cavitation, as it is usually defined, is not common in that service, suspect that you are dealing with line velocity being too high, particulate erosion, or corrosion common to raw water services

For lake water erosion/corrosion --I vote for 316 stainless -- schedule 40S --- or schedule 80S if you got it

This may be the cheapest solution also....

MJCronin
Sr. Process Engineer
Venture Engineering & Construction
www.VentureEngr.com

hacksaw,

I have attached a photo taken of the inside elbow extrados and, although fuzzy, you can see pitting - some through wall.

The strainer backwash lines are connected to a common header which discharges to atmosphere and provides a good siphon, dropping the individual backwash lines below atmospheric pressure.

Vapour bubbles form at the valves and collapse at the first elbow where the static pressure recovers, generating a continous growling noise.

• http://files.engineering.com/getfile.aspx?folder=9e5e571a-1c8c-417f-9b09-7c

The 7000 usgpm continuous flow strainer manufacturer specified 416SST body and Teflon seat material for the 3"nps 150# backwash discharge butterfly valve. Perhaps teflon coating is the way to go?

looks like bacterial pitting,

if you specify the pressure temp and flow rates, you should get responses that will help you resolve the problem

Will agree that looks like MIC probably due to lack of disinfection.

It is not likely that this is cavitation as your system is under pressure. The valves are typically used to throttle down the pressure before discharge.

Consider adding some chlorine to stop the MIC.

hacksaw and bimr,

I am not familiar with growling bacteria or mircrobes...

MJCronin,

I agree with stainless, a lining could disbond.

I just finished reviewing "Erosion of Concrete in Hydraulic Structures" ACI 210R-93 and am intrigued with the idea of supressing cavitation through aeration using a radial array of vacuum breakers.

Any thoughts?

Stainless steel is NOT the answer. MIC will chew up stainless as well:

http://www.nickelinstitute.org/~/Media/Files/Techn...

Here is what MIC looks like"

https://www.google.com/search?q=mic+corrosion+imag...

The corrosion look familiar doesn't it?

You do not have cavitation, you have MIC. If you want to end it, you will have to add chlorine to your water to kill the microorganisms that are responsible for the MIC.

Google "MIC":

http://www.corrosionclinic.com/types_of_corrosion/...

The "growling" is simply the sound caused by the pressure reduction being taken across the valve. It is not the cause of the problem.

I reviewed the paper that you mentioned. Here is a link:

http://civilwares.free.fr/ACI/MCP04/210r_93.pdf

There are a number of reason that your proposed solution would not work:

The system is under pressure and you will not have cavitation. They are talking about gravity spillways and flumes, not pressure pipes.

"A good way to avoid cavitation erosion is to make u large by keeping the pressure high, and the velocity low".

A vacuum breaker will also not produce fine bubbles. You would need some type of bubble diffuser.

You should reread the paper again, specifically section 4.3.2 (Bacterial action) where MIC is discussed.

bimr,

Thank you for all the links, however:

1) I see absolutely no common features (except a hole) in the photographs
2) there is no sign of biological matter in the elbow, just a scoured and pitted moonscape
3) MIC could not survive in a turbulent strainer backwash discharge pipe - it needs normally stagnant or very low flow conditions to thrive
4) When the backwash discharge valve opens, the inner volume of a filter element is exposed to the sub-atmospheric pressure in the siphoning discharge header, causing vapour pockets to form and highly turbulent flow, conditions anathema to MIC
5) gravelly, growling sounds from the elbow are audible symptoms of cavitation
6) dousing the backwash flow with chlorine as it is pumped back into the lake would, I think, be unpopular...

Common features of MIC are "scoured and pitted moonscape", not biological matter.

It is not possible to see the biological agents that cause MIC, just the results, pits, which are the result of biological attack. M stands for microorganism which would indicate that the biological matter is quite small. Studies have shown that a millilitre of lake water commonly contains nearly 10 million microorganisms.

You say there is no stagnant water, but then "When the backwash discharge valve opens."

What is cavitation? Cavitation manifests itself audibly as a grinding noise, a noise that closely resemble gravel being moved around in a cement mixer. It can be heard at pump inlets and control valves. It is due to the fluid being vaporized because of low pressure and then suddenly collapsing due to high pressure produced by a pump impeller for example or the increase in pressure that occurs at the outlet of a control valve. Where in your system is the higher pressure that should be causing the bubbles to collapse?

Here is a recording of cavitation:

https://www.youtube.com/watch?v=1Lbxtjfdat4

Why don't you put a pressure gauge on the filter to determine the pressure. One would be highly skeptical that a "vacuum condition" and cavitation could exist. A pump rotating at thousands of rpms will not cavitate unless you are pulling more than 15 feet of water.

