20 replies to this topic

[shan]

Hi Everybody,

We are working on a ballast water pump with vertical discharge line sections to the platform overboard. When the pump is started up, the discharge line is empty (no back pressure). Will the pump run at the maximum flow (beyond the pump curve) for a while to fill the discharge line? Or should we design a start-up line to fill the pump discharge line section before turning on the motor?

Thanks

Shan

[Nikhiln]

Hi,

I presume the said pump is centrifugal (since you mention a curve). I think it is precisely to avoid this condition that we start the centrifugal pump with discharge valve closed, so that some backpressure is provided to prevent runout condition (beyond max.flow/max power).
Do correct me if my thoughts are not right.

[shan]

Hi,

I presume the said pump is centrifugal (since you mention a curve). I think it is precisely to avoid this condition that we start the centrifugal pump with discharge valve closed, so that some backpressure is provided to prevent runout condition (beyond max.flow/max power).
Do correct me if my thoughts are not right.

If you close the discharge valve, there is still no backpressure because the pipe section is empty. Therefore, the pump is still at the runout condition at the first a few seconds of startup. Then, the pump will switch to the shut off condition with highest pressure and no flow when the upstream of pipe section is filled. When you open the discharge valve, the pump is back to runout condition again because the pipe section of downstream of valve is still empty. Therefore, it is not a good idea to close discharge valve during the startup of pump with vertical discharge line.

[daryon]

Hi Shan,

Here's my thoughts:

It may be possible to run the pump off the curve for a while to fill the line, but it's likely the motor will trip on overload current. Assuming the pump is driven by an electric motor?

You should be able to find out the overload current trip setting (e.g. 690A), if you know the power supply voltage(e.g. 440 V), and motor efficiency you can estimate the power at which the pump motor will trip. Then from the pump curve you can see at what head and flow capacity corresponds to the motor trip. If its above 0 head you may need to fill the line a bit to stop the motor tripping.

Or perhaps you can override the motor overload current trip for a short period to fill the line. Running above the overload current for a sustained period can melt wires and damage the motor, but a short period to fill the line might be o.k. But i'm really not sure if that's acceptable. A start-up fill line maybe required - better to have this than risk damaging the motor.

rgds
daryon

[Nikhiln]

I believe it is still the accepted practice to start centrifugal pumps this way. As for the case of the pump seeing runout again after opening the discharge valve, that is why the usual procedure is to open the discharge valve SLOWLY. This ensures that the pipe section further downstream doesnt get a shock, simultaneously reducing the potential for runout. Alternately, the discharge valve is kept not fully closed, but cracked open, which takes care of both aspects.
As for the problem of experiencing shutoff pressure, the recirculation line is kept open when discharge valve is closed. If there is no recirculation line, then the problem is solved by keeping discharge valve cracked open, instead of fully closed.
Yes, I do accept that a very small section of the discharge line (upto valve)will see runout, but then that is why discharge valves are placed as near as possible to the pump (after providing space for the NRV). Also in many cases, the NRV itself acts as the closed discharge valve reducing the pipe length subject to runout.
Finally, when priming the pump, many a time, we open the discharge vent and allow the fluid to push air out of the line. (of course, only in the case of fluids which can be released like that). Apart from achieving the target of priming, it also reduces the length of empty pipe upto the valve.
It is particularly applicable for vertical discharge lines. So the nearer the discharge vent tapping is to the discharge valve, the lesser length for runout.
I am writing from my experiences of having operated a petrochemical unit for 3 years earlier in my career. I have not come across a case where we decided to open or close the discharge valve based on the orientation of the discharge line. Of course, that does not in anyway mean to say that there is no such consideration to be applied. There are a lot of things I have not yet seen. So I do accept the fact that maybe it could be a point to be considered.

Cheers.

[fallah]

Hi Shan,

Seems the pump start up would be based on open discharge valve.Anyway,for operating as you mentioned it is needed the motor to be sized for flow at the end of curve,also NPSHR value at the end of curve should be acceptable regarding NPSHA (assuming NPSHR would be increasing with increase of flowrate).

[shan]

The question is how you define “SLOWLY”? 20% opening per minute or fully open the valve in more than 10 minutes? There is no option of partial opening discharge valve because the valve is an on/off ball valve. Also, it is not practical to hire a worker to just slowly open/close the discharge valve because the pump will be start/stop many times a day based on the upstream container liquid level, which is different with a continuously operating pump in a petrochemical unit.

[shan]

Hi Shan,

Seems the pump start up would be based on open discharge valve.Anyway,for operating as you mentioned it is needed the motor to be sized for flow at the end of curve,also NPSHR value at the end of curve should be acceptable regarding NPSHA (assuming NPSHR would be increasing with increase of flowrate).

The start-up operation is not at the end of curve. It is off the curve. I am not sure if the vendor allows to extend the pump curve to 0 ft heads to get flow rate and BHP.

[Qalander (Chem)]

Dear,I do not really grasp what is your exact query here since I feel that some info is missing.

