Design Rectangular Low Pressure Stainless Steel Pipe 7

[eeh68]

Hello everybody.
I'm tasked to design a 23" X 54" rectangular pipe, about 100 feet long. The pressure is quite low, at about 12 1/2 psi, but the moments produced astound me. I'm looking at 3/8" 304 L Stainless Steel plate with channel stiffeners. If I could do a round pipe, there'd be no problem. It is comparing the wall thickness of the round to rectangular that has me thinking I'm doing something way wrong. The indeterminancy and five foot side is making me consider a career in real estate. Pipe design books seem to only talk about using existing round pipes and configuring them for flow. I need some basics like: what wall deflection is allowed; Does Mc/I still apply (I'm using it); what references may be available; and Is this really as hard as it seems??

 

[MikeHalloran]

That's a big pipe, and the net force on each face is substantial. The problem becomes worse, of course, when the pressure changes.

I've seen similar sized and similarly reinforced rectangular pipes used for gas turbine exhausts, at much lower pressure and with somewhat thinner walls, and the designers are always surprised when the pipes develop cracks in less than 30 days of service.

PIPES ARE ROUND FOR A REASON.

Write that down and nail it to your manager's face.






Mike Halloran
Pembroke Pines, FL, USA

 

[jte]

Mike-

Would you specify a round or square nail for that?

jt

 

[eeh68]

Thanks,
I appreciate knowing I'm not totally out of it...
I mean, everybody keeps telling me it is no big deal, though none of them seem willing to step up and do it.
Suppose I could just use superglue?
g

 

[JStephen]

I would anticipate that the design could become unreasonable. Even if it's rectangular at each end, it might be worthwhile to transition to round in between.

From a design standpoint, I would try using the rectangular plate formulas from Roark to design the plate between stiffeners. Then I think he has frame formulas that can be applied to the stiffeners themselves. Normally, in designing the stiffeners, you wouldn't necessarily use Mc/I with the full plate width. AISC-ASD includes an awkward design method for "wide" flanges on beams that can be applied. It's not uncommon to assume that 16t either side of the stiffener acts with it.

If the pressure is fairly static, and flow is not terribly high, I wouldn't see any big problems, other than lots of stiffeners (and you might want to use a fairly high bending stress for this case). But if pressure is fluctuating or very high flow rates occur, you could get some vibration or unwanted deflection in the panels.

You say it's 304ss- can you use carbon steel stiffeners on it?

 

[unclesyd]

Look for a book that covers designing duct work.

The only thing I remember from the design of rectangular duct is to keep the welds off the corners.

 

[eeh68]

The "pipe" has to stay rectangular throughout its length. No, the stiffeners (channel/angle/hollow setions, etc) will also be SS to avoid corrosion. May use 316L SS for the stiffeners for the slight additional strength, haven't yet checked availability. I'm not convinced they would change the design much anyway. The pressure will be about 11 psi (pushing out) when full, with possibly 12 psi (pushing in) while emptying. It will be vented, but in the field sometimes the undesirable things we couldn't quite do in the lab happen. I'm planning on an 8-inch vent line at the high point of the line, but a fast evacuation with the help of a plugged vent could result in very low pressure in the 'pipe.' At the elevation where this will be located, atmospheric is about 12.2 psi, so that is what I'm using for all calculations. I'll see what I can find from Roark, and thanks for saying "16t."
All things being relative, I'll just state: 70 to 100 cfs is the anticipated flow, so the velocity through this section would be 8 to 12 feet per second. I do wonder how the water will act in such a tall, thin 'pipe.'

 

[JStephen]

Normally, in a round pipe or vessel, pressure tends to round it out, whereas vacuum tends to buckle it. But here, it's stiffened either way, and it won't make a great deal of difference if it's vacuum or pressure.

If I remember right, the "L" grades of SS have somewhat lower stresses than the non-L versions. If you're not following a specific code, consider using a dual-spec material & design for non-L properties.

 

[eeh68]

Yes, the 304L SS is lower, Fy=30,500 psi, but it is easier to weld as in less steps before, during and after. It has about 0.03% Carbon whereas the regular 304 has about 0.08%. If nothing else, I'm learning about stainless steel. 316L or 316 is much stronger, but cost about 1 1/2 times as much, so my first, thick-walled, 316L design was shot down.
Thanks, I'm hoping to be able to go into work monday and tell them why this thing may work, but it is going to be much more expensive, thicker, etc. than a round, 48" Diameter HDPE would be.

 

[SnTMan]

eeh68, if available, you might take a look at ASME Sec VIII, Div 1, Appendix 13. It deals with pressure vessels of non-curcular cross section. If you don't have software that runs these calculations, they are not too difficult for a simple rectangular section as you describe. The appendix also has calculations for a stiffened section as you describe.

Appendix 13 calculates stresses at the side plate midpoints and at the corners. It may require thicker sections than you'd really like, however your pressure is not that high, and the section is not that large. I don't believe it allows for external pressure design, so you would need to avoid those conditions.

