Sudden Drop of Pump Discharge Pressure

Vapor pressure?

You contradict yourself about running the pumps. First you say that, "When Pump A operating, Pump B stop and vice versa"

Then you say, "When they both running, they delivered almost same discharge pressure".

Finally, " And then we shutted Pump A, that's when the problem begun."

Do you use check valves?

Really, you won't get an answer until you ask an unambiguous question.

Do you use check valves?

Too true my friend...........pumps running in parallel should always have check valves on the discharge sides of all pumps that can be connected to that particular service.

It may also be of help to know what the pressure gauge readings are (if fitted)

......."grease overleakage"........

Is this to be taken as over lubrication?........high bearing temperatures.............and pumping highly flammable liquid ..........Oh! dear............me thinks, I don't think that I would like to be particularly near to this plant............it may suddenly go dark dark, and someone strikes a match to see what happened!!!!!!

Mr Mobi, Thanks for the reply

The discharge pressure gauge readings were :

• when Pump A ran alone, the disch press was 22 kg/cm2
• when they both ran (which was the moment when we wanted to switch pump), disch press of Pump B was 23 kg/cm2
• when we shutted Pump A down, disch press of Pump B suddenly dropped until 15 kg/cm2

The "grease overleakage" problem was not at the pump but at the motor. We only replaced the deflector to correct it.

Hi sir anrava,

I have looked through all the posts and I cannot recall anyone mentioning the checking of the check valve, on pump A to ensure that it is operating correctly. If you have a faulty check valve this may well cause the stated problem of the pressure drop with pump B.

When you changed over pumps did you shut the discharge valve of pump A before you shut the pump down?

The "grease overleakage" problem was not at the pump but at the motor

I had assumed that this is where the problem of over lubrication (if this was the problem) had occurred, at a motor bearing.

It looks as though the pressure/flow characteristic curves of the two pumps differ in some way. Pump B isn't doing a lot of work when pump A is running as well.

Clearly, pump B is not as effective as pump A at delivering fluid into the downstream piping system(s).

Is the direction of rotation of pump B correct?

Hi Mr MOBI, thanks for the reply.

1. Yes we havent checked it yet, but why should we check on pump A, not B?
2. When we changed over pumps, we shutted the discharge valve slowly. And then we slow down the turbine (Pump A is driven by turbine), and then we shutted it completely.
3. Yes maybe it was an overlubrication (I was not at site for a little while when it happened). But I think it didnt have anything that affected the big problem.

Hi Mr PWSlack. To answer your comments:
1. Pump B had run well 2 months ago in 7 days, then stopped for minor maintenance of electric motor. After the motor installed, we started it again and then the discharge pressure of Pump B went down. We didn't do anything with the pump. I think the direction was correct since we didn't move the pump and the motor turned on the right direction.
What are things could make the pressure/flow characteristic curves different? Maybe we can start analyzing with them.

Mr Lynlynch, thanks for your reply and sorry to confuse you.

The system requires 1 pump (and 1 spare pump).

• When Pump A run, Pump B stop
• When Pump B run, Pump A stop
• There's a moment when both of them run, which is when we switch it from A to B or B to A so that we dont lose fluids flowing to the system.

The problem was happened when we run Pump B alone. The discharge pressure was dropped suddenly when we totally shut Pump A.

Yes they have check valves at the discharge side. A simple schematic of the main pipeline is shown below. I'll try to draw the flushing line later.

This isn't a homework problem, is it?

it is a real case, and i dont even have pump at home =)

When both pumps are running, their combined head may be such as to place each pump far enough to the left on its respective curve to avoid cavitation. If so, both pumps can contribute to the total flow. But if one pump, especially the smaller, is running alone, the head it produces may be low enough to move its operation far enough to the right on its curve to result in cavitation--further resulting in low or unstable discharge pressure.

To determine if this is what is actually happening, more details of piping layout and pump curves are needed.

I understand that when you have a pump running and try and start the second pump, the second pump will not pump. Either one of your pumps will pump if the other pump is off. If this is not your condition then disregard the information below.

Are they both sucking from one tank? Perhaps when a pump is on there is a negative pressure in the resavoir that your are pumping from. Maybe you are loosing the prime in one pump when the other is on. Perhaps your plumbing causes one pump to suck air into the second pump if the second pump is off.

If they both are stopped can anyone of them be turned on and pump a head?

I think we are suffering from a language barrier.

Thanks for replying, Mr Captkirkconnel.