One would think that dumping backwash water into the lake would also be unpopular.

bimr,

Why would putting the silt and organics back in the lake be unpopular?
The product information sheet is attached for your edification.

• http://files.engineering.com/getfile.aspx?folder=f8714596-1800-467b-8ba5-bb

It is not legal where I am located to discharge filter backwash to the environment. It depends on the regulations where the facility is located. Here is an example:

"SEPA considers that routine discharges to the water environment from the water treatment works are ‘activities liable to cause pollution’ under Section 20(3) of the Water Environment and Water Services (Scotland) Act 2003"

http://www.google.com/url?sa=t&rct=j&q=&am...

bimr,
To summarize then:
- cavitation cannot occur in 'pressure pipes' attached to strainers
- fresh water MIC builds transparent tubercles in turbulent intermittent flow
- chlorine laced backwash will solve the problem but is probably illegal
?

I've been watching this thread and I go back to hacksaws comment.

The thing missing here is any idea of what your flowrate is as you discharge this product and hence your velocity. Also how many minutes in an hour does this discharge - many filter systems work on both a differential pressure to initiate the cleaning and a regular flush every hour or four hours or once a day? what is the "average" time and duration for your system.

I haven't seen any of these details except a discussion, very useful though it is, on MIC

The "growling" sounds to me like all the dirt, small creatures, gravel etc from the filter hitting your elbow at high velocity. The differential pressure in these systems can be very high as the filter sits at a few bar pressure, but the 3" outlet basically at about atmospheric.

If this line is low pressure why don't you use PE? It's used quite a bit in slurry lines which essentially is what this is.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

LittleInch,

I appreciate your interest in the strainer operation, however, I feel no need to discuss further the improbability of bacterial infestation.

I modelled the backwash lines from three strainers and demonstrated that the elevation drop in the backwash header alone (no dynamic pressure gradient) can remove flows of 275 usgpm per strainer when all three MVs are open up and 400 usgpm when only one strainer MV is open. This equates to velocities of 7 to 10 ft/sec, respectively.

The backwash flow through each strainer element and MV must keep up with this siphoned flow rate or vapour pockets will form at surface irregularities, move downstream and spontaneously collapse (cavitate) at locations where the static pressure increases sufficiently (elbows).

Aerating the flow downstream of the MV using vacuum breakers could introduce too much air and reduce strainer efficiency. Buoyancy forces and secondary vortice flows would make it impossible to ensure the air arrives at the extrados where most of the cavitation is taking place.

I think sch80 stainless is the most cost effective way to prolong the replacement period.

Bambie,

If what you are theorizing is correct, the simple and inexpensive solution is to throttle the backwash discharge with a valve or fixed orifice at the end of pipe discharge point. With this fix, you would maintain pressure in the backwash discharge pipe and eliminate the possibility of cavitation.

That would be the fix to remedy the system problem if your theory is correct. However, the theory is improbable at those fluid velocities.

bimr,

Unfortunately throttling like aeration increases backpressure which reduces strainer efficiency.

Would you care to elaborate on the improbability of cavitation 'at those fluid velocities'?

While I guess I could have done something wrong, I just did a very quick calculation that came up with a velocity of more than 17 fps at 400 gpm in 3" sch 40 Stl pipe. If the velocity has been in that range, and also with non-clear backwash flow conveyed, is it really surprising that a steel elbow has exhibited some metal loss? Also, am curious what is there to lead one to believe that "stainless steel" pipe will survive such conditions really any differently than carbon steel?

rconner,

You are quite right, I used 4”nps sch40; thanks for this verifiable correction.

The literature states that high hardness, yield, tensile and work hardening properties reduce the rate of material loss from cavitation. One paper determined the highest cavitation resistance to be steel with 0.1% carbon and 6% chromium, with a microstructure of 32% martensite and 68% austenite (attached).

• http://files.engineering.com/getfile.aspx?folder=4089ff2c-d8de-4247-9e1b-61

Cavitation produces compressive fatigue stress failures in the pipe flow surface. Cavitation damage is surface spalling rather than "erosion". Cavitation occurs when the local flow conditions temporarily fall below the vapor pressure of the fluid, creating air bubbles that expand and rapidly implode, which produce extremely high compressive stress on the pipe flow surface.

If there is a cavitation condition in the pipe system, then no type of coating will resolve the problem. Switching to a different type of metal for the pipe will not help much either. The only thing that will effectively resolve the cavitation problem is correcting the flow deficiency.

II just don't understand why you're obsessed with cavitation. You might well be producing some bubbles in the water and this continues in the water, but cavitation damage occurs at the point where the bubbles are formed and then collapse within a very small distance and time. The high velocity of liquid with particles in it would appear to be a much more likely cause of your erosion. you might be getting some collapse of some larger bubbles, but I find the whole system difficult to follow. Can you sketch it up as a corss section so we can see what is happening?