Dual certified stainless plates should be readily available if you need the higher allowables. If by HDPE, you mean plastic, yeah, I would figure the stainless, fabricated duct is going to cost more.

Maybe this will be helpful.

Mike

 

[arto]

If you can have "dead space," how about a big round 60" pipe for pressure, with a thin rectangular flow liner with support plates/discs along the length ? {check for flow-induced vibration}

 

[eeh68]

Mike,
Thanks, I'll try to find ASME... and yeah, the HDPE is plastic and carries the water to its ultimate destinations. I just need to get it to where they can connect.

 

[eeh68]

arto etal,
I guess I've not given the limitations. This 'pipe' conduit or what-have-you is limited to the rectangular shape and cannot exceed 60" high, and preferably 24" wide though the stiffeners are going to make it wider. It is going into a 4ft wide by 5ft tall concrete box culvert and both the inside of the pipe and the remaining area outside of the pipe must be accessible for people (smaller than me!) to go through and inspect. i've got a ton of 'better ideas' but the shape, the location, etc. are being dictated. so all i need to do is come up with a design that will work and last at least fifty years.
g

 

[jte]

eeh68-

One thing I haven't seen mentioned here is the possibility of rounding the corners. Do you abosolutely need to have square corners or can you bend the plate and give the corners a 2" radius? That would do wonders for your thickness calc's.

jt

 

[eeh68]

jt,
Would it?? We expect to have the steel bent into shape so rounded corners are a given (at least in my mind). My drawings all show rounded corners and I've seen that as a potential problem due to the high moments at the corners. I've been thinking I need to put some sort of plate from the corners to the stiffeners, but the post from 'unclesyd' had me worried about doing so, though I assume he is referring to longitudinal welds. I'm still at the stage of handling the moment and deflection at the center of the five foot span and researching stainless steel. Apparently I lied to y'all in that 304L apparently isn't any easier to weld and perhaps has no place in this project. Anyway, as is apparent from my first post, I'm in over my depth, and our office library doesn't have too many references that approach this kind of a problem.
Thanks
g

 

[unclesyd]

When we made our duct we made channel sections and welded the side plates to same. The the corner was rolled with about a 1"-2" straight section beyond the tangent line of the rolled corner. This was to take the bending moment at each corner out of play, in other words the weld doesn't see this moment.

 

[MikeHalloran]

I'd be inclined to make the channel sections in quarters, each having just one short bend, to the outside, so all the corner seams are bolted, or plug welded, with all the fasteners on the outside.

But I find the requirement for external inspection access somewhat inconsistent with the fifty year design lifetime.

Inspection suggests that you expect it to fail. But if it does, how will you repair or replace it? And if it were not accessible, and failed, wouldn't you still detect the failure the same way? How often will it actually be inspected? What happens when it leaks?







Mike Halloran
Pembroke Pines, FL, USA

 

[eeh68]

Sorry, this pipe isn't necessarily the subject of the inspection, though it would be inspected. The things requiring inspection are 'upstream.' The gates at the mouth of this new, pressurized pipe, and a free discharging spillway above and ultimately to the side of the new pipe. Also the concrete box culvert the pipe will be in along with its existing waterstops. I think the inspection frequency is every 6 months, though the users may inspect it more frequently. We obviously have more than one entity involved, actually several within our own office. The fifty year life I mentioned, is probably less than expected. We generally look at 200 year lives for the things we design and build...This particular pipe is out of the ordinary for us, and many of you have helped me see it isn't exactly ordinary for anybody.
If it leaks, we fix it somehow...that is a possibility that has been disrupting my sleep. Fortunately this pipe leaking would be inconvenient as opposed to life-threatening. It is part of a larger system that if it failed probably still wouldn't be life threatening, but would be a big problem...you might see it on the news!
I haven't been trying to hide what the overall project is, just didn't think it really mattered for the specific questions I've had. This pipe is being placed in the outlet works of a small dam. there is 25 feet of head, and a fairly good sized pond impounded behind the dam and dike system. the referenced inspections are mainly for the dam. even though dams look like little more than piles of dirt, they are actually fairly complex structures, and a little bit of water in the wrong places can cause horrible damage.
g

 

[jte]

eeh68-

Yes, it would. Go outside and pop the hood on your car/truck. Look at how the radiator header on the top or bottom is fabricated. See how they put a radius at the corner instead of a sharp bend? From another perspective, imagine if you could increase the bend radius to... 11½". Don't you think that would be a stronger design? So, qualitatively, any move from a 0" radius sharp corner towards your ideal 11½" radius is beneficial. If you could run a quick plane strain FEA you'd see just how significant a small radius is. Another example: Look at a 1 gallon milk bottle. It has flat sides with rounded corners. They're rounded for a reason...

jt

 

[eeh68]

JT
Thanks. I have no problem accepting that the rounded corners will be stronger, I just do not have a way of calculating exactly how much. I may be overdesigning as I have the "stiffeners" carrying everything with their spacing being determined by the flat plate stresses.