In-tube Velocity Limit to prevent Erosion 3

[thermmech]

Greetings!

I have found some recommendations on maximum in-tube velocities for exchangers (no clarification whether straight tube or U-tube construction) in order to prevent erosion.

They come from Wolverine Tube Heat Transfer Data Book. For water service and low carbon steel, the limit is 10 ft/s. That seems awfully low to me as it yields in-tube rho-v2 of about 6000 lb/ft-s2.

I would appreciate any comments or suggestions for these velocity limits.

Thank you.

 

[chief]

10 ft/sec is correct for this material and fluid medium. YES , IT IS LOW, BUT IS THE EQUIVALENT OF APPROX 1000GPM VIA A 6 INCH PIPE.

Offshore Engineering&Design

 

[MJCronin]

thermo..

The HEI Guideline contains velocity maxima to be used in the design of steam surface condensers

http://www.heatexchange.org/pub/default.htm

As i recall, the max recommended velocity for SS materials were greater than that of the brasses. Titanium was higher than that of the SS materials.

Of course, you must now realize that the low velocity limit for carbon steel materials is the reason why many designs end up with SS tubes....

-MJC

 

[davefitz]

As I recall, erosive limits for normal pipe geometries are :water in CS < 20 fps
water in 1% cr alloy steel < 27 fps
water in stainless steel < 75 fps

some particular locations have a higher erosive propensity than others, and of course, pH, chlorides, TDS, O2, etc have a role.

The typical "economic" pipe size for piping systems and pumps was estimated at 10 fps over 30 yrs ago ( assuming steel piping).Perhaps that is the basis of the 10 fps that Wolverine used.

 

[thermmech]

Thank you all for your comments!

My experience with air-cooled and S&T heat exchangers is that tubeside in-nozzle rho-v2 should be limited to 4000 & 6000 lb/ft-s2 respectively due to danger of eroding tube ends.

This goes in line with the intent of TEMA paragraphs under RCB-4.6 IMPINGEMENT BAFFLES AND EROSION PROTECTION as they strive to "minimize erosion of tube bundle components at the entrance or exit areas." Also, RCB-4.624: "Consideration shall be given to the need for special devices to prevent erosion of tube ends..."

Back-calculating of water velocity from rho-v2 = 4000-6000 yields 8-10 ft/s. However, those water velocities do not seem applicable for in-tube application.

In once-through steam generators rho-v2 values are far in excess of 6000 lb/ft-s2, and that is for in-pipe wet steam.

How do you apply velocity recommendations for pipe to tubing (yield stress for tubing is lower than for pipe).

I would appreciate any other comments or guideline sources you might have.

 

[davefitz]

Max erosion of S+T HX is at tube inlets at tube sheets due to the vena contracta effect: the actual max velocity at the center of the vena contracta is 1.56 times the average bulk velocity, and this is roughly matched at the recirculating eddies in this area and which contact the eroding tube section. This can be avoid at the tube inlet by providing an inlet bell mouth extension, but is rarely done; instead, a lower bulk velocity is chosen. So 20 fps/1.56 = 12.8 fps max bulk velocity for CS . On CS, it is the erosion/corrosion of the magnetite layer that causes tube failure, and this magnetite layer is related to the water quality ( O2, pH, TDS)

This type of erosion corrosion can be essentially elimiated by switching to the combined oxygenated feedwater treatment method: using a pH of 8-8.5 and 50 ppb O2 and 0.1 uS conductivity will allow a stable layer of protective hematite to cover teh tube, and this avoids the erosion.

 

[vesselguy]

Good topic.

It would be good for all engineers if there is a study report that show the experimental data on in tube errosion effect caused by fluid velocity, stream particle size and tube material. I know all these rules of thumbs from Marks Handbook and various sources were from experiences dating back to God knows when. But for today's work on optimizing equipment design, we need to have those results verified in a study.

Interesting that thermech points out OTSG have high pV2. OTSG is specified to have a max pV2 of 25000 lb/ft-sec2; that's 20 ft/s for a mixed flow regiem of steam and water. I have never seen or hear of any results to justifiy if the 25000 is a good number or not. Does anyone know?

 

[davefitz]

For OTSG, the mass flow G lb/hr/ft2 is used more frequently than rho V ^2. The relevant limit or design value varies all over the lot, is a function of mfr, circuit design, imposed heat flux , tube design ( smooth bore, rifled, vertical, inlcined), max possible overheat, variable pressure or constant pressure ops.

In a radiant furnace with a flame temp of 2500 F, the imposed radiant flux is the governing boundary condition with a max achievable overheat equal to the flame temp of 2500F ( ie burn out), thus it is critical to ensure loss if inside film heat transfer coef due to dry out or DNB either does not occur, of limit its effects so that the tube overheat does not exceed ( 950 F) mean metal temp. For a smooth bore vertical tube, that might imply an average circuit mass flow of 1.5-2 E6 lb/hr /ft2 ( recognizing that it is the overheat of the "worst tube" that governs circuit design). For a vertical rifled tube, variable pressure, that may imply an average mass flow of 0.65-1. E6 lb/hr/ft2, which allows for a positve , or natural ciculation, circuit thermal hydraulic sensitivity characteristic.

For HRSG's or convective evaporators with a flue gas temp of only 800-1000 F, the max overheat is not high enough for a prompt burnout; cyclic fatigue or tube max allowed deflection may govern, and dryout is permitted even in smoooth bore vertical tubes ( assuming holy water feedwater quality); DNB is not likely to occur due to the low max heat flux. Refer to correlations by Doroschuk and Kon'Kov . Low mass flows of below 0.65E6 lb/hr/ft2 are OK

 

[vesselguy]

Davefitz,

I thank you for your response.

So what you are saying is due to high local heat flux rate it is possible for DNB, or even transition to dry film boiling, could occur and so you want high stream velocity (relates back to mass flow rate) to get rid of any steam layer that might develope on the heat transfer surfaces. Am I reading this right? If so, this is the same principle as zero fouling in shell and tube heat exchanger thermal design, which I am a proponent of. However, this still does not address the proper value of kinetic energy parameter (pv2) to use to prevent errosion. Know what I mean?

 

[davefitz]

I suggest you browse thru the papers by KWU Kiefer, Kohler, and Hein in the Intl J multiphase flow ( 1982-1997) and heat transfer texts edited by Sadik Kakac to understand the commercial limits on 2 phase flow heat exchangers.

 

[thermmech]

Thank you all for you comments!