Fire Protection Jockey pump shaft failure Issue

Fire Protection Jockey pump shaft failure Issue

Posted by Mohan Dudani on 20 Dec 2011 at Pump Engineers group at LinkedIn

We have a fire Protection Jockey Pump which operates continuously. It is driven by 60hp / 3600 rpm motor. It is rated for 150 gpm and 154 psi pressure. It maintains system pressure of 154 psi all the time. It takes water from a 30” supply line and keeps fire main loop pressurized. This pump shaft has broken 7 times in last two years. Requesting any comments, suggestions to solve this issue. There are two more electric motor driven pumps in the same system one 75 hp , 1000gpm 100psi and other 125hp , 1000gpm , 120psi which takes water from 300000 gallons storage tank and pressurizes the fire loop but these pumps do not run continuously . They operate only when the system pressure drops below a set point.
Jockey pump has a relief valve at the discharge end it will open up if the water pressure inside the pump increases more than 154 psi and will release the water to the suction of the pump so basically it allows water recirculation.
Can you suggest a solution for this issue?

There are three diesel engine driven pumps in this system, which kicks in at set pressure drop in the system, 90 psi, 80 psi and 70 psi

Mohan Dudani , P.Eng , CMRP • The pump in question is Horizontal overhung centerline mounted, Single stage mounted on a base plate and flexibly coupled to a motor.
Impeller is mounted on the tapered portion of the shaft with a key and tightened with a nut.
The shaft breaks always at the same place, on impeller end bearing seat. The step of the shaft at which it breaks is 1.25” diameter.
Originally the shaft was made of SS 316, it was changed to AI
SI 4140 which was again changed to AISI 4340. Now we have made a spare shaft of Triloy (it is comparable in the chemistry to AISI 4340, but is of high strength and hardness) we have not yet used this shaft.
Original design had sleeve at a step after the step at which this shaft breaks. The diameter of the shaft at which this sleeve was 1” We have eliminated the sleeve and increased the diameter to 1.25” to make shaft more strong .
The pump operates continuously. Its bearings does not have temperature monitoring system installed.
We did not record as left alignment. Now it is in my to do list to develop and implement it including procedures for alignment and its tolerances, standard clearances , fits and runouts ,pump repair checklist , bearing replacement procedure , pump repair procedure , pump internals inspection procedure , soft foot correction procedure , etc etc .
Check valve was replaced sometime back.
The frequency of vibration monitoring has been changed from monthly to weekly.
The shaft failure looks like fatigue rupture. The broken section had “old” section at the outer edge and rest of the section was bright so it looks like it took some time to break. In short it was not immediate complete break

SL Abhyankar • Dear Mr. Mohan Dudhani,

Do the jockey pump run in parallel with other pumps, especially when the other pumps come in ? Does the failure happen when the pump runs in parallel ?

I am making this wild guess, because there have been cases of shaft-failures, especially when dissimilar pumps run in parallel.

William (Bill) Mugford • Dear Mr Mohan Dudani, I would suggest the jockey pump shaft breakages occur due to the following sequence of events.

The jockey pump is to keep the system pressurised allowing for small leaks. It is probable the pump is running for long periods of time very close to its shut off head head at over 154 psi at very low flow. This causes heating of the water in the volute and recirculation cavitation with the attendant vibration and side forces on the impeller and shaft.

The water which comes through the relief valve will be hot and it goes back into the suction of the pump and comes out the discharge even hotter.

Note, the flow from any pressure relief valve arrangement used to protect the pump from overheating MUST be directed back to the suction storage tank so it can get a chance to cool down.

Re-positioning this pipe will help but there is more to the story.

Now when there is a fire and the automatic deluge valve operates for a sprinkler system or a fire hose is connected and the system pressure drops, then the jockey pump cannot keep up and the the first electric pump cuts in. This is rated at 120 psi at 1000 gpm.

Until the set pressure (unknown) where the 120 psi pump turns on, the jockey is trying its best to keep the system pressurised, which is probably why the motor fitted is over sized at 60 hp (as highlighted by Mr Mohamed Albadri) to cope with the low head, high flow the jockey pump sees.

Under these conditions it is possible the jockey pump is running too far to the right of the curve with the attendant vibration, cavitation and the high power being needed by the impeller coming through the pump shaft.

When the first main pump cuts in the system pressure will rise but to an unknown figure and the jocky pump will be running under unknown conditions.

Under these variable conditions, one possible outcome is shaft breakage as suggested by SL!

Performing a system review and watching the motor ammeter and suction and discharge pressure gauges on the jockey pump should tell you a little more on what happens.

When the jockey pump is running on its own the motor amps will be low and the differential pressure (discharge minus suction) can be plotted roughly on the pump curve.

