A limit switch on the valve to shut down the pump or alarm if it is closed while pump is running? Temperature monitor on the product.... too much lag probably to help much. Monitor the amperage of the motor drive, it will go down when pumping against a closed valve.

Good luck
Ralph

You could have a minimum flow recirc with flow restrictor always in service... wastes a little energy.

Or you could put an auto-controlled valve in the recirc path. One scheme we have is a recirc valve which is controlled at a setpoint based on total flow (recirc plus normal discharge). During normal operation the valve is closed. If the discharge valve is later closed the valve modulates open.  If pump is secured valve will go full open.  Then it will be full open at time of next pump start. If full-open recirc inteferes with motor starting time limitations, other controls may be required.

I'd support Electricpete's recommendation.  

Continuous recirculation is usually only done for smaller pumps due to the energy loss and upsizing required on the pump.  

One other alternative is a Yarway (or equivalent) ARC valve if your product is clean.  Under normal operation, the valve is closed and only opens under low flow conditions.  With an ARC valve, you avoid the need for another flow meter and stand-alone control valve manifold.

If possible, run the recirculation line back to the tank rather than directly to pump suction.

If you want to attack the problem with instrumentation, you should measure pump electrical horsepower (not current).  Set the power monitor to trip the pump out on low HP that corresponds to the deadhead condition.  If you want a 2nd line of defense, back up the HP monitor with a thermocouple on the pump that will trip it out.  The thermocouple should not be the primary trip, because HP is much quicker to respond.

All that being said, proper system design such that the pump won't be deadheaded in the first place is better.

Hi,

Thank you guys for your advise.

Butelja, I have not much knowledge about instrumentation, what do you mean exactly with electrical horsepower. If the pump is running against a closed valve -> head = 0 m. If I look at the centr. pump curve,then I see that the power usage isn't at its maximum, when head =0 m.
How do you measure "electrical horsepower"?

Thank you in advance.

Scyzoo

Scyzoo, you would trip the pump on low amperage to do what Butelja suggests.  

Determine what is the minimum flow you want to trip the pump at, estimate the Hp and then determine the corresponding amperage.

I havnen't heard of the approach of tripping on low motor power, but it seems like it might work.  It is a little disruptive to the system and may not be an acceptable alternative at a remote location where no-one is availabe to reset the trip. Especially if normaly fluctuations in demand might frequently take you into the low-flow area. Maybe in this case if you want to use the trip you can give thought to incorporating an ability to remotely restart.

One thing you would have to be careful of when setting your current relay... the relation between current and motor output power is fairly linear at high power (near full load), but woefully non-linear at low load. A low-speed motor may have 50% current at zero percent load.

Another thing that comes to mind... the shape of bhp vs flow curve was discussed in recent thread Thread407-27055 .  To my understanding, single-stage axial flow pumps will draw the highest current at shutoff.  Perhaps in that case instead of undercurrent, we simply adjust the overload setpoint down lower?  Other varieties of pumps may have differing shapes of bhp vs flow curves which will require individual consideration when setting your relay.

Electricpete & TD2K,

As you rightfully mentioned, the current vs. horsepower relationship at low loads is very nonlinear.  This is the reason for measuring true power [V*I*sin(phi)], not just current.  Current monitoring does not work well at all for low power cutoffs.  One manufacturer's device for measuring power can be found at: http://www.loadcontrols.com/.  An article on power monitoring from another company can be found at http://www.flowserve.com/pumps/technical_lib/index.stm#.

Something you may want to take a look at...
http://www.rainbird.com/Commercial.asp?Page=Product&LinkID=D56E8FD1-A843-488C-9193-FD7774F3CB54&SubCat=F8892584-79CE-4C05-A727-CEC3C16CA389

Perhaps you'd like to analyze the following alternative to do this:
A flow detector switch (ANSI code: 80) wire it as NO (normally open - closed with good flow).
Install the switch immediately downstream of the pump (before the check valve or any other valve). This means you will use one flow detector for each pump in the same service.
In the control system, add a 10 sec time delay if no flow is detected for 15sec (or a loose wire) the pump will trip.

Note: depending on how critical the service is, you'd like to have 3 switches installed for each pump. and do a 2 out of 3 logic voting in the control panel.
Hope this helps.
saludos.
a.

If to control on instrumentation I feel it is better to have flow transmitter rather. Transmitters based on calorimetric principle are very cheap and work well.

Is the power expressed by VISin(phi)or VICos(phi)?

Regards,

Quark,

OOPS, my mistake for posting too quickly.  Yes, power is expressed by V*I*COS(phi), NOT SIN(phi).

Additional information on power monitoring is at: http://www.loadcontrols.com/products/pmp_25_products.html.  

I prefer ARC Automatic Recirculation valve( mechancial automatic valve) to maintain the minimum flow which is simple and automatic in nature. It is not advisable to use instrument oriented  flow control for this service since there is a possibiliy of instrument malfunction.

Simon