Safeguarding a pump against too low flow

[Yobbo]

Hi,
I have got a problem with a periphery pump. It ran for about six months and then broke down. It is a pump of the brand Sero. We were warned by the supplier, that under no circumstances the flow should decrease below 1 m3/hr. Originally the pump was supposted to be running continuously, but because of this minimum flow limit we installed a switch on/of control of the pump based on the level in the suction tank. This measure apparently wasn't effective. Therefore I am looking for a more effective safeguarding method, so that we can apply the pump foolproof. Questions:

1. Is there a proven robust method to assure that the pump flow will never decrease below the minimum flow limit?
2. Are you familiar with pumps like these and what is your opinion about the sensitivity to the minimal flow limit?


Thanking you up front for your time and attention and wishing you a healthy and prosperous 2011

Karel Postulart, The Netherlands
Nuon Power Generation

 

[Yobbo]

I forgot to add some of the process parameters. The design flow should be about 2 m3/h and the design head about 270 meter. The pump in question has got avery low NHSPr ( 1 meter ) and the suction inlet is axially orientated. The bearing at that side is depending on the medium flow for lubrication. Could that last fact be the sole cause to the problem we experienced?

Karel Postulart, The Netherlands
Nuon Power Generation

 

[JJPellin]

The preferred method in our plant would be to measure the flow with a flow transmitter. Build a spill-back line from the pump back to the tank. Adjust the flow controller to open the spill-back to maintain the required minimum flow. If the outgoing flow drops below the set-point, the spill-back will open and protect the pump. Admittedly, this is complicated and expensive.

For less critical services, we have used a low amp trip to protect the pump. Assuming that the pump has a horsepower curve that drops to minimum as the flow approaches zero, this can prevent pump damage. This installs in the switch gear and can be adjusted based on the measured amp draw with the pump running at minimum flow. If the flow drops below minimum, the amps will drop below the trip set-point and the pump will shut down. This would require operator intervention to reset the trip and restart the pump manually. A low amp trip is an off-the-shelf item that could be installed in a few hours if there is room in the switch gear/motor control room.

Johnny Pellin

 

[ScottyUK]

A quick comment on 'low amp trips'. Current monitoring can certainly be used, but for a few pounds / dollars more you can buy a device which measures the true power consumed by the pump which is a far better indicator of how hard the pump is working. Slow speed motors in particular can draw a lot of current even when idling, so current is a rather poor measurand. The principle described by Johhny still applies though.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[BigInch]

I'd much prefer monitoring the proper variable to guarantee the most effective control in the most direct manner possible. If you want to control flow, monitor flow and control flow. Since you appear to have several pumps there, I'd guess this is a rather critical service and Johnny did say did say that current measure is only used for noncritical services, probably exactly for that reason mentioned by Scotty; it's not always reliable, as a locked rotor condition may very well give the highest current draw of all.

Given your 2 m3/h normal flowrate, the 1 m3/h minimum flow specified by the mfgr seems very high for a typical centrifugal pump. It is a centrifugal pump right? Normally one would expect to see warnings at < 20% flow.

And ... How are you controlling flowrate to these pumps now?

 

[micalbrch]

It's a side channel pump, German brand. See:

http://www.seroweb.de/english/index.htm

Monitoring the power consumptions as suggested before seems to be the best safeguard.

 

[ccfowler]

I agree with BigInch. Monitoring the flow would be my first choice for assuring the required minimum flow.

If the cost of measuring the flow is too great for this situation, my second choice would be to monitor the temperature rise from suction to discharge. If the flow rate is too low, the temperature rise will be too great. Monitoring the two separate temperatures (suction and discharge) and comparing them is not a good choice due to instrument error, drift, ..., plus it will cost much more due to the amount of hardware required.

The temperature rise is more reliably and cheaply measured by using two insulated junction thermocouples configured to simply measure the temperature difference. Very simple circuitry could be used to open a solenoid valve on a recirculation line whenever the voltage generated by the differential thermocouples exceeds the chosen maximum value.

It would be wise to include some "dead band" in the control scheme to keep the solenoid valve open until the differential thermocouple voltage drops below some suitably smaller value than the voltage that is used to open the solenoid valve.

If I understand this pump properly, it is a variant of the regenerative turbine pump configuration. I have limited experience with these, but as I recall, they tend to have a fairly significant temperature rise at lower flow rates. I presume that this is the primary concern for establishing the minimum flow requirement.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[tr1ntx]

What was the nature of the pump's breakdown?

 

[Yobbo]

Hello,

Thank you for your answers so far. Indeed we are talking about a sidechannel pump. It sucks its fluid out of a suction tank, that is level controlled. The manipulating actuator is a control valve in the discharge line. If I get it right the preferred way of safeguarding is measuring the flow and comparing its value with the minimum allowed limit. Fact is that this pump in a way is only of minor significance to the main process. So I would prefer acting on the amperage of the motor. Using the temperature difference might be well applicable as well. What temperature differences are we talking about here? Does it provide significant reliable differences regarding the possible impact of inaccuracy of temperature measurements? What I would like to know apart from these solutions is if I should consider a completly different type of pump? I have got the feeling that sidechannel pumps are far more sensitive then the usual centrifugal pumps.

Karel Postulart, The Netherlands
Nuon Power Generation

 

[hahor]

Yobbo,

why not just implement a minimum flow recycle line with a restriction orifice (Mindestmengenbypaß)? Just have a look at the pump curve and calculate an orifice diameter for pressure and flow conditions at the minimum duty point. When flow increases dp goes down and less flow will recycle back to the suction drum. We used that set-up a lot when minimum flow was required for cooling reasons especially for canned motor pumps.

Kind Regards,
hahor

 

[BigInch]

The problem with orifices is that they are also open when you don't need them.

If you ONLY have a choice between temperature and amps, use temperature! Pumps don't care about amps and if they get to be too much, the breaker trips to keep the motor healthy, but pumps can and do get pretty excited when fluid temperature rises and its a sure indicator of Lo-Lo flow. Temp is not going to help you keep away from low NPSHA, but neither will monitoring amps.

 

[JJPellin]

Since you have a downstream level control valve, you have one other option. Since you know the Cv of the valve, you can infer flow from the valve position if downstream pressure is constant. You could limit valve position with a hard stop or a controls limit. Or, you could just alarm on low valve position.

I still like amps for our applications that are similar to what you describe. Amps will go low if the pump is gassed up completely and running dry. Amps works if the pump is completely dead-headed with zero flow. Depending on the location of the temperature probes, temperature might miss this. Temperature could also be a problem for a very high flow, low head pump. Even at very low flow, the temperature rise might be quite low.

The steeper the horsepower curve, the easier it is to protect with amps. I am not familiar with a side-channel pump. So, I may be missing something.


Johnny Pellin

 

[Yobbo]

Thank you for your view on the case. The pump characteristic is quite steep ( almost linear ), so the amperage monitoring should work properly then.

Karel Postulart, The Netherlands
Nuon Power Generation

 

[1gibson]

What about a pressure switch on the discharge that opens a solenoid valve to smaller orificed line. You have your orifice only when needed.

Suppose you could trigger the solenoid valve with any measurement you'd like.