Creating a simulated fire by opening slowly some hydrants should show you where the jockey pump operates at before and after the first main pump cuts in.

Randal Ferman • Operating at or near shutoff, the impeller radial loading will be high. Shaft failure may be due to cyclic bending fatigue at the shaft "stress riser" step location.

ARINDOM BORAH • Hi Mohan,
Pump doctoring requires a medical history too like human beings. You have mentioned that the pump shaft always breaks at a certain point. Two things need consideration here.
(1) Has the pump shaft problem been there during the entire pump life. If no, then you should compare the working before the first break and the working now. Check the flow, heads etc( but i feel that you do not have flowmeters etc installed as is the case in most old installations).
(2) The dynamic balance of your pump has gone haywire. Maybe some repair works were carried out to the impeller earlier. resulting in a loss of balance. The shaft is experiencing a torque near the bearing and is shearing off. Check the bearing and see if is the OEM recommended bearing. Changing the MOC will not help you here.

Actually, virtual trouble shooting has its limitations. Suggest you delve into the pump history and/or appoint a local pumpman. We are now in the age of specialization and if specialists are available, hiring them will save you money and time. I have worked in fire fighting systems and each of the main fire pumps will kick in in case of a pressure drop (in India it is 8.8Kg/sq cm and then increases to 144). The main pumps will start working only if the jockey pump is unable to maintain the pressure in the ring main. Hence it does not appear to me to be system related.

If your delivery NRV is non functioning, then when pump stops, there will be huge back pressure on the impeller. But this will depend upon the elevation difference.
Sorry, I could not be of more help.
Another thing I would like to say is that for a 60HP pump, the motor should be 4pole. 3000 rpm(syn) motors are not recommended for motors of such rating.

Ahmed Raza • The hydraulics that you've listed 150gpm @ 154psi requires no more than 25hp motor. And since this is a single-stage pump, fitting a 60hp motor means that the pump is fitted with NOL (non-overloading) motor over the entire pump's performance curve. This suggests that you're operating the jockey pump far away from its BEP (best efficiency point) (close to shut-off). Running the pump continuously over this point is not advisable. Would suggest that take the performance curve of your jockey pump and verify this.

If your pump is operating close to shut-off, I would recommend get another jockey pump (preferably multistage), because in your scenario you will get a better operating range in terms of pressure. Moreover, the payback cost of the jockey pump will be no more than a year considering your current power consumption.

Also note that unless it is a low-flow pump, a single-stage pump to act as jockey pump in your scenario is not recommended. You'll get a vertical multistage pump in SS internals for approx. $4500. You may calculate your energy figures being paid on the current pump based on this cost and have the new pump approved...best of luck!

PS: As William suggested, ensure that when your main pumps kick in the system, your jockey pump switches off...it is not advisable to run different size pumps in parallel unless they are selected to do so by the people who know what they're doing and in your case it seems the selection of jockey was incorrect...

ARINDOM BORAH • Hi Ahmed,
This is a fire fighting pump which means that
1) shut-off head should not be more then 120% of rated head (140% in case of VTP)
2) At 150% rated flow, head should not be less then 65% of rated head.
Because of these conditions, the drive motor is generally of higher rating ( the worst condition of 150% flow has to be considered ) . Again- a margin of safety is taken over that.The motor rating can be 60 HP therefore without the pump running far away from bep towards shut off as you have concluded.
You are right in advising Mohan to go for a multi-stage pump. The speed can also be brought down then to 1500 RPM. But I don't understand the MOC. Why SS? CI casing, Bronze Impeller and EN-8 or EN-9 shaft should be OK as the pump will handle water only.SS pump is not required for the jockey pump or am I missing something?

Ahmed Raza • Arindom, the guidelines that you are referring to, are valid for main pumps not for jockey pumps.

Pump in either 2P or 4P operation will not play any decisive role here considering the size of the pump. Low speed pump might last long but high speed pump will utilize less capital cost in comparison.

For MOC it doesn't matter much. My arguments were based on fabricated sheet metal pumps which in our region costs less as compared to a casted one.

ARINDOM BORAH • Hi Ahmed,
That is the crux of the matter. In India, we have the TAC regulations which is mandatory.
(see link
http://en.wikipedia.org/wiki/Fire_pump#Jockey_Pump

In NFPA-20 or UL..whether the same is applicable for jockey pump is mentioned somewhat ambiguously. TAC does not mention it either. Customers generally prefer to be safe and jockey pumps can be selected on the basis of Q or 150% Q. In this case, the selection must have been as per the TAC norms, since otherwise no one would have selected a 60HP drive motor.Some calculation errors may have taken place but I feel that it will be absurd if we conclude that much of error.n India, wherever I have supplied or quoted or designed, the jockey pump was considered on the basis of 150%Q in 75% of the cases.

