Reinforcing silo with exterior welded plate

[StructureMan44]

A silo has section loss of up to 40%. We are remedying the situation by using curved A36 plate to form a horizontal segmental ring. We are planning to bevel the edges of the 1/4" plate to allow for a full-pen weld at each vertical interface while the horizontals will be fillet welded to the existing silo. My question is, if we would like the full-pen weld to engage the existing silo, is anything more needed other than making the gap between plates slightly wider?

 

[JStephen]

Check the allowable compressive stress in the new plate assuming no support from the existing tank. If it's adequate like that, it shouldn't matter if the verts attach to the existing shell or not. If it's not adequate like that, consider putting holes in it and plug-welding to the existing shell. Or possibly increase the thickness slightly, and see if the extra material is cheaper than the extra labor.

 

[SKJ25POL]

StructureMan44 (Structural)
(OP)
Is it a silo or a tank? The title says silo, you stated in the text existing tank?
What the structure contains?

Can you provide some pics of the location of the damaged part in relation to full structre?

 

[StructureMan44]

JStephen: Thank you; this is great approach; assuming the existing silo walls no longer contribute to the vertical capacity. The silo is outdoors so the less welds the better because of corrosion and having to go back and paint over the welds.

SKJ25POL: Thank you for catching that, it is a silo (vertical cylinder), I have corrected my original post. I don't have any pictures as of yet. The damage is halfway up the 60ft silo and appears to be isolated to one 10ft high band.

Does any resource provide guidance on hoop stress or an equation for checking silo buckling strength or dent resistance strength?

 

[JStephen]

In the US, there's not atandard for silo construction. Typically, the allowable stresses would be similar to those used in the API or AWWA tank standards, depending on the whims of the designer.

The problem is that silo loading is a combination of hoop stress and compression. You can make certain assumptions about the product and how it behaves and get design loads. You can make different, but equally valid, assumptions, and get different loadings. Ideally, you do tests on the actual product stored and base any design on that. In reality, that usually doesn't happen. Gaylord and Gaylord's Structural Engineering Handbook has some information on silo design, and they also published a book specifically on bin and silo design. John Buzek has a silo book out, I believe, and there may be others.

For small silos, it is usually easiest to treat the product as a liquid and check for hoop stress, then assume entire product is supported by the shell and check for compressive stress. For larger silos, that will be overly conservative.

 

[dhengr]

StructureMan44:
Be careful, when you put those patch rings on the silo, that you don’t trap or allow moisture btwn. the two layers of steel. Or, you might create a corrosion problem more damaging than what you are trying to fix. Also, consider the fact that when you connect the two layers at discrete locations circumferentially and vertically, loads and stresses (over stresses on the inner tank wall) will be transferred at these discrete locations, and then flow into the new reinforcing shell. I don’t think you should assume that you can just split the loads and stresses in proportion to the two thicknesses. For example, the hoop stress will primarily be taken by the inner shell until it yields and comes into intimate contact with the new outer shell by stretching. I don’t think you can apply the outer shell tight enough (almost prestressing it around the inner shell) to make them share this stress initially. I suspect JStephen knows more about this aspect of tank design/construction than I do, so this post may be partly to jog his expertise and thought. Compressive loads from the stored product, roof loads and self weight might be transferred more uniformly every 4' at girth welds. And, there might be construction schemes and temporary unloading means during the reinforcing ring application which cause the loads to be better redistributed to the new reinforcing shell.

 

[StructureMan44]

JStephen: Thank you. I’ll look for those references, it might be worth buying them for use in the future. The silo is fairly small, I’ll use your approach of treating the product like a liquid for hoop stress and the entire vertical load for a compressive stress on the shell.

Dhengr: Thank you. That’s a very good point about moisture. Is there any way to eliminate this moisture concern; possibly through heating during the exterior reinforcing ring application or through some sort of coating between the layers? For design purposes I am conservatively assuming the inner wall will no longer take any (compressive or hoop) load; it’ll be taken entirely by the new reinforcing shell. I agree in actuality the hoop stress will still be taken by the inner ring and then it’ll transfer to the outer new reinforcing shell. To your point about girth widths; are you thinking that the horizontal full-pen welds between each existing 5’ tall silo ring will be where hoop stress is transferred to the new reinforcing ring because this weld is likely wider than the steel that is being welded together?

 

[StructureMan44]

With regards to moisture AISC 360-10 B3.13 states “In a sealed HSS, internal corrosion cannot progress beyond the point where the oxygen or moisture necessary for chemical oxidation is consumed” If the air gap between the existing corroded plates and the new steel is fully contained by welds can we assume that corrosion will be limited by the oxygen and moisture content of the air at the time of construction?