Yes they are pumping from one "tank", they were bottoms pumps of a refinery distillation column. We couldnt stop both of them because it would shutdown the unit.

Before we start the pump, first we filled the pump with cold fluids, and then hot fluids (which is its service). In my opinion it will eliminate any vapor pockets in the pump and also the pipeline.

@Mr Lynlynch , I'm sorry if my English were bad and confusing. I seldom use it after finished my education.

I believe that "unredundant" is correct. When you ran both pumps- which you stated that you normally do not do- pump b triggered a vapor lock because it was not moving any fluid- just turning and running against the head from pump A and the friction caused some of the volatile liquid to boil- but the existing head pressure would not let the pump clear the vaporized liquid into the discharge header.

When you shut off pump A, pump B had been compromised by the vapor that had built up and began to force fluid back up the intake and to occupy some of the interior part of the volute so the pump could not operate with a full impeller and generate full velocity lift on the fluid. That is why the pump looked OK when you opened it up later.

Pump B restarted (flooded) and was able to carry the full load again. If you had shut pump B down and then restarted pump B, it probably would have worked normally with full pressure and flow because it would also have been flooded before the restart.

Thanks for the reply Mr energygod

The simplify procedure of starting the pump were :

1. fill it with flushing oil (around 176 Fahrenheit / 80 degC)

2. slowly fill it with the fluid service (around 660 deg Fahrenheit / 350 degC) and wait until the temperature of both upper and lower case almost same.

3. if another pump is currently running (i.e. Pump A), open the discharge valve of Pump B slowly and start the pump.

4. Then shut the discharge valve of Pump A and shut the pump off.

So, "both pump running condition" is normal because we had to do it to switch the pumps.

and about :

"When you shut off pump A, pump B had been compromised by the vapor that had built up and began to force fluid back up the intake and to occupy some of the interior part of the volute so the pump could not operate with a full impeller and generate full velocity lift on the fluid. That is why the pump looked OK when you opened it up later."

Is that mean there is a chance the pump deliver normal discharge pressure when we wait until few minutes?

Mr Unredundant:

Thanks for replying. It may be Class 1 div 1, and fire was ever happened at the plant when mechanical seal leaked.

If you do not have non return valves at the delivery of each pump and if both are centrifugal pumps you may be shorting one pump through the other. There will be a pressure drop.

Bioramani

Hello Mr bioramani, thanks for replying

I am sorry I didnt mention at the first post about the type of the pump.
The pump both are centrifugal pumps with between bearings mounting.
There are check valves at both of discharge side of the pumps.

Hi Sir anrava,

I think that you are mixing between two different situations:

1-2*100%each stand-by Pumps,where one Pump is nunning and the other is in stand still.The duty point is the intersection of single Pump Q-H curve with the system curve.

2-Two Pumps running in parallel,where both Pumps are running at the same time to give more discharge at the same head.The duty point is the intersection of Q-H curve of both Pumps with the system curve.Take into consideration that in such a case you have to plot the Q-H curve of both Pumps running at the same time together with the system curve.

Which situation you are working with?

Also provide us with details about the piping.

Regards

Sayed Sarhan

Thanks for replying Mr sayed sarhan

It is the Situation #1. Only 1 pump required in the system.

about the piping, the simple one is above at my comment #17

I'll try to give the detail later

The pressure drops, but what about the flow rate,

When the first and second pumps are running in series you will have the sum of pressures pumps is running you will. When the first pump is shut down, that pressure output of that pump will drop

Do you have a pump curve on these pumps

p911

missed parallel, you need to post schematics of the hookups.

About the flow rate, it happened quickly (only few seconds because we afraid it would cause problem to mechanical seal and it's hazardous) so that we couldnt get an accurate data of Pump B's flow.

We run them parallel, not series. And normally only 1 pump runs. They both run when we want to switch the pumps (only in a little time)

below is the pump performance curve

First confirm that Pump B when running on its own, does produce a pressure of ~22kg/cm2 on the same circuit. (this was not clear from your description).

If this is the case: Then you either have a non return valve on pump A not closing properly (leaking) or completely missing, OR something like Unredundant suggested and confirmed by Energygod.

If pump B running single doe produce only 15kg/cm2 then that is expected.

