Need some reference on centrifugal pump design

[ecfedele]

Little bit of a disclaimer: if you take a look at my first (introductory) post, you'll see that I'm working towards aerospace engineering; therefore, this is leaning towards turbopump design, but for the time being I'll start out slow.

Basically, I'd like it if anyone could help me get a sense of the engineering flow for a centrifugal pump. I've already examined multiple books on the subject and consulted any number of web PDFs about it, so I have some exposure to many of the fundamental concepts (impeller velocity triangles, head, NPSH considerations, cavitation), but in a really disparate manner that doesn't amount to anything enabling me to start sitting down and working on a trial or first design.

I'm not asking for anyone to walk me through the calculations in a textbook, cookie-cutter or formulaic manner, but I'd like some explanation of the design flow in a way that helps me put all the formulas and concepts I've read about together into a way that would actually enable me to feel like I could go and crank out a design, however inefficient or idealistic.

Now, let me approach this from my desired angle, which is a rocket turbopump, so you can see the way I'm seeing the design and the conundrums I'm looking at. I know that a turbopump is (usually) a large centrifugal pump with an emphasis on light weight and high efficiency that is driven by a turbine, but from looking at the basic pump equations it appears that since density is a primary consideration, the theoretical, basic design shouldn't be all that different. The fundamentals are similar.

My main issue is the idea of the head from source and to exit of the pump and with the pressure requirement. Rocket engines don't usually bother with stating their flow requirements as a unit of head. They require that the pump can develop the designed chamber pressure with the required flow rate. Obviously, flow rate, Q, is a direct parameter of pump design, but not pressure. I asked around on Engineering Stack Exchange and found that the pressure is usually a related concept to the head (via the Bernoulli equation), but the head is what is throwing me.

In a rocket, the fuel tanks mount above the pumps which mount above the motor. This almost implies that the head requirements would be negative. The design of a pump that would function in that manner is the hardest thing I'm trying to wrap my head around.

For anyone who can help me work through all this, many thanks. It feels here like I simultaneously have a good grasp of the basic concepts but still know nothing to the point of having to be walked through those basics.

 

[billy_joule]

For a given density and gravitational field strength head and pressure are interchangeable terms. You can use whichever you like, just convert as required.

 

[ecfedele]

For a given density and gravitational field strength head and pressure are interchangeable terms. You can use whichever you like, just convert as required.

Which, as I said, I understand that element of it. But pressure generated from head and the head as the computed requirement don't appear to be the same thing. So let me know if what I've got so far is correct or even in the right ballpark.

According to Wikipedia, hydraulic head appears to be the combination of pressure head and elevation head, so:

$h = \psi + z = \frac{P}{\rho g_0} + z$

Now, for computation's sake I'll make an example where water needs to be supplied to a nozzle at 1MPa where the nozzle is situated one meter below the pump inlet. So, if my ideas are right, the elevation head ($z$) is -1m. If this is the case, is this head calculation correct? Or if I'm missing something, what would it be?

$h = \frac{1000000}{9806.65} - 1 = 100.972m$

What I'm much more worried about though, is how to translate the head and flow-rate requirements into an actual pump design. I know about some of the impeller geometry and velocity triangles, but most books on the subject are more pump selection guides than design texts, so I have no idea what the link is between the input requirements and the actual pump/impeller/volute design process. Can you shed any light on that?