Speed-yes theoretically-the advantages and disadvantages cancel out each other. But I am more comfortable with 4pole as I have seen for myself that 2P requires much more attention then 4P. At site, conditions are not as per theory and in case of foundation/base-plate problems, vibrations are transmitted at a faster rate. Therefore, the higher the speed-the more the vibration-the more are the adverse effects due to alignment problems getting compounded. You can call it a hunch but I have found that 1500 RPM requires less troubleshooting then 3000 (syn) . India is a huge country and sites are sometimes at remote areas. Hence,considering the importance of the application I prefer to go for a 4p motor.

Anyway, I had the chance of looking up this matter again so thanks! I don't think that our discussion was argumentative-it was /is constructive.. Looking at things from other points of view can only result in value addition. We become better engineers and maybe better human beings too.

Ahmed Raza • By arguments I meant "discussion".

Do you mean that jockey is being selected at the rated capacity of the main pump? Or 150% capacity of the main pump? For instance, the main pump is of 1000 USgpm. How would you design jockey pump flow?

ARINDOM BORAH • Abhayankar Saab.. What do u think? I would appreciate it highly if you can give your opinion considering your position of a pioneer in the development of the Indian Pump Industry'.

SL Abhyankar • I was away on family-outing since 24th Dec. Could not access the email. To refresh myself on the problem, I read through the original post of Mr. Mohan all over again. Mr. Mohan has mentioned pumps by the duties punched on their nameplates. Interestingly, he has mentioned values of pressures - 100 psi, 120 psi, 154 psi with electric motor-driven pumps and 'pressure-drop' values of 70 psi, 80 psi and 90 psi for engine-driven pumps.

First of all, shall I say that pumps work in a system. So their operating point (flow and pressure) is dictated by the system and not by the duties mentioned on their nameplates. So, to analyze the problem, one needs to understand the system-design.

Again he has stated that the jockey pump runs continuously. A pump running continuously does not mean that it is continuously delivering into the pipeline. Since its job is to keep the pipeline pressurized, once the desired pressure is reached, why should it keep running ? Mr. Mohan has mentioned that the pump has a safety valve to cut the pump off at 154 psi. "inside the pump". I am putting this phrase of Mr. Mohan between inverted commas, because the phrase is related to a fallacy that pressure inside the pump will be different from pressure outside the pump. Basic law of Pascal is "pressure acts equally in all directions". So 154 psi is pressure, not just inside the pump, but also in the system. in the pipeline. Does the system need that high a pressure ?

Since the other pumps are rated for pressures such as 100 psi, 120 psi, it seems 154 psi is too high a pressure. With such pressure being maintained by the continuously running jockey pump, there is no chance for the other pumps of lower pressure-rating being able to deliver any water into a pipeline pressurized to higher pressure as 154 psi.

At 154 psi the jockey pump is also reaching the limit of its safety valve setting. The system and any of the pumps are not expected to be developing a high pressure as 154 psi.

Simple solution that comes to mind is to bring down the setting of the relief valve to a lower value as 120 psi. if such pressure is appropriate for the system. Requirements of the system should be the point of reference and never the values on the nameplates of pumps.

Ahmed Raza • I agree with the lower pressure values but I still think that unless it is a low-flow pump; it will keep troubling even if it is operated between a range of 120~125psi.

SL Abhyankar • @ Ahmed Raza - When you say, "...I still think that ....it will keep troubling ...." is it just a hunch ? What is the technical logic ?

Ahmed Raza • Same reason as in my first post and based on my experience but since I don't have the performance curve and pump data that is why I used the term think...

SL Abhyankar • @Mr. Ahmed Raza - Such a prompt response ! Looks like we are on-line at the same time, myself in India, yourself where ?

My logic of suggesting lowering of pressure-setting of the relief valve was that, at lower pressure-setting, the jockey pump will be safely away from Minimum Safe Flow (MSF). By that, the pump may not suffer temperature rise. It could be the temperature rise which can be stressing the creep properties of the shaft and cause it to fail.

Ahmed Raza • Pakistan.

I do understand your point of view...however, all of these (including mine) are wild guesses unless we look at the pump details...looking at it alternately, may be it is a low-flow pump and operating already close to run-out...lets see if Mohan gets back on this...

