13 replies to this topic

[ethanhan]

Dear all Experts

With refer to Exxon design practice, standby (2X100%) Centrifugal pump driver shall be size up to end of curve when auto start operation. Would you mind to explain the reason behind this?

Regards
Ethanhan

[Zauberberg]

There are two very good reasons behind this requirement:

1. Normally you want both your reflux pumps to be able to handle full/maximum rated flow, which is on the most right side of the curve.

2. The starting torque of the motor is much higher than in the case of a steady, continuous flow delivery by the pump. The explanation is that more energy is required to accelerate a body (of fluid), than to keep it in motion. Try pushing your car by your own hands on the parking and you will be convinced in this statement. That is why large pumps are usually started with the discharge valve partially throttled.

The best design practice is to choose such standard motor size that can provide the power sufficient to deliver maximum rated flow at the maximum possible impeller size, for given pump casing. That way, in case you ever need to trim the impeller in future, you will not need to replace the motor as well.

[sheiko]

I agree with Zauberberg. Drivers are generally selected for the so-called "end of curve" (and usually for the maximum impeller size for the case) for several reasons. One is that if the pump just doesn't make the required head and flow, an impeller change can cure that without having to change the driver side of the pump. The other is that increasing the capacity of the pump is a relatively cheap "revamp" if the motor is already sized for the maximum (i.e. end of curve with maximum impeller diameter) horsepower requirement.

I also permit myself to take the opportunity of this discussion to ask why drivers are selected for the "end of curve" for:
- pumps under level control?
- pumps in parallel operation?

Thanks

Edited by sheiko, 25 January 2011 - 06:10 AM.

[ethanhan]

Hi Sheiko/Zauberberg

But I guess because the opearation is auto starting and no operator is opening the discharge valve gradually therefore size up to EOC is required, do you agree?

For parellel pump operation, in the event of one pump trip, the other pump operating point or intersect point will be shifted to the right of the curve hence more power is required.

Regards

Ethanhan

[Zauberberg]

Sheiko,

I haven't encountered this requirement for the two services you mentioned - at least not explicitly stated as such. To me personally, it is more like a general good practice, a "rule of thumb" that had arisen amongst engineers long time ago after accumulating years of field experience. And if I am about to size the pump/motor for any application involving large variations in flow demand, or when resistance to flow could change in future, I'd always go for the motor size that can handle the biggest impeller, and the highest rated flow.

My answer to your query would be the following:

- Pumps under level control may be required to pump more liquid when unexpected surge (or change of process conditions upstream of the drum feeding the pump) cause high inflow of liquid, and subsequently high liquid level in the drum. In order to avoid upsets - or even process shutdown, depending on severity of high level in the drum - these pumps would need to evacuate extremely high volumes or liquid and probably this will be happening through 100% open level control valve. This is just by itself an extremely high-flow condition calling for the motor capable of delivering the required power to the pump.

- Pumps operating in parallel are a somewhat similar example, in the sense that unexpected shutdown of the first pump will automatically put high flow on the second pump that has remained in operation. Again, large flow that needs to be handled but now with a single pump, will also call for more power delivered by the motor.

Perhaps there is some more sound reasoning behind the standard/practice you asked about - this is just my answer based on common logic.

[Zauberberg]

Ethanhan,

I missed your reply but I concur with what you are saying. Your line of thinking with respect to parallel pumps is exactly what I had in mind.

As for the automatic restart, basically you are right - there is no operator in the field and the only thing that will create resistance to flow (apart from piping itself and the difference in pressure between source/receiver) will be the actual position of the control valve on the discharge side of the pump.

Best regards,

[fallah]

Ethanhan:

Auto start operation of the pumps would be normally accompanying with Open Discharge Valve due to not presenting the operator at the time of start-up. Therefore, the resistance against pump start-up would be minimum leading to have highest flowrate in pump operating range,i.e. flow at the end of pump curve.

Hence,in such cases the motor should be sized based on required power at the pump end of curve.

Edited by Art Montemayor, 27 March 2012 - 03:04 PM.

[JoVi]

Dear all respected members,
I have read through this forum and agree that for centrifugal pump parallel operation with autostart, motor shall be sized for end of curve, but take a case, there is a pressure control valve (not recirculation) in the pump common discharge header. So in case of pump failure, the system curve will not shift to right. In such case is there need to size motor for end of curve?

[fallah]

_{JoVi,
Yes, due to the possibility of the PCV failure to be in full open position...
Fallah}

Edited by fallah, 28 March 2012 - 03:04 AM.

[JoVi]

the standy comes in operation due to failure of running pump, and at the same time the PCV fails in open position, will it be double jeopardy?

[kkala]

I have offered services to a Process Dept of a local engineering company, following notes can be useful on the subject.
1. In elaborating process duty specs for refinery boilers, autostart standby boiler feed water pump was not specified for end of curve operation. There is control valve upstream (between pump and steam drum), while at normal conditions most of pump head is due to steam pressure in steam drum.
2. You had better clarify whether end of curve operation concerns motor size, or also NPSH requirement (see http://www.cheresour...-can-type-pump/ '> http://www.cheresour...-can-type-pump/ , post No 9, also valid for horizontal centrifugal pumps). NPSHr increases with flow, but suction level may not be at minimum when "end of curve" occurs.

[S.AHMAD]

the standy comes in operation due to failure of running pump, and at the same time the PCV fails in open position, will it be double jeopardy?

1. The case you mentioned is a double jeopardy.
2. However, we need to consider the possibility the control valve stuck in an open position. Not due to utility failure, but due to mechanical damage of the valve itself or the utility failure is localized that is specific to the control valve.

Edited by S.AHMAD, 30 March 2012 - 10:59 PM.

[fallah]

the standy comes in operation due to failure of running pump, and at the same time the PCV fails in open position, will it be double jeopardy?

JoVi,

Your concern isn't overpressure protection, hence being or not being double jeopardy might not be considered as a basis for motor sizing. Nevertheless, referring to sec 4.2.2 of API 521 you can consider absence of favourable instrumentation response as a "latent failure" in combination with running pump failure, then the matter couldn't be considered double jeopardy events.

Fallah

[S.AHMAD]

1. i have two cases in the past, where i was asked to increase pump capacity due to higher capacity requirement.
2. The pump can accommodate new capacity by increasing the impeller size but the motor power could not cope with the higher power required by the pump. As a result I have to specify a new motor.
3. I would very much appreciate if the "previous process designer" be more generous by specifying bigger motor which was cheaper at that time as compared to current cost of motor.
4. The moral of the story is that, do not be too fussy about over sizing of pump motor. It has certainly some advantages especially when most plant management is trynig very hard to maximize production up to the maximum attainable capacity.

Edited by S.AHMAD, 02 April 2012 - 02:38 AM.