Please confirm:

• the drive is turbine type or VFD electric motor thereby RPM and respectively the flow rate will vary?
• there is motorized discharge valve provided to be opened whenever operation need arises?
• Multi-stage centrifugal pumps are at times operated through a minimum flow line initially and subsequently main discharge line is commissioned.
• Pump in question has sufficient NPSHA at all times and all flow rate conditions encountered!

Hope these may lead you to some way forward in resolving your query,best of luck.

[shan]

• The drive is electrical motor without VFD. Therefore, RPM is 2950 rpm constantly.
• There is no minimum flow by-pass line because the discharge is overboard and piping design pressure is higher than the pump shut off pressure
• There is a check valve, an on/off control valve and a hand ball valve on the discharge line.
• NPSHA is OK for the designed rate and even OK at the maximum rate at the right end of pump curve. Not sure the if it is allowed to extend the flow, BHP, and NPSHR to the 0 ft head discharge point because the pump curve is pretty flat.

[Qalander (Chem)]

Dear shan

As I recall my visiting on oceangoing Oil Tankers and their manifolds and various pipelines connected it just reminds me that:

• normally and most usually vertical sections should be fully primed with the under pumping liquid (in your case ballast water).
• Having said that the system is most usually designed in such away to have complete drain outs at such isolating points and
• washing out to slops system for ensure "switch loading" of various cargoes without jeopardizing quality in each case.and
• thus Most presumably your case shall be similar and the normal back pressure should be the vertical water head &
• if I understood correctly the hand operated ball valve(as mentioned by you) should provide needful gradual opening of discharge
• thus avoiding of extra ordinary "Initial starting torque" from the concerned E/motor driven pump at start up.

Hope this proves useful and guides to something positive!

[kkala]

Having worked for almost 6 years in a fertilizer plant, I can say that starting a centrifugal pump with the first discharge valve closed was the recommended practice. Pumps were mainly transferring acid or ammonia salt solutions (no minimum flow at that time) at rates 40-120 m3/h, distances 20-200 m, altitude differences up to +/- 20 m. After starting the pump up, the valve was gradually opened in (say) a minute. We neglected to apply it in all cases, but this was the recommended practice.

An operating manual of another installation recommended this practice later on.
When you start the pumps from the Control Room (or automatically) this is not so easy, but you can have remote manual operation of first discharge valve, or even automatic gradual opening. In recent years I did not meet anybody dealing with these "details", at least from the design viewpoint (I have been in the office, far from operations).

Concerning time of discharge valve opening, it may get fully open when the discharge line gets full of liquid in downstream piping.

Edited by kkala, 23 May 2010 - 04:44 AM.

[Art Montemayor]

I have covered this topic and type of centrifugal pump startup operation many, many times in the past in these forums and elsewhere. I've done it because I have come across this application in the field in many circumstances and in many countries in the past. I've written many startup manuals and edited even more in the past regarding this very subject. Not all centrifugal pump startup procedures fall into the same general instructions. Everyone is different. But all startup instructions and operating recommendations always center about some very simple and specific properties and characteristics regarding centrifugal pumps.

I fully agree with the general statements offered by Kkala. He has followed the very basic principles of centrifugal pumps. Allow me to state some of these:

There is no such thing as a true and realistic "beyond the pump curve" zone. There isn't even a realistic and credible "end of pump curve". Once you start approaching the end of the depicted pump performance curve on most centrifugal pumps you should be on the alert and not trust the predicted pump performance anymore. A centrifugal pump can only truly be trusted to operate on the empirical curve that it has shown to follow on the test stand. And most manufacturers don't bother to define and prove the performance of the entire theoretical performance curve because it isn't expected of them and it doesn't make for practical sense. The manufacturer isn't expected to perform beyond the operating points that are beyond those guaranteed. The point here is that everyone should be wary of pump performances that approach the "end of the curve". The pump performance doesn't necessarily have to follow the very smooth and theoretical curve depicted on the performance curve.

The GENERAL method of starting up centrifugal pumps is by having a discharge valve (preferably a globe valve) closed at the very outset of the pump's startup. This is done in order to achieve the pump's expected BOP (Best Operating Point) as soon as possible. Of course, the discharge line downstream of the globe valve may be void or empty. THIS DOESN'T MAKE ANY DIFFERENCE TO THE PUMP. The pump follows the predicted performance shown on the performance curve up to and around the BOP and the TDH (Total Developed Head) is set by the opening of the globe valve on the discharge at the startup outset. After initial startup, this globe valve is throttled to develop the design or desired TDH. The downstream line will start to fill up and reach the TDH expected. At that point, the globe valve is totally opened.