Can you advise where you get the flow figures from? My experience of back wash filters is that the flow is directly related to the operating pressure of the filter and given that the pressure loss is essentially taken across the filter elements can lead to much higher flows than you think. Also remember that the initial flow from each filter element will contain the largest particles. Hence if you operate your filters at pressure much above 1 or 2 barg, the flow can increase from that which the vendor has assumed.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

this discussion is a perfect example for the need for getting a piping/mechanical/civil engineer involved in the system design, rather than patching a bad design after the fact.

Materials selection, piping layout, coordination with the pumping/backwash system and where you route the backwash, is only the beginning of the design issues that come up

good luck

LittleInch,

Cavitation erosion has a history of producing through wall pits (not large areas of wall thinning) in the extrados of the first elbow downstream of the strainer backwash discharge MVs (sketch attached).

My task is to reduce the elbow replacement frequency, not re-design the configuration.

• http://files.engineering.com/getfile.aspx?folder=68a66662-7100-4405-9a33-b2

Hutton's paper on cavitation erosion mitigation through aeration is also attached.
I am impressed with the tiny amount of free air that is required to significantly reduce erosion rate.

Now if I could only find someone who has employed this technique in short backwash discharge lines.

• http://files.engineering.com/getfile.aspx?folder=2af68680-bdab-434f-a37d-93

Sounds like you'll be the first, give it a go! sketch up your design, test it and keep us updated

Thanks for posting the sketch as it clarified the issue. There were several posts that were misleading prior to the introduction of the sketch.

The initial post said there was pitting in the steel pipe and elbows.

The velocity through the 3-Inch elbow is 12.5 ft/sec at 275 gpm. The velocity is not excessive in a waste pipeline where you are discharging to atmosphere.

The strainer is operating at 120 psig according to your sketch, which means that it is possible that you have cavitation at the butterfly valves. Cavitation in valves is caused by a reduction in flow area that increases the velocity head through the valve while at the same time decreasing the fluid pressure.

The strainer backwash lines are connected to a common header which discharges to atmosphere and provides a good siphon, dropping the individual backwash lines below atmospheric pressure.” This statement is also incorrect. The siphon is not an issue at all. You only have a drop of 10.5-feet which calculates to a static head of 4.5 psi. That will not produce a significant siphon, especially at the fluid velocity that you have.

The problem is the valves.

You need to replace the backwash discharge valves with valves that are a full port type.

You need to increase the downstream backpressure by installing an orifice at the end of pipe. Size the orifice to pass the required flow while at the same time reducing the discharge pressure.

Install sch. 80 steel pipe, it will last longer. I don't believe the pipe material is an issue at all.

http://www.valmatic.com/pdfs/Cavitation_in_Valves_...

bimr,

You might have forgotten Mr. Bernoulli's velocity term...

Bambie,

Are you throttling the motor operated valves or opening them completely?

Please explain "Mr. Bernoulli's velocity term" and how it was overlooked.

Bernoulli states "The kinetic energy increases at the expense of the fluid pressure". Isn't that was posted above "Cavitation in valves is caused by a reduction in flow area that increases the velocity head through the valve while at the same time decreasing the fluid pressure."

In your application, the static pressure is inconsequential and does not need to be addressed.

Here is another reference for you:

http://www.engineeringtoolbox.com/cavitation-d_407...

Cavitation can in general be avoided by:
"reengineering components initiating high speed velocities" ie components meaning valves
"increasing the total or local static pressure in the system" ie add backpressure valve or orifice plate

http://www.engineeringtoolbox.com/control-valves-c...

• http://files.engineering.com/getfile.aspx?folder=3cd0e002-8469-4205-b163-91

While I am trying real hard not to claim specific expertise with what you are running into here (and probably convincingly!), I noticed now your "turbine blade" paper advocating some specific chemistry/microstructure of metal for alleged best "cavitation" resistance came up with a conclusion that the best resistance was what sounded like really a sort of specialty steel formulation (w/ it appears "6% chrome", and 489 HB hardness) . I think the article you provided is even some confusing to me with its reference to "stainless steel", as I thought that term is generally applied to alloys with far more chrome than that.

While I guess I wouldn't argue that e.g. very high hardness steel (e.g. near 500 Brinell, if available reasonably in the piping configurations you need?) could conceivably provide at least a little better general wear resistance than softer, I don't think that's what you would really be getting with at least the most common grade stainless steel (flanged?) piping mentioned so far on this thread. E.g. I think you most common ferritic SS pipes have in range of instead 16-20% chrome, and are generally much softer than this.

