Reverse Operating Pumps Part 3

It occurred to me last night that Engineers overall lead a charmed existence. It was while I was (re)watching the 2011 movie - Margin Call (a movie I highly recommend and will return to in a moment).

Now of course I'm biased and hence you should treat my analysis with a deep and abiding skepticism. However by my way of thinking I was lucky enough to grow up with sufficient nurturing and talent that I had the choice after completing secondary school of becoming a Doctor, a Lawyer or an Engineer. For me the choice was pretty easy:

• If I became a Doctor not only would I have to deal with various human excretions but in the end all my treatments would ultimately result in failure (we all die in the end).
• If I became a Lawyer not only would I have to spend an excessive amount of time writing English in a way nobody else could read or comprehend, but statistically I'd lose half my cases.
• So in the end the choice wasn't really a choice at all. Become an Engineer where I get to design and make cool things that people didn't even realize they needed. Sometimes we even make the world a better place.

Back to the movie Margin Call. In it a young analyst happens upon data that shows the financial market is about to implode. It is revealed in the course of the movie that the analyst is in fact an Engineer who ended up in Wall Street because the "money here is considerably more attractive". Ultimately his company uses the information to do some very bad things. The analyst gets promoted and hence his fate is sealed to chase Mammon forever. In his place I like to think I'd have made a very different choice.

With that thought, welcome back to Part 3 of this article series on the topic of reverse operation in pumps. In this installment, we will take a look at a the effect of a degrading isolation valve over time and what kinds of problems it might create. We'll also take a quick look at what happens if you drive the pump with reverse rotation due to miswiring the electric motor or commanding the wrong VFD rotation.

When does isolation valve leakage become a problem ?

Here we are considering the scenario where the isolation valve (or NRV) on the pump discharge side becomes eroded and damaged over time. This can simply be due wear from the number of cycles the valve sees as well as erosion and corrosion. Whatever the cause it can be expected that at some point the valve will leak such as the butterfly valve seat failure shown below.

Assuming the fluid being handled isn't toxic or hazardous, small amounts of leakage back through the pump are generally not a cause for concern other than representing a loss of system efficiency. What we are interested in is predicting when the leakage will have a negative impact on the pump.

To determine that we will need to combine two pieces of information that we've already utilized in Part 1 and Part 2 of this series.

The first piece of information is to determine the breakaway torque of the pump. This is the torque required to cause the pump to start rotating from a standstill. Typically the pump manufacturer can (and should) supply this. It will vary depending on the pump design, suction pressure and type of mechanical seal plan utilized.

If the information on pump breakaway torque is unavailable you can utilize the standard pump speed-torque curve below taken from?Hydraulic Institute Standard 14.3. Keep in mind this is a general guide and you should not treat it as an exact value for your specific pump. For the purposes of this example we will utilize this curve, meaning that the pump breakaway torque is 15% of full speed torque.

The final way you can obtain the breakaway torque is to perform a test of your pump. You'll need a means of accurately applying and measuring torque to do this. In a pinch, a calibrated torque wrench connected to the end of the pump shaft will work.

The second piece of information you will need is a four quadrant curve for the pump or one of a similar specific speed. Since four quadrant curves are extremely expensive to make, most likely you will need to settle for one of the publicly available curves for a similar specific speed to your pump. This of course introduces additional potential for error. For the purposes of this exercise we will utilize the four quadrant curve used in Part 1. This is shown below.

What valve leakage will make my pump rotate ?

So given for this exercise we need 15% of full speed torque we can establish the flow required to set the pump in motion in reverse. There is no 15% iso-torque line, so we'll need to extrapolate using the 25% torque line. The corresponding reverse flow at the 25% torque line is 58%. I have marked this with X.

By linear extrapolation, the reverse flow required to generate 15% torque is

58% * 15/25 = 35% reverse flow

I have marked this with S.

Ok now my pump is running backwards - is this bad ?

As is often the case, the answer is "It depends"

In this example, once we reach 35% reverse leakage flow (point S), the pump will start to rotate backwards. Applying (and assuming) a line of constant head at this point (pink dotted line), we can see that the pump will accelerate until the zero torque line is reached resulting in a reverse speed of 47% of normal forward speed. This point is marked with F

If your pump is a single stage OH2 or BB1 with a classically stiff rotor, simple single seal plan, and rolling element bearings, this probably isn't a concern. However if you have a multistage pump or a pump with sleeve bearings or a double seal plan it may well cause significant damage for one or more of the following reasons:

1. Multistage rotors rely on the Lomakin Effect to provide adequate stiffness at the wear ring and bushing locations. At slow speeds and low head, the effect may be insufficient to prevent contact between the rotor and stator. The resulting surface contact can cause significant wear or even seizure of the pump if it continues for more than a short time.
2. Similarly with less Lomakin Effect the pump rotor natural frequency will be very low and if it coincides with the reverse running speed the resulting rotor motion can quickly damage the pump.
3. If the sleeve bearings are designed for forced lubrication, they will likely fail rapidly in the absence of this. (Sleeve bearings with oil rings - either as primary or backup lubrication, will probably be ok). The same problem exists for any tilting pad thrust bearings utilized in the pump.
4. Seal systems which require active circulation (plan 54) or seal systems that rely on pumping rings may not work as intended and seal failures can result.

Depending on your system, pump and the resulting risk profile, you should consider reverse rotation detection for critical service pumps.

What happens if I wire my electric motor wrong and run my pump backwards ?

Here we are consider a situation where by accident the driver of the pump runs with reverse rotation. Typically this happens if the 3 phase electric motor has 2 phases swapped or if the VFD is set for the wrong rotation.

(This is why when first commissioning a pump or after any major work on the electric motor you should always check the electric motor for correct rotation while uncoupled from the pump).

If you are unlucky enough to skip the step above and your pump uses an impeller that has a thread to attach it to the shaft or uses threaded line-shaft couplings, then odds are you will seize and severely damage the pump as soon as the driver is energized. This is because the driving torque direction is reversed causing any threaded connections to unscrew. Try NOT to be the person who skips the uncoupled motor rotation check and causes this easily avoidable failure.

For pumps that use keyed or splined connections for the rotor parts, the pump will operate with reverse rotation in sector E of our four quadrant curve.

It is a common misconception that powering the impeller in reverse will result in reverse flow. It won't in a normal system with a NRV, however the pump performance will be severely compromised. To see how much let's look again at the four quadrant curve we used earlier.

We start with -100% speed on the Y axis since we are running at full speed in reverse. Drawing a horizontal line we can see that the pump develops:

• 67% head at shutoff
• 50% head @ 23% of normal BEP flow. Torque is 105% of normal
• 25% head @ 45% of normal BEP flow. Torque is >135% of normal

From this we can see the pump operates in a very constrained and inefficient way which will quickly overload the motor. If we were to compare the forward rotation and reverse rotation performance HQ curves, this would be the result:

So if you are measuring pump head and flow that is much lower than expected together with high motor power, check the rotation to verify it is correct.

That's all for Part 3. In Part 4 I will cover how to estimate the performance of a PAT (Pump As Turbine) when only the standard pump HQ curve is available as well as any other scenarios I receive questions for. Until then all comments, questions and criticisms are gratefully received.

Obsesio?positivum