present readings we have choosen 6400 lines and 15 average setting to collect signals.

i m attaching jpeg of system to get an idea.
pipe line seen is a suction line bet. evaporator and comp

thanks riyaz and alex for nice suggetions.

riyaz happy to hear that u studied from mumbai.
which college

vivek