As I'm not sure the point was understood, I will ask a couple more questions - while I understand the desire to backwash quickly (and therefore e.g. one might not want to just throttle flow quantity way back as perhaps another has suggested) and your statement your job is not to mess with "design", is this backwash discharge piping correctly sized given the resulting quite high flow velocity with it appears for at least some time with some sort of solids you've had? Regardless what the experts determine is going on here, I wonder since the piping and valves are to be replaced anyway would really hurt to also put an increaser/reducer and a little larger e.g. lined piping off this backwash discharge (that would appear to result in a much more normal water or slurry flow or whatever velocity beating around the elbows etc.)? I noticed the first thing the folks who have appeared to look at arguably at least some similar-sounding problems at length but in the energy field at http://www.hse.gov.uk/research/rrpdf/rr115.pdf mentioned under the heading "Design of Pipework" was:

"Pipework should be designed to minimise flow velocities..." [immediately followed by "...and avoid sudden changes in flow direction (e.g. at elbows, constrictions and valves).]

bimr and hacksaw,

Okay, we all agree that cavitation is very likely the cause of pitting in the steel pipe and elbows.

The solution cannot be to increase backpressure - that will directly impact on strainer efficiency.

No matter how 'slippery' the valves are, vapour bubbles will still form on the filter elements, travel downstream and cavitate.

The solution is to live with cavitation by increasing resistance to cavitation erosion.

The possibilities are: stainless steel, aeration or polymeric linings.

Any ideas?

rconner,

I agree, the materials in that paper were not readily available, the attachment illustrates that stainless steels have higher resistance to cavitation erosion than plain carbon steels.

• http://files.engineering.com/getfile.aspx?folder=8991496f-bb02-47c5-bdb6-b0

OK, just trying to learn here - what was this last figure an "attachment" to (I guess I missed any reference to it in any of the papers I quickly glanced through on this thread)? Thanks.

The damage done by cavitation is similar in nature to that done by electrical discharge machining. Hard or soft doesn't matter much. If it did the material would be common knowledge, and cavitation would not be a problem.

If you can't up the backpressure enough to stop cavitation, then put in a sacrificial piece or change the routing so the bubbles can't collapse against what you don't want eroded. Since the pressure is low, you do have the chance to put in a viewing window to watch them to determine a method that works for your circumstance.

If you are looking to replace the elbows and materials hard or soft still erode, either pick something cheap or something thick.

The cavitation problem I was involved with was in a hydraulic manifold where one of the valves would whistle a bit as it opened. The bubble stream went across a passage and drilled an extra path through the manifold block to produce control actions that weren't appreciated. The ultimate fix was for the valve maker to redesign the valve, but a short term bandaid was a steel sleeve, which would wear through a little slower than the aluminum.

Where did you get this idea? "The solution cannot be to increase backpressure - that will directly impact on strainer efficiency."

Have you discussed this issue with the manufacturer of the filter?

Are the valves used for control or are the valves on/off?

If the valves are used for control, replace the valves with a control valve that has a cage allowing the bubbles to collapse within the cage.

http://www.cla-val.com/documents/pdf/B-KO_Anti-Cav...

If the valves are used for on/off, replace the valves with types such as full-port ball valves, plug valves, or pinch valves.

Bambie,

You clearly have an issue with your elbows on the backwash side and appear to be trying to solve the symptom, not the cause. We disagree about what is causing your erosion, but no matter - use a Teflon coated elbow or a HDPE elbow and you will replace them much less often.

If you want to fix the problem rather than just replace elbows on a less frequent basis, you need to start from the basics. First - What exactly is the flow rate coming out of these things. You are running a fairly high filter pressure (120 psig / 8 barg). Given that your backwash is effectively at 0psig, you need to determine from the filter vendor what the flowrate is for that DP. When I worked on a similar system, we worked out that during the backwash cycle, we lost about 10% of the flow... Now if your butterfly valves were manually set to some partial opening to reduce the flowrate, then your flow may be less than that, but given that the valve is being subject to impact with dirt and hard bits, there may not be much of the valve left by now. Any cavitation effect will be on the valve itself, not further downstream. Your assertion that providing some sort of flow restriction further downstream would impact on the filter performance is not correct. You have 8 barg driving the flow - 1 barg back pressure won't make any difference

It is a well known issue on "dirty" flow that any changes of direction need to be as gentle as possible and as slow as possible.

So to fix the problem you really need to stick in a reducer and go to 6" in a straight line before bending down into your header. By the way your discharge piping looks Ok, but it is not a syphon. ?I can't see if your discharge pipe goes below the lake level, but if it does then as bimr has noted, your static pressure in the header is still about 10psia - nowhere near low enough to cause an issue when flowing.

