NPSH – Discover Dredging

Graduation Of Alex De Rooij: Pumps Actually Fly Like An Unusual Airplane

Alex de Rooij receiving flowers from Suman Sapkota for his graduation

When you hire a carpenter, he repairs everything with a hammer. So, what happens when you ask an aeronautical student to solve some issues in a dredge pump? He models the pump as a badly behaving airplane. And with success, Alex de Rooij joined our company as a graduation student and recently graduated on the topic of ‘Numerical Study on NACA Profiles Sensitivity in Dredge Pump Impellers’.

The normal procedure for designing pumps is relatively straight forward. Set the performance specifications and try to hit that mark with the simulated behaviour from an iteratively improved design. This is well documented and I’ve been writing about this process before.1

Pump design workflow (inspired by Suman Sapkota)

One of the design parameters is the NPSHr. This is basically the amount of absolute pre-pressure the pump requires to operate.2 The system and operating conditions will result in a certain available pre-pressure: NPSHa. When the NPSHa drops below the NPSHr, the pump will experience cavitation at the low pressure side of the blade. The flow of the medium will be disturbed and the performance of the pump will abruptly crash. There is some warning. Operating near the NPSHr, there will be an incipient cavitation where the vapour bubbles start to form, but do not cause any issue. The imploding vapour bubbles may be detected audibly for the trained listener.3 Next will be a stage on the NPSHr, where the bubbles get larger and they loudly implode. At this stage, the pump sounds like it is tumbling nuts and bolts inside. These imploding bubbles will definitely cause damage to the impeller. At last, working below the NPSHr, the bubbles will become so large, they will block the passage between the blades. The result is an immediate drop in delivered head.

Explanation of NSPHr, NPSHa and cavitation

The flow of the medium through the impeller can be simulated in a CFD program. Specifically for impellers, you will need to model a rotating frame of reference. And that is the usual representation of the results. However, with mathematics being one of the most powerful tools invented by humanity, we can have an alternative view on these results. We can cut the impeller along a radial and stretch open the meridional passage and the blades to a row of foils. And that is exactly where our young aspiring engineer comes in. In literature, the blades in the impeller are modelled having a constant thickness. But, Alex has been investigating what the influence will be when we model the blades as foils. Selecting a proper profile makes the blade less sensitive to stalling.

Conversion of axial view to blade to blade view

Alex, thank you very much for your work here at Damen. We’ve learned the influence of certain profiles on the performance and geometry of the pumps. You have the right mindset to pass your time at the TU Delft and graduate successfully over there also. And whenever you have some days of the month left after you spend your allowance, know that we can give you a warm reception at our office.

Alex working at the reception desk at Damen Dredging Equipment

References

Our Interview About New Pump Designs In The Latest Damen Nieuws

Headline of our interview in ‘Damen Nieuws’

Another magazine dropped on my doormat, albeit a digital edition of ‘Damen Nieuws’1. The internal magazine for Damen colleagues. It featured an article with Suman Sapkota and me. Suman is our pump design specialist2 and at a Damen wide R&D convention he presented a poster on his pump design workflow within Damen Dredging Equipment. This caught the attention of the editorial board and we were interviewed on what we actually do for a living. Although we can’t share the exact details of the article or the poster, it is still an interesting message that we can highlight here.

The design of a dredge pump is based on the required specifications(1). The most important properties of the pump are: efficiency, NPSH, wear and ball passage. The first important property we try to fix is the ball passage3. We do use our own geometry generator(2) that assist us in creating a pump with a big ball passage. Unlike normal pumps, dredge pumps have to cope with debris and boulders that have to pass the impeller. The bigger chunks that can pass, the more uptime the dredge will have. Once we are satisfied with the geometry, we feed this through a file format converter(3). The resulting 3D file can be used on several platforms. This will enable us to create the digital solid for the engineering4, but it also gives us the negative volume, also known as fluid. Then to do mathematical operations on the digital fluid, we have to divide the volume into tiny cells. This process is called meshing.(4) When the mesh is available, the fluid flow through the mesh can be simulated with computational fluid dynamics.(5) All the fluid properties of every cell are calculated and the results are shared with the adjoining cells. This can be repeated until all properties of the cells don’t change very much anymore, a stable solution. Integrating all the properties of the cells give the resulting performance of the pump.

