BLDC pump suddenly runs dry...what happens next?

[Flyback]

In our Tulip field irrigation system, we have 3 phase BLDC water pumps. They spin at 8000rpm to pump water. (that's 400Hz bridge transistor switching frequency).....this is the maximum frequency that the bridge can switch at, and so even when the pump suddenly runs dry, the BLDC speed cannot increase....therefore, how does the BLDC inverter controller know that it must reduce the PWM duty cycle of the bridge transistors when the pump suddenly runs dry...after all, the motor speed does not increase?

 

[alec_t]

Can you add a flow sensor?

 

[ronv]

Current goes down?

 

[4pyros]

Back EMF

 

[crutschow]

Why do you think it has to reduce the PWM duty-cycle? The speed of the motor is determined by the frequency of the drive and is not much affected by the load.

 

[Flyback]

YES BUT IF NOT ENOUGH DUTY, THEN NOT ENOUGH CURRENT TO HAVE THE TORQUE TO ACTUALLY SPIN AT THE REQUISITE FREQUENCY (sorry about caps, my caps button gone)

 

[alec_t]

I'm not clear what you're getting at. If the pump goes dry surely the requisite frequency is zero, i.e. stop the pump altogether to prevent bearings seizing up?

 

[shortbus=]

I'm not clear what you're getting at. If the pump goes dry surely the requisite frequency is zero, i.e. stop the pump altogether to prevent bearings seizing up?

I'd be more worried about the impeller seals than the bearings.

 

[Flyback]

yes , you want zero speed when it runs dry...but how do you achieve that.?.......

 

[tcmtech]

A flow and or pressure sensor that shuts off the power.

Seriously?

 

[kubeek]

If you want to do it just inside the bldc driver, then measure anything you can with the pump dry and with the pump floodedl over the whole load range. Then pick something that will let you decide whether the pump is dry or not.
I doubt anyone can give you anything but theoretical behavior. The reality will allways be different, so go measure with the specific pump, motor, controller and load.

 

[BobW]

The ones we used to use on our projects, would sense current. When they run dry, there's a big drop in current. Pretty basic stuff. If the current drops, turn off the power. Of course the bearings and seals must be designed so that they can run dry for whatever amount of time it takes to get it shut down. I'm actually surprised they're using BLDC motors on something like this. It seems excessively complex for a water pump.

 

[alec_t]

But with a BLDC drive at a fixed frequency will the current change with the change in pump load? I would guess not significantly; hence the suggestion to use a flow sensor.

 

[Reloadron]

I would be in the group that suggest the use of a basic flow or pressure sensing switch. There are considerations depending on the application. For example at start up there is obviously no flow till the pump gets up to speed and you certainly don't want your pump to immediately shut down if an air bubble passes through so some small delay is likely in order. Without knowing some details like pump discharge pressure, pipe diameter and normal flow rate it is difficult to be more than generic in suggestions. Anyway, I would be considering a flow switch as a sensor similar to some of those I linked to.

Just My Take....
Ron

 

[4pyros]

Its done without sensors all the time using feed back from the motor, I just dont know how error prone it is.

 

[Reloadron]

Its done without sensors all the time using feed back from the motor, I just dont know how error prone it is.

I agree with that although I seldom used those methods. My reasoning runs a little along these lines. When a pump driven by an electric motor is moving water it draws current, that is a given. How much current is going to be a function of how much water the pump is moving and to what line pressure at the outlet. In this case the motor is a 3 phase motor so while measuring current isn't difficult it really isn't all that easy. I want to keep this simple, as simple as possible anyway. I just don't see back EMF as a viable solution for an application like this. In this case with the motor being driven by a VF (Variable Frequency) source I can't get my head around using back EMF or any other feedback from the motor.

This is why I would suggest using a simple flow switch as the sensor. There either is or is not flow exceeding a preset number. It's a yes or no thing with little room for error. I really don't care what the rate of flow is, just if the flow exceeds a predetermined set limit.

Just My Take
Ron

 

[JimB]

I am not sure if anyone has said this already but just use a flow sensor.

The flow sensor will protect against:
No water
Closed inlet valve
Closed outlet valve
Blocked filter

JimB

 

[alec_t]

@ 4pyros

Its done without sensors all the time using feed back from the motor

Can you point to an example, where the motor is a BLDC type driven at a fixed frequency?

 

[Flyback]

Sorry we don't have flow sensors, and its too complicated to install them in the water tanks...this is just a water pump which drops into the tank, and pushes water along.

 

[Reloadron]

Sorry we don't have flow sensors, and its too complicated to install them in the water tanks...this is just a water pump which drops into the tank, and pushes water long.

So it's a submersible pump? Now you mention that? What's the part number for the pump drive? You wouldn't install a flow sensor in the tanks but anywhere in the pump discharge line. You are telling me there is nowhere in the discharge line to install a sensor? Nowhere to install any of the types of sensors I linked to?

Ron