I've done my best to try and look beyond your current issue, but there are so many holes in the data that I can do no more so good luck and keep the spanners handy for changing your elbows on a regular basis.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

1. strainer blow-down in not continuous but intermittent, the op's presumption is that "bubbles" are being formed in a liquid-full pipe, with insufficient pressure drops to cause cavitation...yet claims the raw water is not capable of support MIC...

2. the mov's, if full opened have neligible pressure drop and can only support incipient cavitation at best if partially opened.

The discharge piping sizes are responsible for neglible pressure drop<15 psi at best.

The op has not addressed the valve operation; though butterfly style valves rarely used or needed for high pressure drop control

3. Strainers or strainer flushing do not support cavitation processes.

4. The low cost alternative is to hire a qualified engineer!

The vendor literature states that 7000 usgpm will pass through 15, 750 micron filters or 467 usgpm per filter with 7.2 psid when dirty, which gives each one a Cv of 174 and flow of 1905 usgpm at 120 psid.

These values are based on inside-out filter flow, so I am endevouring to find out if this is also the outside-in flow characteristic. That would be a discharge velocity of 87 ft/sec...yikes.

Bambie,

run a material balance on your system, showing the normal flows and flush flows, calculate the various piping drops, valve drops along with the valve settings,

specify the details of the flushing cycle (timing and how many strainers flush at any time).

you also need to specify the percent open of all valves

It is incorrect to describe the damage done to a metal surface by cavitation as "erosion". Erosion would imply some form of abrasive wear produced by the fluid flow. The pitting and loss of material produced by cavitation on a metal flow surface is spalling due to compressive fatigue failure. All other things being equal, a metal flow surface with higher compressive strength (ie. hardness) will be more resistant to spalling failures, and thus cavitation damage, than one with lower compressive strength. The reason materials like stainless steel or bronze often exhibit better tolerance to cavitation is due to their greater corrosion resistance. The surface spalling process resulting from cavitation can be greatly exacerbated by surface damage caused by corrosive effects of the fluid on the metal.

That is a good start point and it will be interesting to see what the vendor says. To be fair to them, the filters are often run at relatively low pressures of a 2-3barg so they need to have the ability / size in the back wash to have a good flow rate under those conditions. It would be worth trying to find the original data sheet from both you and the vendor to see what they had assumed for the operating pressure.

What the filter vendor sometimes does is on commissioning sets up a throttle valve or some other type of flow restriction downstream to get an acceptable backwash flowrate (say 5-10% of normal throughput). What then happens is that either that valve or restriction quickly erodes or gets "readjusted" by someone who doesn't realise what it is doing or gets blocked and then the operator just removes it as it "keeps getting blocked". The reason for it then gets lost in the mists of time until the elbows start to erode...

Given the interest in this post, please let us know how you get on.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

Are you seeing just localised pitting on the extrados? Or does general wall loss also accompaniy this? Perform UT to confirm, not just visual.

Looks to me to be erosion. You can see the grooves and gullies being caused by it.

If it was MIC, the entire piping system would have been affected. Doesn't look to be cavitation either, but there's only so much one can tell from one pic.

...

But in order for you to fix the problem, you need to know the causes. Cut out an elbow and send it for failure (metallugical) analysis. This would conclusively tell you the cause.

If it's cavitation or erosion, upgrading to Stainless Steel elbows should fix your problem.

Or you can go further and look at eliminating the root causes of the erosion or cavitation.

Ripz,

Elbow replacement is planned very soon (sample photo attached). I will obtain better photos then.

LittleInch and hacksaw,

I have little faith in being able to eliminate cavitation via throttling - it will remove material somewhere in the system. Moving it downstream to the isolation valve or common header would be more disruptive to replace.

• http://files.engineering.com/getfile.aspx?folder=69d9b86a-1396-4736-b5d5-d1

In looking at the original photo, I also tend to think that you are seeing erosion damage rather than cavitation damage. Probably some corrosion issues also. It will be interesting to see more photos of the affected components.

First, I am not an expert in cavitation but have had some past interest in the subject. While I work in the piping field, I think I furthermore have only run across only a time or two of instances I can remember where I was told some lining/piping destruction problem was caused by "cavitation" in a great many years. While I have probably been some sheltered, this could also mean at least the real deal happens rather infrequently.
When I first opened up this photograph reported as cavitation damage, I likewise have to admit however what I thought I saw in the picture (some sort of what appeared to be axially oriented flow lines?/grooving? and some deep, but widely separated/discontinuous "pits"?, and perhaps even with much more normal metal thickness surrounding?) did not appear to fit the image I had in my "minds eye" of cavitation caused by very abrupt pressure drop etc.
While that certainly doesn't mean this is not cavitation, my previous minds image, right or wrong, was of a porous, pock-marked surface deeply scarfing an area of a valve or very closeby piping, I thought to be caused caused by sort of blasting or splattering due to implosion of gas bubbles.
Incidentally, in the one past instance I thought I remembered where I was advised someone had encountered severe cavitation, I thought I read or heard in my research way back then that the violence of the vapor implosion events were akin to tiny "nuclear explosions" that really no lining or pipe could withstand (sort of like what I have read from at least a few on this thread). While I had no idea then where or how they had come up with that analogy (and it seemed to be a sort of non-obvious over-statement to at least a few of us sort of semi-educated folk back then), I now read a more specific statement in another technical reference,