When Pump B running on its own, it only produce 15 kg/cm2 vs 22 kg/cm2 expected (and maybe would drop less and less pressure because it only happened in a few seconds and then we turn Pump A on again for safety -- that was if the mechanical seal damaged and leaked, it would be very dangerous situation)

Around 1-2 months before the problem occured, Pump B ran well in 5-7 days, then we switched it to Pump A because Pump B needed a little maintenance at the driver (electric motor). The mysterious problem were when we tried to run Pump B after the maintenance, the discharge pressure of Pump B suddenly drop.

Thank you for your reply. (and by the way you have been very responsfull).

To try and help you I have few suggestions:

- If the electric motor of pump B was repaired: Was it rewound?

If Yes, then maybe you are not achieving the correct speed at the required load. The graph for the pump curve is not readable as submitted. What I mean is that it might be possible that the motor was not rewound properly and your efficiency and power factor are low, therfore giving a lower rpm (significantly the pressure is proportional to the square of the speed).

When checking pump B you seem to be afraid to run it long enough to check the rpm and pressure drop---> If the mechanical seal is leaking, you should be able to see the leak ar at least findout if any vapor from the fluid is coming out. YOU NEED to be able to run for enough time to check the speed and if the pressure drops gradually.---> then the electric motor might be a problem if the rpm drops as it is running---- At least dismantle the pump B and test it on a different fluid ...? Any leaking mechanical seal should be replaced{When repairing, if the pump was dismatled, care should be taken with the mechanical seal as it can easily be broken while removing the shaft --- depending on location within the pump etc.}

Thanks for replying.

1. It was not rewound, there's no major repair.

2. We were afraid because the seniors were traumatic of this pumps. Fire were ever happened in the past according to them. About the rpm, I'll try to check it later. In the past we didn't install scotchlight to reflect the rotation for the tachometer to read. We had installed it and check the motor speed with no load, it was 3015 rpm. Is it normal for a motor designed at 2969 rpm? Will the speed of electric motor drop when we load it with pump?

And about the performance curve, I uploaded it directly from the"Insert/edit image" facility in this forum and it became bad in resolution. I uploaded it again in external site :

[IMG]http://i46.tinypic.com/ffa58k.png[/IMG]

In reply, after the new information received:

1- What is the frequency of the power supply to the motor?

2- Is the motor directly coupled to the Pump?

A- If Directly coupled and 50Hz, then the motor cannot turn at 3015rpm! the Synchronous speed will be 3000 rpm and the nameplate of 2969rpm will be normal speed at full load of the motor (slip = 0.989 ? Acceptable for such a big motor...): no load, motor only, will have a speed between 2969 and 2999. If coupled to pump, it should not drop below 2969 otherwise it means overloading the electric motor!?

B- If directly coupled and 60Hz, then the synchronous speed will be 3600rpm and the nameplate speed of 2969rpm looks too low (slip = 0.824 ? maybe?). But the no load speed you said measured 3015 will be suspicious since it indicates that the motor was drawing ~90% of its full load power at no load(?!pump not coupled to motor.): more likely should be 3500~

(3600-3015)/(3600-2969) = 0.927 this is approximation just to show the proportion of the loading on the motor at no load ~90%.

Conclusion: I am not sure of what to say about the speeds submitted, specially the 3015rpm. Please reply on the 2 point above.

Third Point: Is the pump single stage or multistage centrifugal?

If multistages, the assembly must be checked for any excessive wear between the stages shaft to stage casing /clearances.

If single stage (high pressure for that...?), then the clearance between the casing and the impeller, and specially at the inlet EYE should be very small: check for any wear, rubbing or any missing spacer that could have been forgotten during re-assembly (if done).Because of the high pressure, the efficiency on pressure will be critical and dependent on the clearances!

I hope i was helpful here.

if the pump itself was opened and dismantled, you should check if the clearances between the Impeller and casing conforms with the one printed on the chart (i could not see if shown or just mentioned) . The clearance is very important for such a high pressure pump: Is it a single stage or, most probably multistage? In any case, if the motor is running within the rated Amps and the rpm issue is sorted out to be within the load range of the motor, AND if there are no leaks on the pumps mechanical seals, the last issue will be the Flow rate required by the process: The pressure outlet will be academic if the flow is achieved. NORmally, the pressure value for the process equipment will have a minimum value to achieve the required minimum flow.

Sorry.

In my reply i forgot to remove the last paragraph! Ignore it since it is repetitive.

Thanks

Ok, I forgot one more possibility:

---> Check the rotation direction of the Motor with the Pump direction of rotation! Maybe, on re-connection, there was a phase swapping and the motor is reversed. THIS will make the pump output drop and reduce the pressure dramatically. Please check.

