5 replies to this topic

[thorium90]

Hi all,

 

I found myself in an interesting debate with other engineers recently. It concerns the starting of a centrifugal pump and its discharge valve position. I have compiled some of the points mentioned, below.

Should a centrifugal pump be started up with its discharge valve closed or open or slightly open? If started closed, how long before opening it?

What if the pump is equipped with a VSD? Does the rule still apply?

What if the pump is a submersible rather than a horizontal or vertical type?

What if the pump is quite large?

What if the discharge is to an atmospheric tank instead of a pressurized line?

I have noted that there are wildly varying opinions out there on how to start a centrifugal pump and thus I felt a good discussion is in order.Do share your opinions and experiences.

 

[curious_cat]

Without attempting a justification, what I thought was the "right" way was to have the discharge valve slightly open and then wait for the pump to come up to speed and then smoothly / gradually open the discharge fully. 

 

PS. I assumed this was to limit the draw of power initially and also to aid suction in not cavitating by having a large flow rate while the static fluid column  was accelerating up to speed. 

 

On the other hand, wait too long and you might get heating up. Especially bad with the high vapor pressure fluids. 

 

If the pump did have a VFD I assume the throttle discharging procedure is redundant; you'd start using a low speed & slowly ramp up the motor to the rated flow. The discharge valve would stay open all through. 

 

Again, not an expert. Just putting my thinking out there. I might be totally wrong. 

Edited by curious_cat, 05 September 2015 - 12:15 PM.

[Bobby Strain]

I wonder what Norm Liberman would suggest.

 

Bobby

[curious_cat]

I wonder what Norm Liberman would suggest.

 

Indeed! A lot of the useful insights about pump operations I've learnt are from Norm Liberman's books. 

 

Hence, I decided to dig up my books & here's what I found Norm saying ( Process Equipment Malfunctions: Techniques to Identify and Correct Plant Problems.):

 

"The basic idea is that when we start a pump, the liquid in the suction line is stagnant. To accelerate the mass of liquid takes energy. The energy cannot come from the pump. It comes from the available NPSH. This starting NPSH is subtracted from the NPSH needed to run the pump. So, to avoid cavitating the pump on startup, the good operator opens the discharge valve slowly—which is fine, up to a point. The difficulty is that running the pump at a very low rate for a long time causes the barrier fluid, or seal flush, between the seal faces to dry out. This also will damage the seal faces..........There's a widespread misconception that starting a centrifugal pump with the discharge shut is a good idea. This too will overheat the liquid inside the pump case. The discharge valve needs to be opened one-quarter to one-half turn, and not shut on startup."

Edited by curious_cat, 06 September 2015 - 12:02 AM.

[thorium90]

What if the pump is a large VFD submersible? Since the liquid is technically in a chamber rather than in the casing, the overheating of the liquid in the pump casing is not quite relevant then? If it also has a VFD, then no control of the discharge valve is required during startup? If the discharge piping is large enough, then it should be able to take some fluctuation? Therefore, for submersible pumps the discharge valve can be closed at startup and since it has a VFD, it can also be opened at startup?

Edited by thorium90, 06 September 2015 - 12:17 AM.

[PingPong]

As far as I know pumps are usually started with a closed discharge valve to avoid that the electric motor driver gets overloaded.

 

If one looks at a typical centrifugal pump curve with also a power curve on it, one usually sees that power goes up towards end of curve

 

So unless the electric motor is sized for end of curve, one has to start a centrifugal pump with a (nearly) closed valve to avoid motor overload.

 

In my design experience, where sometimes a pump is to be started automatically on high level in suction vessel, or too low flow in discharge system (automatic start of spare when main fails), or whatever reason, the motor of that pump is either sized for end of curve, or the pump gets an automatic on/off valve, connected with the pump start/stop signal, that is normally closed when pump is not running, but opens after the pump is running. And automatically closes when the pump stops.

 

It is also safer to have the discharge valve of a non-running pump always closed, because otherwise the pump check valve is the only protection of the suction system against backflow from the higher pressure of the discharge system. Check valves are unreliable as they are often damaged or even break due to wear and tear. 

Edited by PingPong, 06 September 2015 - 04:51 AM.