"When the cavitation bubble collapses it generates a temperature of 5,000 degrees C and a shock wave that travels over 500 miles per hour."

If this is in fact true (and for someone to even quantify such in this fashion I suspect they would have had to have some rather heavy duty brainpacks, $ and instrumentation behind them!) I'm not sure I would assume that the problem (cavitation or erosion) would be magically solved economically or functionally by just throwing in instead at least common austenitic steel pipes and fittings, I think that have about a maximum 88-95 HRB (maximum Brinell ~151-175 hardness). I guess this is what LittleInch talked about in treating the symptoms as opposed to the problem.
As butterfly valves are apparently involved with low discharge pressure, you may also be interested in the guidance at http://uwrl.usu.edu/facilities/hydraulics/projects... I also noticed in the latest pic that the 90 elbows are obviously VERY "close-coupled" off what appears to be a discharge valve, a practice I think had also seen mentioned in past in some literature.

On a definitely lighter note, there is homework for the weekend at http://www.youtube.com/watch?v=gy7PTl6UJfI . Everyone have a good one!

In at least one location, I'd try putting in a Tee or Wye so that cavitation bubbles will either collapse against a more easily replaced capping plate or when they reach the increased stagnant pressure in the dead end, rather than carrying into the bend of the elbow. There is recognition of an increase of back-pressure, but this may be small enough.

Anecdotal: I saw a desk-toy in the form of a sealed glass tube partially filled with water, the remainder water vapor. If the tube was rocked back and forth the water would slosh, but if it was shaken it made a plink sound, very much like a ball bearing slamming into the ends of the tube. As the vapor and liquid are in equilibrium any mechanical change causes an instantaneous phase change at the interface.

rconner,

The youtube link demonstrates the cavitation mitigating potential of aeration (beer carbination in that case).

Could this be achieved by locating a flanged orifice plate, with pressure taps connected to an air injector, downstream of the motorized butterfly valve? The orifice plate would need to be sized to provide 26 psid in order to generate the 5 usgpm required for air entrainment (by the model attached).

• http://files.engineering.com/getfile.aspx?folder=56fcf702-dce9-4e76-9a25-f3

Bambie, glad to see you finally understood, the product you have selected is used to increase D.O. and thus to reduce MIC, not cavitation...

hacksaw,

I always try to hear what the machine, rather than the tool, is trying to tell me... "Maybe Its Cavitation?"

The Lancaster product is used to add oxygen in order to oxidize soluble iron to insoluble iron so that the insoluble iron can be filtered out.

http://www.youtube.com/watch?v=vRk1SWucy1s "Cavitation in a globe valve"

Video of cavitation into a stub-Tee.

Bambie,

Oxide formation is essential in reducing Fe corrosion, and Dissolved Oxygen is also an essential step in removing the microbes that solubalize iron in the first place.

Some of the posters dealing with water treating can better explain what happens in increasing the D.O. and reducing the D.O. demand.

Good luck with your planned changes, they are right on target.

hacksaw and bimr,

I love a good teasing...you're the best.

PS I attached an excellent paper on "Mitigating Irritating Cavitation" thru aeration.

• http://files.engineering.com/getfile.aspx?folder=559fbf1a-0dc5-4c7f-b81e-3e

Here is a tease for you from Bela:

"Because no known material can remain indefinitely undamaged by severe cavitation, the only sure solution is to eliminate it."

• http://files.engineering.com/getfile.aspx?folder=5aa7f47b-65ae-4528-89a0-0d

bimr,

I didn't answer your question about "Mr. Bernoulli's velocity term" and I'm not certain he should get the credit, however, I was thinking of the following:

A long siphon connected to many filters will have unsteady flow.

From first principles: delta P = rho x c x delta V.

For water: delta P (psi) = rho (62.4 lbm/cuft)/32.2/144) x c (4720 ft/s) x delta V (ft/s)

delta P = 63.5 delta V

You only need a quick change in velocity of 1.6 ft/s to drop 100 psi in a water filled pipe.

The siphon drops static pressure close to vapour pressure and the shunting columns of water continuously open and close vapour pockets, which ravage my pipe.