Thanks for the Reply Mr LAA Lucke

To answer the question :

1. The Rated Voltage of electric motor : 3000 V, 3-phase, 50 Hz. We thought it couldnt be over 3000 that's why we checked it again with 2 infrared tachometers. Both of the tools show same number (3014-3015 rpm). We haven't check the speed yet when coupled with pump. We'll do it later after pump assembly finished.

2. Yes the motor is directly coupled to the pump (no gearbox).

3. The pump is single stage centrifugal closed-impeller. The maximum clearance of wearing ring is 1.22 mm and still in tolerance range. And about the clearance between casing and impeller, we didn't check it. Now the pump has been assemblied (only missed tighting of the bolts, we're still waiting for the hydratight to do the job. The internal parts (impeller, wearing ring, packing box bushing, etc) are in good condition (no scratches).

About the flowrate, it's in normal condition (current capacity of the system was 98%). And if the flow reached minimum value, there would be problem too with Pump A.

About the direction of motor, I have checked it when we do solo-run test of the motor. It was on the right direction (CCW from driver-side).

--

I'm now thinking about NPSHa. Could it be different between two pumps if they had different speeds? The datasheet tells that NPSHa = 7.4 m, while NPSHr (water) = 7 m at 3% drop. I'm thinking about cavitation.

I am still baffled:

1) The rpm of the motor should not be more than 3000 rpm. This is dependent on the Frequency 50Hz. The Motor is obviously One pair of poles and this is the minimum number of poles for a motor:

Asychronous motor, 1 pair of poles, 50Hz can only reach < 3000rpm. There will be a Slip necessary to produce the torque and this slip is from 0.0 to <1.0: Final Loaded speed is 3000 x slip (in the range 0.9 to 0.98 ?).

If the motor is Synchronous type, then also, the speed will be 3000 rpm and remains there (Synchronous). There could be a small variation but usually less than 3000rpm and not above Unless something is driving the motor!!? Therefore, this is beyond my perception of the system to have a speed above 3000 at no load. At load, the Synchronous motor keeps the speed at the Sync Spedd of 3000 rpm here (50Hz x 60sec = 3000 rpm)

2) Direction of rotation: OK ... but I must insist that you check the Pump's direction of rotation (usually an arrow on the casing will indicate, Otherwise, you need to look at the pump output opening and see how it is curved. ... needs some kind of drawing or I rely that you are familiar with centrifugal pumps and their cowl to confirm the direction (??) The Motor's direction of rotation can be changed with the wiring. It must rotate the pump's Impeller in the correct direction indicated on the casing (or if you can guess it by looking at the pump's casing and impeller type). That is after coupling: Take your time to check this carefully because many times people can get it wrong. I cannot tell you if CCW or CW from any point unless I can see the pump or have a drawing of the pump-motor assembly (Picture...).

3) Cavitation: This requires the following:

a) Is the pumping circuit a closed circuit? means that the same fluid is returning from the process back to the inlet of the pump? Or you are pumping from a reservoir which is at atmospheric pressure (the tank is open to the outside atmosphere?).

b) Whatever the case, if the inlet pressure is above 0 bar gauge pressure (not absolute pressure otherwise >1bar), then cavitation will occur only if the fluid temperature under such pressure can reach the low Vapor pressure: example, if water, at 70deg C and pressure < 0 bar (partial vacuum) some steam will be produced and cavitation can occur. Therefore, if ther is a back pressure (high tank or return pressure of the system is > 1bar, ) then cavitation does not occur. Put a pressure gauge just at the suction of the pump and measure the pressure. if you have an h,s diagram or chart for the fluid, check that you are still in the liquid phase (you need the temperature).

Normally, you can hear the noise of cavitation ( something like a shower noise indicating that some vapor is impinging on the impeller. I don't know how noisy the area is, but you can compare the noise with the other pump (a stetoscope can help).

In short: A pump is a pump. If it cannot produce the expected pressure, then: 1- the flow is too high (look at the curve) then overload of the motor can happen (draws more power.

2- There is a restriction at the Inlet side: Then you can cavitate since there will be a suction pressure < 0bar gauge preaaure and depends on the fluid/temp

3- The direction of rotation of the Impellers is wrong: then the performance curve of the pump is changed to produce lower pressure and flow. There will be no cavitation but lower performance. This can be checked also by looking at the powwer drawn by the motor and compare to previous or to the pump's curves chart you have for the pump.