Balancing the four dredge pump performance properties

The estimated performance can be evaluated against the four properties.(6) The head times the capacity divided by the power required will give the efficiency. That is one of the items we wanted to know, as it relates to how much fuel will be consumed. The other parameter obtained from the CFD is the NPSH, or roughly: the suction performance. Wear cannot be estimated yet, but we are working on that2. Although the calculated turbulence might give a clue what wear to expect. If the properties are not satisfying our requirements we make an iteration in the geometry for improving the performance. However, changing the geometry will usually result in a smaller ball passage. If the parameters are OK to our requirements we have a pump design.(7) Manufacturing it is a completely different game.5

The design process of the dredge pump takes quite some effort and we are continually looking to improve the workflow6. Eventually we would like to be able to cater for all special requirements each individual customer might have.

Working for a dredge manufacturer, I am happy we design and produce our own pumps. It gives us the confidence, that when we supply dredges, they are as we like them to be. Another benefit is in discussions with the customer. It is easier when we can sit at the table as experts on their equipment assit them in finding a solution for their dredge.

Pump experts immersing themselves in checking the design of their pumps

References

When does your pump suck?

One of the key process indicators for the performance of your dredge pump, is the capability to work with low suction pressure. The parameter involved is called ‘Required Net Positive Suction Head’. Which translates more or less to: ‘the head value at a specific point required to keep the fluid from cavitating.1’ Effectively, this is the extra pressure above the vapour pressure. From the pump inlet to the blade, there still is a pressure drop. And the geometry and the form of the blade influence this pressure drop. The operator will notice this as when the blade wears down, the pressure drop becomes greater and the required suction pressure goes up. Resulting in less performance and less production. Regular inspection of the pump will warn the operator of prospective deterioration.

Normally, the measurement of the NPSHr requires a valve in the suction pipe and a valve in the discharge pipe to control the flow. Every time you want a data point, you have to adjust both valves and iteratively return to the same flow conditions, albeit with a different suction pressure. This usually takes a lot of time and one hour per data point is not uncommon. Klaas Slager presented an alternative method at the CEDA Dredging Days2. His method is more suitable for testing the NPSHr as installed in a dredge. It does not involve the dredge valves and is quicker to execute. It is optimised to check if the NPSHr wanders off nominal and thus will yield an indication on the condition of the pump. If the internal pressure drop increases, there is less differential pressure available in the suction pipe for the dredging process. Less concentration or less capacity, or less in the combination of the two: less production.

Instead of varying the flow conditions, he proposes to vary the pump speed. This will influence both flow and suction pressure at the same time. However, by cleverly applying the affinity laws and presenting the operating conditions in a dimensionless scale, the cavition is immediately visible. A quick post processing will reveal any wandering of the NPSHr conditions. As this can be implemented in the PLC and executed during start-up every day, the operator will receive a daily update on the suction condition of his pump and can plan actions accordingly. This will prevent unnecessary delays and downtime.

Worn down suction side of a dredge pump deteriorates NPSHr

This concludes my scheduled series of posts about the CEDA Dredging Days. There was much more to discover. The interactive session was fun. There were a couple of interesting presentations. And I’ve seen some innovations at the exhibition. So, I will write some more reports, although at a more leisurely pace of about once a week. Later on, the other promised topics will be covered3. I’ll keep you posted.

Tag: NPSH

Graduation Of Alex De Rooij: Pumps Actually Fly Like An Unusual Airplane

References

See also

Our Interview About New Pump Designs In The Latest Damen Nieuws

References

See also

When does your pump suck?

References

See also