Would you agree with this argument?

Don't know about bimr, but I don't agree with that. Your "syphon" won't do any of these things and, IMHO, the thing ravaging you pipe isn't cavitation, but a high velocity of liquid containing dirt and sand.

Did you get an answer back from your filter vendor on the flowrate anticipated through the back wash line at 120 psig to zero? You anticipated some very high veleocities.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

LittleInch,

I have learned the following:

1) These strainers and pumps are normally on "standby" and only see flow once per month during testing to confirm availability, which is done without starting the pumps.
2) There could, therefore be an accumulation of silt on the ouside of the 750 micron (.03" mesh) elements, but how would this get through?
3) The vendor suggested that velocities would be approximately 40 ft/s

Perhaps one month is time enough for MIC to build tubercles and create pits which are then removed via cavitation erosion?

Bambie,

Can you just provide us with a clear set of information on these elbows.

we need to know:
How often are the filters back flushed?
what is their purpose / normal operation?
Is there any biological inhibitor injection?
Does the discharge pipe remain full of water (stagnant) in what you now say is one month between flushing operations

your points above
1) I cannot understand how the system is tested without the pumps being operated
2) Any dirt etc would be on the inside of the filter (normal flow in to out) which then gets flushed out during back wash (Out to in)
3) That sounds about right, but is a high veleocity when it also has particles in it

MIC can occur quite quickly and if it builds any items would simply get flushed away by the flow and any dirt particles in the back wash - the cavitation issue is a red herring.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

It not completely clear, for example what is the 4720 ft/s term?

It may be possible that this siphon occurs during transitions when the systems is starting up or shutting down. But you have witnessed the noise during continuous operation and the operation has been described as generally continuous rather than being a batch.

I am not confident that the siphon is the problem.

You said the problem is at the elbows. At the elbow, you have 12 ft/sec velocity which would generally be defined as full flow which would prevent the loss in pressure from traveling upstream. Since that location is also the location where the problem is the worst, one would suspect the problem originates with the valve not the siphon.

Your sheet also shows the siphon being approximately 10 feet or so, which has a significantly lower pressure than the inlet pressure.

If you were having the cavitation problem in the larger 6-Inch pipe, you may have an argument.

To resolve the siphon question, consider adding a siphon break on the 6-Inch pipe. The cost of the siphon break will be very inexpensive compared to the other solutions proposed.

Maybe you can explain the flow period of once per month? How long does the flow occur?

The 40 ft/sec velocity (suggested) is too high to be believable. If you know the flow and pipe diameter, it is simple to determine the velocity.

Since the strainers are backwashing to remove the crud and detritus that are present in the process water, this material will eventually be backwashed out into the backwash line.

I can believe 40 ft sec in the 3" backwash nozzle. The filter has 120 psig inside it remember and there's not a whole lot of resistance to the flow back through the filter. Use a PE or a Teflon lined pipe and it should make all the issues go away - or put in a bigger pipe / elbow to reduce the velocity.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

Looks like you might have room to weld two 45* reducing elbows together, with the large ends in the middle. Worth a shot to reduce velocity and see if life improves. That change might help supplement a better solution. Of course then you have a weld right in the middle and that feature could make things worse.

Let's see:

1) These strainers and pumps are normally on "standby" and only see flow once per month during testing to confirm availability, which is done without starting the pumps.
2) There could, therefore be an accumulation of silt on the ouside of the 750 micron (.03" mesh) elements, but how would this get through?
3) The vendor suggested that velocities would be approximately 40 ft/s

Curious to see how with the pumps and strainers normally on standby, and the pumps are not turned on for your tests, that you are having problems in all of the discharge piping.

It is also curious that you are only associating filter blowdown with cavitation, but not the pumps.

You've clearly stated that the blow down flow of 10 ft/s or so in the 3" pipe, though you later increased this to 4" piping, now it is 40 ft/s

Sound like you owe the forum a bit of an explanation

Until better pics and less contradictory information is available it may be hard for anyone to tell exactly what is going on here. While perhaps flow erosion or other may not end up being the dominant issue (since we have learned of some periods that could involve perhaps significant stagnation etc. as well?), you may be interested in the lined pipe flow velocity guidance e.g. on pages 9-1&2 (I saw PTFE listed) at http://www.publications.usace.army.mil/Portals/76/... which states,

"Liners should not be affected by erosion with liquid velocities of less than or equal to 3.66 m/s (12 ft/s) when abrasives are not present. If slurries are to be handled lined piping is best used with a 50% or greater solids content and liquid velocities in the range of 0.61 to 1.22 m/s (2 to 4 ft/s)..." I think DIPRA suggests a similar magnitude/maximum of about 14 fps for lined ductile iron water pipes in normal service. For design numbers beyond such guidance, ​I guess ​it might be best to talk with the manufacturers involved.