Good Luck.

Mr LAA Lucke,

1. Theoretically yes it shouldnt be more than 3000 rpm, that's why we checked it with 2 infrared tachometer with different brands, but still they show 3015 rpm. Is there any possibility that the motor is not in good condition?

2. About the direction of rotation, because we didnt do anything with the pump after it ran for 5 days normally in the past, the pump shouldnt be turned on the wrong direction.

3.

a. No, the pump is a residue-of-a-column pump which pumps the residue to another plant.

b. We have pressure gauge at the suction and discharge. It was 1.2 kg/cm2. About the diagram of the fluid, this is what we're missing. The feed of the system has changed from the original design, but the process engineer said that the Specific Gravity is still in tolerance. I am suspicious with the changing of fluid properties. Is there any way to predict the pressure at the impeller eye? (because it would be the lowest pressure at the pump, wouldnt it?)

FYI I attached the operating condition of the pump

[IMG]http://i49.tinypic.com/2u633o0.png[/IMG]

About the noise, my colleague said there was a knocking sound when the problem occured.

About the piping, We tried to look inside as far as we could but because dont have tools to examine the piping completely. To shut down both of the pumps to check them all means plant shutdown.In the next 3-5 days we'll test the pump again. Do you have any suggestion what to check when the pump B is running?

Thanks

Sir Anrava,

1- On the rpm, if you are producing your own Electrical power, the frequency of the supply could be different from 50Hz. A genset could be wandering to about 51 or 51.2Hz and that will give a higher rpm: at 51.2Hz --> x60 = 3072 * 0.9815 =~3015rpm (since no load, the efficiency at 98.15% ...). Anything between 51 and 51.5Hz will give the result you are getting. If on the Mains Grid (national grid), then the frequency is more stable at 50Hz +/- 0.5Hz at worst. The speed will then remain below just 3000 rpm.

2- Direction of rotation: From the data you sent last, it should be CCW looking at the pump coupling side (to the motor). I hope that is correct.

3-a,b) If you are going to run the pump B soon, then Check and tabulate the pressures both at suction and delivery of the pump since you have indicators at both.Before pumping and after starting the pump and then after runing few minutes (if possible for the process to keep on...)

Also tabulate the rpm of the pump with the current readings (if your operating voltage is 3000V, then some CT and panel meter will be available maybe).

If the Inlet pressure is Above Gauge 1bar, then cavitation should be very difficult (unless fluid is very light and volatile...low vapour pressure. Any way, the noise to listen to is during runing and if there is any cavitation, then it will be continuously happening.

Mr LAA Lucke,

We have done running test of Pump B as follows :

1. Initial parameter :

Condition : Pump A run, Pump B stop

Pump A (turbine driven)

• Suction Pressure (Ps) = 1.6 kg/cm2
• Discharge Pressure (Pd) = 21 kg/cm2
  
• Speed : 2850 rpm
  

Pump B (motor driven)

• Ps = 1.6 kg/cm2
  
• Pd = 2 kg/cm2
  

2. Pump B start running

Condition : Pump A run, Pump B run

Pump A

• Ps = 1.6 kg/cm2
  
• Pd = 26 kg/cm2
  
• Speed : 2920 rpm (because both pumps run, the speed of turbine increased automatically due to decrease of its load)
  

Pump B

• Ps = 1.6 kg/cm2
  
• Pd = 26 kg/cm2
  
• Speed : 2996 rpm

3. Pump A slow down

Condition : Steam inlet of steam turbine pump A was shutted until low speed

Pump B

• Ps = 1.6 kg/cm2
  
• Pd = dropped until 11 kg/cm2 (after about 15 seconds it stayed at 11 kg/cm2)

About the sound, i'll upload it later.

We still don't know what the suspect is, is it possible check valve of Pump A is abnormal causing discharge from pump B flow through discharge pipe of pump A?

The last item you mention is quite possible. You may be able to check this by closing a manual valve to Pump A. If Pump B picks up pressure, it will be because Pump A's check valve was leaking backwards.

Thanks for the info supplied. There are still some data missing in the trial to determine if possible, the problem.

The "???" places to be filled.

1- If the non return valve on B is OK, you should have the Pdis = Psuc when B is OFF. The difference now showing = 0.4 kg/cm² is probably a gauge error and not significant.

2- The Rpm of B when running Alone is also important together with the Amps being taken when on Load. Please provide the Motor's Nameplate Volts and Amps. This will enable us to see if the motor is being overloaded or running too free, with a drop on the load.

