With the slowing of the economy, I am finding myself with a little free time. One way I fill it is by looking at technical discussion boards and (occasionally) participating.
One caught my eye recently. A person was having problems with pump cavitation, and using an online discussion board was asking for solutions. It should be noted that the person was in operations, and when in operations, the ability to fix cavitation is quite limited (as opposed to design).
The person said all the “correct” things (NPSH calculation, HYSYS simulation, etc), and the discussion took off. And I found this amusing, because …
They were only talking about one way to cavitate a pump.
Are there more than one way to cavitate a pump? Well, actually there are three ways to cavitate a pump. Technically, I learned about all of them in my university days, but I doubt I understood them from a classroom perspective.
And over the years, I have developed what I think is a bulletproof definition of cavitation – it took me over 25 years to fully get my mind around it.
My definition … Cavitation is a two-step process. Step 1 is the introduction of or creation of bubbles. Step 2 is the iso-thermal (non-iso-baric) condensation / collapse of the bubbles.
The first way I learned from working in the oil and gas industry. It is the classical if NPSHA is less than NPSHR then the pump will cavitate. The basis is that the liquid is relatively close to the boiling point, and there are enough hydraulic changes in the liquid that it boils isothermally (by changes in pressure). This is the basis for the CREATION of the bubbles. This is the method of cavitation that can be managed by calculation, and the one most engineers understand (because it has an equation). Unfortunately, this also seems to be the ONLY source of cavitation that most people consider. (ASIDE: with very few exceptions, I think this is essentially the only way to cavitate valves, flow meters, and fittings).
The second way to cavitate a pump is a method I learned from working in the mining industry. Many pumps have the liquid BELOW the pump, and the act of lifting the liquid created a vacuum inside the pump. Since many mining fluids are water or water-based, the pumps use packing instead of mechanical seals. Naturally, the packing will leak a little, but the vacuum would allow air (and air bubbles) to enter – on the inlet side of the pump impeller. This is one basis for INTRODUCTION of bubbles. While I think it is possible to use calculations to estimate cavitation tendencies, it is not a nice calculation (like the NPSHA is less than NPSHR equation).
The third way to cavitate a pump is a method I learned while commissioning a new pumping system. We discovered that when our tank reached a certain level (above the pump LSLL, and WELL above the calculated minimum level for NPSHA cavitation) our pump started cavitating. And it was flowrate sensitive. Because it was an atmospheric tank, we were able to open a hatch and watch what was happening … at a certain level a vortex was forming on the liquid outlet. This allowed gas (not bubbles, but actual gas) to enter the pump. Some was condensed as the pressure rose across the pump, thus the cavitation. Again, this is cavitation caused by the INTRODUCTION of bubbles, but the bubbles are from the process. Again, this is difficult to predict by calculation.
So three ways to cavitate a pump, but because we can only perform a calculation on one, that is the one we tend to focus on.
The discussion board? I made a comment. After my comment, the discussion drifted back to the NPSHA/NPSHR calculation. <shrugs>