Everyone have a good weekend. ​

LittleInch,

How often are the filters back flushed? Once per month.
what is their purpose / normal operation? Standbye (not operating).
Is there any biological inhibitor injection? None.
Does the discharge pipe remain full of water (stagnant) in what you now say is one month between flushing operations Yes.

your points above
1) I cannot understand how the system is tested without the pumps being operated (Three normally operating pumps pressurize the common discharge header).
2) Any dirt etc would be on the inside of the filter (normal flow in to out) which then gets flushed out during back wash (Out to in). Since the pump is not run, there is no accumulation of organics or silt on the inside of the filter elements.

hacksaw,
1) These strainers and pumps are normally on "standby" and only see flow once per month during testing to confirm availability, which is done without starting the pumps. (Pump discharge header is pressurized by three normally operating pumps)
2) There could, therefore be an accumulation of silt on the ouside of the 750 micron (.03" mesh) elements, but how would this get through?
(Backwash removes silt and organics from the inside surface accumulated during normal operation, not standbye mode. Silt and organics accumulated on the outside element surface would not enter the backwash discharge - to sandblast the elbows).
3) The vendor suggested that velocities would be approximately 40 ft/s
(This is due to 120 psig).

Curious to see how with the pumps and strainers normally on standby, and the pumps are not turned on for your tests, that you are having problems in all of the discharge piping.
(The two strainers at the end of the line, closest to the header discharge are on standbye, the three upstream pumps and filters are normally operating).

It is also curious that you are only associating filter blowdown with cavitation, but not the pumps. (no capeesh).

You've clearly stated that the blow down flow of 10 ft/s or so in the 3" pipe, though you later increased this to 4" piping, now it is 40 ft/s
(I clearly stated that the 10 ft/s - corrected to 17 ft/s was due to the syphon alone, not due to 120 psig)

bimr,

4720 ft/s is the speed of sound in water (c).

Once per month each of the 15 filters elements is cycled thru the backwash in the two standby filters.

We are seeing cavitation pitting in the 154' long 6"nps header, just not as aggressive as in the 3"nps standby filter elbows.

I like your idea of vacuum breakers in the 6"nps header - that would prevent cavitation due to unsteady flow pressure transients in the header and, perhaps, the 3"nps elbows too.

I'm puzzled as to what the speed of sound in a liquid has to do with delta P.

q=1/2*density*V^2. dq = density*V*dV. This neglects delta P due to change in height.

3DDave,

This equation predicts decelerative head, which (I think) should be happening in a very long siphon of incompressible fluid with many sources of constantly changing flows (and no vacuum breaker).

It is attributed to Newton's law of motion and the continuity equation.

The equation seems to apply with to instantaneous change in flow, not a continuous flow. It is related to sharply closing valves, which is not the condition previously mentioned.

Thank you 3DDave for answering the Bernoulli question.

I would add that the speed of the pressure wave in such situations (a) varies with the fluid, pipe size, and pipe material. For a medium sized steel line, it has a value of about 3500 ft/sec, slightly less than the 4720 ft/s that you used.

3DDave,

The filter backwash on each of the three continuously operating pumps will cycle on whenever the filter pressure drop reaches 7.2 psid.

The motorized butterfly valves close in less than one second.

Inertance or rigid column theory predicts delta P = rho x v x L / delta t where:

rho = .0135 (lbfx(s/ft/in)^2, v = 40 x 9/36 (ft/s) and L = 154 ft

delta P = 21 psid per second at 3"nps backwash velocity of 40 ft/s, which is 10 ft/s in the 6"nps header.

What is going here - first you say the filters get flushed once a month and now three pumps are running continuously with regular flushing once they reach their diff pressure limit, flushing at 40ft/sec (12m/sec). Please stick to one story or the other. Is the damage occurring in all the elbows? DO you change around which filters / pumps are working?

Your system is not suffering from cavitation, but pure erosion in the working units and probably MIC induced corrosion in the virtually static units.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

LittleInch,

If you refer to my sketch, strainers STR1, STR2 & STR3 are backwashing whenever required by their continuously operating pumps. These elbows have never been replaced, which I assume means low wall thinning rate.

Strainers SRT4 & STR5 are backwashed once per month, however since their pumps are on standby, there may be organics but no silt in the backwash flow. The first elbows are being aggressively pitted.

I agree with LittleInc. If you are experiencing pitting in the strainer piping where the strainers are mostly in standby mode, the pitting is most likely due to MIC. The periodic backwashing will tend to wash off the exposed corrosion detritus exposing the cleaned areas to additional corrosion.