3- If the rpm of the motor, running alone, is higher than 2999 then the supply frequency is needed.

4- You did not say if you were generating your own power supply and what is the frequency setting at no load on the generators (or at least whenthe pumpB is running).

5- The Psuc & Pdis of pump A when OFF. Are you still allowing some steam into the turbine when running B Alone? If yes, the turbine should be not turning the pump or at least very slowly (maybe less than 30rpm?) or completely stopped(?).

6- If there is any cavitation occurring on pumpB, then surely it could happen to pumpA. If not, then piping to pumpB is definitely obstructed at least partially.

This will show from the missing data not yet provided. If the pump B Impeller has been removed and replaced, not knowing the design of this impeller, I suggest you re-check if it was installed the right way round. Some pumps impeller design can be symmetrical and a mistake could happen: If the blades are curved, then it will greatly affect if put the wrong way. (This happened with centrifugal fans at one place near me…).

As suggested by you, the non return valve on A could be defective and not closing completely against a backflow on A. If so, then you will have a return flow cycling back to the B inlet ( this should show a higher pressure at A's outlet. Since some steam is still passing through A's turbine, it will probably resist the flow and show a pressure on A's inlet…!?). Also, if the flow resistance on B's branch before merging with the A's branch IS lower because there is an unexpected flow back, it will account in for the lower pressure. There are many possibilities to eliminate before eroing on the real problem. Not knowing or being unfamiliar with the pipework and the devices in line will not permit listing all the possibilities. I can only guide you by trying to eliminate the unlikely reasons:

a) If possible, you can run the B with a partially closed outlet valve to build up the pressure on the outlet and see if it can pressurise up to that (running together with A and producing an increase of pressure indicates that the pump is able to produce the required pressure. Otherwise the pressure of A would have remained 21 -22 kg/cm2)

b) If that is correct, the gathered data will also permit to show that there is a back-leak or not (the motor load will decrease instead of increasing). If the load increases or remains the same (on the motor… kW absorbed), then you have a by-pass flow due probably to an open route back to the tank or somewhere else, reducing the flow resistance while increasing the flow (H lower-à Higher flowà more kW on motor).

c) etc… you must use all the imagination possible with some knowledge of the pump and motor relations theories and practicals…

Awaiting more info and feedback. This is interesting and hope we get to the bottom of it.

LAA-lucke

As An After taught: I could not identify the purpose of some of the items in the drawing you sent(site): like the part drawn just above [Note 8].

IF THERE IS A KIND OF RELIEF VALVE INSTALLED ON THE PUMP OR JUST AFTER THE PUMP {this is possibly required on such an installation with pressures of 20bars + and big pumps handling flammable (possibly volatile...) liquids} Then check that the relief valve did not get STUCK open ! The flow can be redirected to the suction side of the pump, within the pump body.

Just a possibility that should not be overlooked in this case (You wrote that there was a sudden noise at the start of this problem but that was not solved as to what was the cause: It could be a non relief valve acting and getting stuck when both pumps run together and the pressure went temporarily higher (a peak?).

Please check.

Hello all, I'm sorry I've not responding in long time because I was unwell. About the data I'll complete it when we do another running test today.
2. The motor's nameplate volts and amps are 3000 V and 146 A (rated).
3. I havent checked for the supply frequency.
4. Yes we generating our own power supply here.
5. Yes the steam was still allowed enter the turbine when pump B running alone, and the discharge valve was still fully open. Later we'll try to shut the discharge valve first.
6. In my opinion, cavitation may happen on pump B but not on pump A because the speed of both pumps were different (the lower speed had lower NPSHR) and there was only slight difference between NPSHR & NPSHA on the datasheet.
The impeller hasnt been replaced so I think we can eliminate this probable cause.

about your suggestion, I'll try to bring up later after today's running test. but I'm still not very clear.
a) so the pump B is partially closed when running together with pump A or alone?

about the next running test, we'll try to eliminate 1 possibility. There's a chance lighter components have a rather big percentage in the fluids that enter the pump. So we'll try to reduce it and see the effect of it. If the problem still there, we'll gather again to coordinate another running test with another possibility.

Thanks for replying and it helps me a lot.

SIR anrav,

I am sorry about the table shown in my reply ~45. It was distorted because I copied it from spread sheet and it got shifted on the 2n row. I think you can understand it any way.

