Vortex Shedding and Tube Frequency Assumptions

[yamoffathoo]

A flow induced vibration analysis of a steam surface condenser uses the following methods to determine the propensity to vibrate:

1) how close the tube structural frequencies lie to the vortex shedding frequency
2) the maximum tube vibration amplitude at resonance
3) Connor's fluid elastic instability gap assessment

My question is, how are these assessments affected by the presence of single phase water droplets in the steam vapour? Could the presence of single phase drops change vortex shedding or structural frequencies?

 

[hacksaw]

Very little effect.

The general rule is that the forcing frequency must be outside +/- 20% of the tube natural frequency, for single tubes.

If you are dealing with tube bundles you have to deal with self-excited, fluid-elastic motion of the bundle. This is a non-linear effect and depends on the deflection amplitudes. ASME N1300 is a good source and is covered in texts that deal with flow-induced vibration.

 

[yamoffathoo]

Hacksaw, the reason I ask is that the inner core of tubes are fretting much faster than the outer rows of tubes. This contradicts the (Manufacturer's) assumption that peripheral tubes are under the most severe conditions: pure cross-flow at the highest velocity.

 

[hacksaw]

that maybe true for isolated tubes, but with a bundle, the entire assembly can be excited at much lower fluid velocities than expected for single tube.

there is a considerble body of work dealing with the problem, what standard is used for design: Heat Exchanger Inst, ASME, proprietary, etc.?

 

[yamoffathoo]

TEMA R and ASME are the configuration and design codes.
Among six similar condensers, only one tube bundle contains core tubes that are fretting at supports and, coincidentally, the same one has been running in both spray and tube condensing mode, while the others only in tube condensing mode.

I think there must be a correlation. Do you know of any work dealing with condensing vapour quality and FIV?

 

[hacksaw]

there is very little other than in some cases it tends to interfere with vortex shedding, but that is for a true mixture

direct spray on the tube bundle can mean several things.

if the spray was introduced upstream, and well mixed before entering the hex, ok you have "two phase", but if liquid is being directed against the tubes directly, the governing condition involves liquid phase at high velocity

most heat exchangers I've worked with did not use spray injection and in fact required premixing and baffle plates to prevent direct impingement

look at the spray nozzles, estimate the liquid phase exit velocity, identify if the spray nozzles are functioning and properly placed

simply jetting high velocity spray against the tubes without proper mixing is going to create all sorts of issues

vortex shedding in a liquid phase is a bit more complex than gas or vapor phase, not only because of the increased desity, but also because there is a tendency for the sheeding process to occur at a fraction of velocity normally associated with the strouhal condition

 

[yamoffathoo]

The steam condenser in question is a vertical, four pass, U-tube design with a spray header. The ring steam header is mounted inside of the ring spray header and both headers sit above the centrally hung tube bundle. The 10% quality steam either converges radially inward to the coils or radially outward to the spray curtain. Mixed spray/coil condensation should cause water droplets to get entrained in the coil flow.

The fundamental single tube frequency is 60 hz and the highest vortex shedding frequency (according to the Manufacturer) is 1/2 that (30hz). So either the liquid phase is lowering structural response or raising shedding frequency - or both.

Could this happen and why would it affect only the coldest, first pass coils deep inside the bundle?

 

[hacksaw]

With tube bundles, the first layer of tubes forms jets that are synchronized in pairs. These synchronized jets have a greater effect on the subsequent tubes. The gap velocity governs.

The damage described suggest that flow induced resonance is occurring at a lower velocity than predicted by the vendors calculation. Could be a number of reasons including the actual construction

 

[hacksaw]

added mass does have an effect but more importantly you also have what are called subresonant behavior. This has more to do with the vortex shedding process. There is a dominant component of the vortex shedding process which is in the flow direction and occurs at roughly one half the strouhal frequency.

 

[hacksaw]

surely the manf has taken into account the added mass of the water "inside the tube", if not then you've got a mis-applied design.

 

[yamoffathoo]

The assumptions were:
1) flow is in vapour phase, droplets are gravimetrically separated from the vapour
2) peripheral tubes are under the most severe conditions because as the vapour penetrates into the tube bundle, a portion of it condenses, reducing its velocity, the remainder moves upward
3) the weight of the film condensate on tube surface per inch length is not included in equivalent tube weight

The effective weight of the tube per unit length is the sum of contained fluid, the metal (4.6 x more) and displaced fluid (6 x less).

 

[yamoffathoo]

Hacksaw, There is an operating history detail that might help - the coil flow was constant and a small fraction of what was needed for the steam load. The spray flow modulated as required to efficiently maintain saturation pressure of approximately 300 psig.

Rather than droplet entrainment, could the result be an extremely dry, high quality vapour that arrives, with high velocity at the inner cold coils, driving the vortex shedding frequency up to the first fundamental for the tube spans just in that area of the coils?

 

[hacksaw]

check the steam velocity without the spray, i.e. a worse case condition, that will set your design limit

 

[yamoffathoo]

Both spray and coils have been designed for the same maximum steam flow - I have no issue with either mode of operation, only continuous, simultaneous operation of both - which is not part of the design basis.

Perhaps having too much condensing capacity and really coarse controls could lead to transient coil flooding? Could this lead to a 'steam cracking' phenomenon where plumes of vapour rise between core tubes that are coated with a thick descendiing film of condensate and rapid collapse sends shock waves through the tube bundle?

 

[hacksaw]

far more is going on that FIV or vortex shedding, have your supplier check the case for simultaneous case

good luck

 

[yamoffathoo]

My 'supplier' wants to sell me replacement tube bundles - no help trying to solve this operating puzzle.

If I understood the mechanism, then I could set limits on acceptable leakage past spray and coil control valves in all the other condensers. Running on sprays with leaky coil flow or running on coils with leaky spray flow could cause the same fretting degradation.