The above operation is the BEST way to startup a centrifugal pump. The manual operation can easily be duplicated by automatic, remote startup with a PLC or with adequate instrumentation. I know because I've done it many times and it works. Beside the discharge globe valve, I've always used a minimum flow by-pass circuit back to the source tank. This minimum flow line helps to establish a positivet pumping situation immediately prior to putting the pump on line. I highly recommend this type of centrifugal pump startup - whether manual or automatic because it imposes the minimum mechanical and process load on the pump and ensures a steady and even liquid flow from the very outset.

I agree with Kkala on every point with the only exception being that I give this method importance NOT because it is stated in operating and instruction manuals, but because it makes logical and common fluid flow sense when dealing with centrifugal pumps. I know this because I either wrote or helped to write a lot of startup manuals on this subject. My advice would be to not burn incense or geneflect before operating manuals. These manuals were written by engineers just like you. There are some manuals that are pieces of engineering excellence. Most of these were written by some of the big international and renown engineering firms. They are excellent because they are carefully and technically written with care and precision using the very best experience known. This is not because big engineering companies like to write engineering opuses. This is because they can't afford to make a mistake and give instructions that can produce a wrong result. They are legally liable for engineering performance and they can't afford to make a stupid mistake. My point here is that YOU - just like them - can make the same decisions and designs when confronted with designing a dependable and reliable centrifugal pump startup plan. All you have to do is make certain and be sure that you apply good and experienced engineering decisions.

[Qalander (Chem)]

Thanks "Art" for your final elaborative explanation as usually expected. In a Nice Cordial way.

An Admirer of Yours, indeed.

[chemsac2]

Hi Shan,

Situation you described is very common in refinery drain systems. Drain pump is normally off, but interlock starts the pump on high level in drain vessel. Valve on pump discharge is kept locked open.

We have specified motor suitable for end of curve operation for such systems and I have seen it working fine.

Here is how I think pump system with open discharge valve would behave:

I don't think all systems with discharge valve open would shift to end-of-curve during start-up. It would do so when system is friction dominated and static/pressure heads are less in comparison. When static and/or pressure head dominate, system resistance that does not vary with flow itself may never be beyond head of curve. Frictional resistance on increase in flow would only bring it closer to BEP.

Even when system without significant static or pressure head, it may not get to end of curve. This has to be seen as a transient i.e. pump is also picking up speed and fluid is moving in the pipe. For example assume pump takes 20 seconds to reach full speed. Then each second can be viewed as a pump system with its static head, frictional head and pressure head specification. Pressure head specification would be constant for entire duration as only gas-liquid front would move. Static and frictional terms would change with time. It would be an involving calculation and may not be necessary for most of the systems as duration is very small.

Even if pump does get beyond end-of-curve, it may not be that big an issue for short period of time (provided motor is suitable). I have heard of a case study in which pump was operating beyond end of curve (reason: designer calculated pump head too conservatively for a distribution system without flow control). The said pump was erratic with motor overcurrent issues, but otherwise continued to operate (problem was later rectified by addition resistance).

Regards

[Qalander (Chem)]

Dear chemsac2 Correct,

We also had at our previous employer's refinery main waste water system:

• a submersible type/flooded suction design pump operated with level float actuated motor switch; however
• this had a much smaller i.e. Probably 4" diameter immediate pipeline taking flow to the discharge valve 6" Diameter, NRV Upstream of connection point with main pumping pipeline to the sea discharge
• located on top (ground level) platform.

[shan]

Hi Art,

Your start-up procedure is good for pump with a minimum by-pass line and/or a discharge throttling valves, which we don’t have. Your mentions on pump end point are probably true in the real lift. However, we, as design engineers, do not want to take the responsibilities to advise the operators to run out off pump curve without written agreement from the pump manufacturer. As you know, if we do so, we will be the target of finger pointing in the case something is wrong.

The solution we figured out is to lower the pump (multi-stage vertical pump) so that a section of vertical discharge line is submerged into lower level of the suction tank. Therefore, the section of line is always filled with ballast water as the start-up discharge head.

Thanks

Shan

[Qalander (Chem)]

That's almost exactly the case envisaged It to be resolved "Flooded suction and discharge though the annuler section along side the driving shaft" to top level and thereafter throttle control whichever way one feels better, indeed!

Best of luck shan with the proposed schem ending up successful.

[agorag]

Interesting discussion! And a good ending!

Since ballast water requirement is usually huge, axial flow pumps are employed. And unlike the common radial flow centrifugal pumps, the characteristic curve of axial flow pumps is such that, the power is higher at lower flows and lower at higher flows. And this is precisely the reason why axial flow pumps are never started with their discharge valve throttled. If done so, since the volume of flow is huge, the pipe is bound to burst and so is the motor.

Had the questioner provided the characteristic curve and a drawing of the pump along with the first question, such an interesting discussion would not have taken place.

Good, in a way...

[Qalander (Chem)]

Thanks agorag.

Although I have probably not come across any of these, but probably functioning is similar to multi-stage turbine type pumps.

Kindly educate/improve awareness on this issue,

Thanks in advance.

[sheiko]

An article on pump startup FYI: http://www.chemicalp...s/2005/415.html