On the subject of trials: When pump B is running Alone, you can close partially(gradually) its discharge valve to see if the pressure is building up at the discharge point (where the pressure gauge is located): (please read my addition #48 in conjunction...)

1- If the pressure does not show an immediate increase, then there must be a backflow within the pump (Pressure relief valve stuck Open partially, if present!). Try this as much as possible with the conditions on site...

2- If the pressure start to increase, then:

*** Either the flow coming into the pump from the suction is too low to keep up with the discharge volume (Cavitation or restriction.... resulting in vapor formation, or vapor lock, etc) and by reducing the output flow we reduce the effect of the NPSH and suction cavitation etc..

*** Or there is a backflow through the delivery pipework, or the pump A via non return valve (stuck open partially also?).

Now, if the pressure gets better, you can keep the valve closed partially to the point where the pressure starts to increase and wait and see if it stabilise and start to go back up: Check that you do not exceed the 22kg/cm2 or the rated pressure. Also, check that there is a flow and it is not too low. It will be lower than expected anyway.

from the above, you should be able to decide what the problem is or in which direction to carry-On.

I hope we get somwhere from here and let me know.

Mr LAA Lucke, that's ok, I understand that. About the running test I mentioned before, it was cancelled because the management rejected it to be done. So we're now stuck and couldnt do anything with the pump until someone with power approved the running test. About the partially-closed discharge valve of Pump B that you mentioned, should we do this when the pump both running or after the pump A shutted down? And also when we're gonna switch the pump, is it better to shut the discharge of Pump A first then shutted the inlet steam valve, or to shut the inlet steam valve first to reduce the speed of pump A and then shut the discharge valve? Our Standard Operating Procedure mentioned the former should be done, but in these days the operators do the latter one. About the information that are missing, I'm sorry I didnt bring the documents home so that I cant give it to you now. I'm planning to upload the video when the pressure suddenly drop and also the "Notes" of the drawing that may be needed. Thanks for your suggestions until now.

Since there is a non return valve on Pump A discharge, And pump B will be running, You can shut the steam inlet gradually so that pump B takes over fully, then if required the Discharge valve of pump A can be shut (normally, The non return valve should be enough and no need to shut the discharge valve of the pump. This is an extra safety, and unless the non return valve is leaky, there should be no need except for maintenance purpose, to shut discharge valve of pump A.)... Now, if the Original designers of the system have given instructions to do differently, then maybe they have other reasons (to do with the fluid nature, temperature...risks).

When Shutting the steam first, the pump A output will reduce and when pumpB output pressure becomes higher, the non return valve on Pump A discharge line will start to close: The BACK PRESSURE from B will help shut the non return valve properly as A output pressure drops below B's.

WHAT I AM INTERESTED IN: The design of pump B (Which I presume is the same as pump A except for the electric motor against Steam Turbine)==> DOES IT HAVE A PRESSURE RELIEF VALVE? (Built into the housing/body of the pump or attached to it ?? This unit/device Will start to open and re-direct the flow from discharge to Inlet of the pump (or to the back where the pump seal is located...) when ever the pressure of the discharge on pump B is higher than the setting of this device. When this happens, the output pressure of the pump will drop. THIS IS A SAFETY DEVICE against Overpressure. On such big pumps with high pressures, they are usually included.

Now, If for some reason this relisf valve got stuck OPEN even partially, because of some mechanical failure inside or because of inadvertant tampering with the setting (MOST UNLIKELY), then the problem you are having will occur!

To test as per my previous reply: You have to do it with pump A OFF. If the Plant is running and the management does not want to stop the process, you will not be able to do it since it will involve reducing the output of B (which in any case is out of commission because of the problem...). You will need to do this test if there is an opportunity to run it when the output can be reduced or interrupted etc. BUT IF YOU CAN IDENTIFY the Pressure relief valve on the pump B, then maybe this can be opened and checked (Makers mechanical repair hand book? or maintenance recommendations on how to dismantle and check/repair or replace). There are springs involved and some kind of plunger to open or close an aperture. either the spring(s) are damaged, or the plunger is stuck in a position...

This as far as I can go at the moment. If you can give me some details : Pump Maker and the model number etc. maybe I can get some info from Internet like a description or design diagram...

Mr LAA Lucke, FINALLY We've managed to run Pump B alone! What we did was shut the discharge valve of Pump A until around 10 % open and together shut it completely while we shut the inlet steam valve of Pump A. There werent sudden pressure drop on Pump B. So, can we conclude that the check valve of Pump A leaks? (About the PSV, the pumps didnt have it).