Mohan Dudani , P.Eng , CMRP The pump in question is not pumping in a closed system. There is a continuous demand in certain areas of the plant. So although it is fire water system but same system supplies water to certain area of the plant where it is used. This explains reason for its continuous operation. This is an old system with drawing dated 1990.
There had been pump failures since 1990 but I do not have the data. From the interviews I had with the personnel I do not think that problem started two years ago.
I have the pump curve. I am trying to find the current operating point on the curve. I am looking for suitable flow meter for this. Suction and discharge lines are 4”. Any advice on the flow meter – Portable one which I can get from a suitable vendor in Canada or USA
I will install calibrated suction and discharge pressure gauge on the pump to find out head developed
Pump is Paramount Make, Model 4-M3-A
RPM: 3510
Original Motor was 50 hp which was changed to 60 hp
Pump is rated at 150gpm and 150psi (I do not know that this 150 psi is pump shut off head or the pressure equivalent of the head at 150gpm, how can I find this Tim?)
I do not have information about last servicing of relief valve.
There is no heard “chatter” in the relief valve.
I will find out other pumps if they run straight up to full speed, when they come in operation.
The relief valve at the discharge end is Schroeder make. Do not have the details of the valve in operation other than the name (eg set point??), but new valve to replace this has been procured and awaiting replacement, it is same make and is Automatic recirculation valve made out of forged steel material 1.0460/ P250GH (A105) CS with modulating minimum flow regulation for vertical installation with bypass valve head and non-return valve in the bypass
Last vibration reading on pump bearing was 0.23 IPS. There is no change in this since it is monitored weekly.

I tried to find out the correlation between shaft breakage time and starting of any other pump in the system but did not get the information.
The relief valve at the discharge end is piped back to a 12” header. This 12” header is connected to upstream 8” pipe which is piped to a 4” pipe which is connected to pump suction.
The coupling used is Kop- flex, Elastomeric type

Operational information immediately after the pump shaft was replaced last on December 15, 2011: The initial pressure developed by the pump in question was not enough so 75 hp Pump kicked in which is rated for 1000gpm and 100 psi. Opening a 1 ½ “ fire hose was causing a pressure drop below 90 psi , which prompted diesel fire pump to kick in after its 9 seconds timer delay designed to absorb hydraulic disturbance in the system . This in turn caused a fire standpipe flow switch in one of the building to go into alarm. The electric motor fire pump cut in and cut out of 100 psi and 120 psi respectively was increased to cut in at 115 psi on a pressure drop and cut out at 130psi. After the pressure has been sustained above 130 psi the 75 hp pump will stop running in 5 minutes. This will give enough of a margin to prevent unnecessary start of the diesel fire pump. Since then only 60 hp pump is running continuously.

SL Abhyankar on 9^th Jan 2012• Thanks Mr. Mohan Dudani for the exhaustive details.
(1) Since you have the pump-curves, even pressure gauge reading will give you some idea of the flow-rate, without installing a flowmeter.
(1a) Among flowmeters, ultrasonic ones are portable, more universal (fit on to range of pipe-sizes) and non-intrusive. But inaccuracies in the installation can influence accuracy of measurement.
(1b) Since it is not a closed system, there could be flow at open end of a pipe. If you do Google search on "measuring flow at open end of pipe", you will get interesting simple (though not accurate) methods.

(2) Jockey pumps are used to take care of leakages in a pressurized pipeline. Since in your system there is some flow for using fire water as process water, Will it be good to use the main pump itself continuously, instead of running the jockey pump continuously ? This is another option, you may consider.

(3) Since diesel pumps also tend to kick in, it seems, you will end up working out an O&M manual for the system ! Actually, there should have been one from the people who designed and installed the system !!

Aristides Nuñez • on 11 Jan 2011.
Its been an interesting discussion in this tread.
I just got surprised about the design criteria for jockey pump in India. (Talking about 150%)
I had some experience in fire fighting set in Spain, and according to Spanish UNE regulations and Europeans as well there is not indication that the jockey pump should be selected under that rigorous standard.

Chris Graham • Is it possible that the sleeve that has been replaced with extra shaft diameter was actually there to 'spread' stress? Perhaps a shaft of this design has a lifetime that is made reasonable by the sleeve but now stress is focussed at the step and excessive strain occurs? Just a thought...

Adam Jung • I would also look into the impeller abutment step you mentioned make sure it has a ample radius and the impeller bore has a chamfer so it will evenly contact the abutment face. It is important to make sure everything is parrallel and perpendicular so verify the impeller abutment face is true. I would say 0.001 in' max.
Try and eliminate any stress risers, make sure the shaft has been deburred properly and has chamfers and sharp corners have been knocked down where applicable.
All keyways should have a radius at the bottom, as should all diameter changes or split ring grooves.
From my experience many of these failures occur do to poor machining practice's
Also verify the taper of the impeller and shaft are correct and will have proper contact, also verify runnouts of the shaft alone and with the impeller mounted to the shaft. If it changes some binding might be occuring. If the impeller is keyed also verify the keyover clearance. A good material with excellent toughness and corrosion resistance is Nitronic 50 or Carpenter 625 material. Best of luck