Yes, The check vlve of pump A must have been leaking.

Now that pump B is running OK, then maybe you can service the check valve on A and see what has gone wrong in it.

Thanks for keeping us posted.

OK Sir, Thank u very much for the suggestions that you've given to me.

.........and you were told a number of times earlier in the piece to check the non-return or check valve............were you not.............and you obviously did not follow some of the procedures to check it..........anyhow I am pleased that you appear to have solved the problem.

Mr Mobi, i'm sorry for that, in that moment I was concentrating in Pump B only, that's my fault. And about the procedure to check, what were they?

Thanks for following this thread until now and the suggestions you've mentioned.

No problem.......everyone is here to help everyone else.

The outlet pressure of a centrifugal pump depends on inlet pressure, impeller size and RPM. The pressure of two pumps in parallel do not add up. When running both pumps, pump B with a presumably smaller impeller could not overcome pressure of pump A and just turned liquid on itself without circulation and became hot. When resuming pumping, the low vapor pressure caused cavitation.

Hi there, thanks for replying.

If that happened, what we should do to switch the pump from Pump A to B and vice versa?

This is all a waste of our time until the OP gives us competent, meaningful information about the pumps/system.

As a reference this is a small view of P&ID diagram..

[IMG]http://i45.tinypic.com/5mzn1i.png[/IMG]

From Drawing: The inlet to the pump B is through a 14" pipework reduced to 12" for the pump inlet. Also, I can see a Filter Or Strainer just before, with a (Note 16) on it. Then there is a check valve towards the main supply pipe (14") from the tank.

The outlet is 10" expanding to a 12" pipe, then a non-return valve, then a check valve. A branching to the left is not clear where to ( lack of any legends, I presume there is an orifice plate…).

The 2 items marked (11 PI 050) And (11 PI 147) will be Pressure Indicators at Discharge and Inlet respectively.

Here I can see that the Supply tank is there and there should be some feed-pressure if the tank is higher than the pumps. Normally, the design of the height of the tank should take into consideration the Vapour pressure of the fluid under the process temperatures involved, so that the suction of the pump does not get it to the Vapour pressure condition and cavitation (normal operation).

I cannot say anything about the direction of rotation from the diagram. This has to be checked on site with the pump and motor …

Suggestions: Check the Strainer/Filter at the suction. (Blocked or not properly fitted). Note 16…?

There is a branch connection just before the strainer: Check that it does not come from somewhere that can suck in Air?

Read the 2 Pressure Indicators: The suction pressure should not be below the Vapour Pressure of the fluid at the operating temperature. Otherwise it will mean that you definitely have a restriction somewhere on the suction line.

Look at the full drawing of the installation and check all the lines coming into pump B suction, that are not common with pump A. Somewhere there will be the problem:

-Restriction (partially closed valves or faulty valves…lose item inside etc.), Filters and strainers.

- Unexpected input of vapour or air, leaking from somewhere into the system and mixing with the fluid at the inlet line.

- Finally, recheck direction of rotation.

WHAT IS THE ITEM MARKED (11 XL 013) and note 8.?

two pumps in series will double the flow rate, two in parallel will double the head pressure. In both cases the pipe dia. and run of the suction, is very important.

H T H cheers

Alas, it is exactly the opposite. Series increases head (pressure); parallel increases flow.

(Just like electricity: series increases voltage; parallel increases current.)

Dear Mr. anrava,

From your posting I understand the following.

1. Pumps A and B are not RUNNING IN PARALLEL.

2. Only one Pump is operated at any time.

3. Pump B was running well for 7 days.

MY REASON FOR THE PUMP B NOT PERFORMING PROPERLY IS,

1. Pump "B" is STARTED BEFORE STOPPING PUMP "A" and PRIMING MIGHT NOT HAVE BEEN DONE CORRECTLY.

2. By chance Parallel operation means pressure difference at delivery will be observed since speeds are different, and Pump B will develop 8.598% more pressure than Pump A

(OR) Pump A will develop only 92.08% of the Pressure of Pump B. Hence proper Parallel Operation for Starting Pump B and then Stopping Pump A will not work. Did you start Pump B before stopping Pump A.?

3. Vapour Pressure issuue should be ruled out since Pump B worked well later as stated by you.

HENCE PRIMING ASPECT SHOULD BE LOOKED INTO. A SKETCH of the Installation would give more idea about the working aspect.

Thanks,

DHAYANANDHAN.S