Load-bearing capacity of aluminum L-beam

[Violina]

You're right,

What is the area between the short sides called?

That's where it also dips.

Long side

Edit: or do you mean the whole deck??

 

[Derbyboy]

Well, both spans are sagging.

1 beam is sagging

14 (?) floor joists are sagging

 

[Bernieberg]

Can we call it that the roof bends down both longitudinally and laterally?

 

[Derbyboy]

Sounds good.

If TS were to place a beam in the middle, splitting the span from just over 4 meters to just over 2 meters, he could reuse some of the frame timber.

 

[Huggedugge1]

I imagine that a main beam in this style would be nice (seen from the short side). For now, I assume that the floor joists in the transverse direction can be solved with 170 mm studs and that one must accept that it will not be super rigid (was it 120 mm in the existing one?)

[image]

I imagine that the beam is made of stainless steel sheet, two strips and a fillet weld on each side. The steel beam becomes quite integrated into the construction and only the edges will be visible. It will be quite nice to assemble as well without a lot of wood-metal fastenings.

Regarding dimensions, the goal is a moment of inertia of 1895 cm4 for 15 mm deflection according to previous calculations. If the two wooden studs are properly spliced along the length, they can be counted towards the stiffness. This reduces the requirement for the steel beam to 1660 cm4 (assuming 10 GPa E modulus for wood). The bottom flange can be chosen width-wise a bit freely and let 45 mm protrude under the outer stud. It can be chosen to fit some standard for flat iron but let's say 120 mm. Then the sheet needs to be 12 mm to achieve the desired deflection. I have not accounted for the weight of the beam in the example, it is about 250 kg which increases deflection to 20 mm.

I think it might be a bit bulky to work with, but it might be reasonable. The next step is whether to pre-cam the beam. If it's welded, it shouldn't be too troublesome other than cutting the beam web with a slight curve and then clamping down the flange when welding. When the deflections are this small compared to the span, you can calculate the same way even if the beam is pre-cammed, there won't be any non-linear effects to speak of. You simply subtract the pre-cam from the deflection you calculate. So if you pre-cam by, for example, 30 mm, the allowed deflection in the calculation is 45 mm instead of 15 mm. In the same example, you can then reduce to one third the moment of inertia. There will be less material but more expensive to manufacture. Personally, I have no feel for the balance, it's probably a business economics question

These were some gathered thoughts to add to the idea fire anyway.

I think I understand in principle what you mean. Mounting an upside down T-beam seems possible and the existing look would be maintained. Replacing the existing 120*45 studs with 170*45 is no problem.

Then I am unsure about what you mean by cutting out in the beam and how to solve the attachment on the short sides. Do I understand you correctly that the beam should be tensioned in the middle by about 30 mm when it's welded?

Should there be a similar beam in the middle?

Stainless, can't you paint the steel sheet with rust protection?

 

[Huggedugge1]

Yes, both spans are sagging.

1 bearer is sagging

14 (?) floor joists are sagging

No, the front and middle bearer are sagging. There are 13 double joists, one of which is full-length that is set into the bearer and tightened.

 

[Huggedugge1]

Can we say that the roof bends down both lengthwise and crosswise?

Yes, depending on the main beam in the middle.

 

[Huggedugge1]

No, a rectangle has FOUR sides.

Two short sides and two long sides.

The short sides and one of the long sides don't sag, they have support underneath.
The last long side sags because it has no support.

That's right. Additionally, the supporting beam in the middle sags.

 

[Huggedugge1]

Have you considered a slatted cover?

Considered filling the pool with concrete…

 

[Derbyboy]

I can imagine that...but you keep fighting, is it... strong or stupid, I don't know

 

[Huggedugge1]

I think we mean the same thing. A rectangle has two sides. One side is called the long side (at least if the other side is called the short side). In your case, the long side is just over 8 meters and the short side is just over 4 meters. The short side sags (clearly visible) except right at the end where the short side is supported by wheels.

You need to redo the short side as well.

The short side doesn't sag. It might look like it because we've tried different wheel heights.

 

[Huggedugge1]

A stainless "beam" of that type with only a fillet weld on one side will become like a banana and in the wrong direction after it is welded..

Then the question is whether you are thinking of ferritic or austenitic stainless steel? The E-module is not the same for them..

Straightening stainless steel is not at all the same as with carbon steel, so no, not a good idea..

Moreover, the coefficient of thermal expansion in stainless steel means that great attention must be paid to the choice of dimensions of the pieces to be welded together, as well as the welding sequence..

I would say that the cost of this solution will be very expensive..

Do you have to use stainless steel? Can't you rust-proof regular "carbon steel"?

 

[Bernieberg]

I think I understand in principle what you mean. Mounting an upside-down T-beam seems possible and the existing appearance would be retained. Replacing the existing joists from 120*45 to 170*45 is no problem.

Then I become unsure of what you mean by cutting out in the beam and how to solve the attachment on the short sides. Do I understand you correctly that the beam should be tensioned in the middle with about 30 mm when welded?

Should there be a similar beam in the middle?

Stainless steel, can't one paint the sheet metal with rust protection?

I'm thinking of a solution with the reinforced support beam on one side and keeping the support beam on the other long side as it is now. Just one support beam on each long side and none in the middle. Instead, all joists running across should be continuous between the support beams. The dimension on them, considering the height, seems appropriate to set to 170 but might need to be doubled or set closer than cc600 to get ok deflection in the transverse direction.

When I talked about cutting into the beam, I meant during manufacturing if you choose a cambered beam. Then I thought you cut out the web (the vertical part of the beam) so it gets a half-moon shape unloaded and then weld the flange on. This as an alternative to starting from a straight finished beam that needs to be bent.

I don't see any major problems with attachments on the short side? The steel beam can be integrated by making holes in it and then screwing into one joist, through the hole in the beam, and into the next joist. Then you attach the other wood to the 170-joist that lies against the beam.

For deflection, it makes little difference whether it's stainless or regular steel; regular steel is even slightly stiffer but marginally. I suggested stainless because everything by a pool tends to be stainless, but it might turn out to be too expensive. If you can protect regular steel from corrosion, it works just as well and is perhaps an even more reasonable way forward.

 

[Huggedugge1]

I would consider approaching the problem from a completely different angle. The fact that it bends when it rolls out is not "really" a problem until it reaches the other side and hits in. If you were to put a small "ramp" in the form of a wedge in the middle, preferably with a series of ball-bearing wheels and a small "wear surface" in the form of steel or a plastic with good wear resistance, then it would solve it by lifting the cover when it reaches its final position.

The disadvantage would be that you get a protruding part in the middle of the short side of the roof.

An alternative would be to have some mechanism on the edge to lift the roof when it comes. Say you have an L-shaped bracket where the lower part of the L sticks in under the roof when they arrive. As the roof arrives so that "edge meets edge," the L starts to rotate backward, so that the lower part of the L lifts the roof up and is completely upside down when the roof is at the right height to roll further. The simplest way would be to have two Ls that are back to back (like a rigid Z or a half swastika) with an axle in the middle. Then you can have a motor or similar that rotates it half a turn at the right pace to match the rollout.

If you want to reduce the risk of getting stuck in it during other times, you can either have a cover over it that folds away or let the entire mechanism be pushed out when the roof is sent out.

Difficult with children and dogs unfortunately, as well as the concrete pool wall. However, we were considering welding on rails and sanding them to build a "hill" that would lift up the pool side. But I don't think it will work.

 

[Derbyboy]

I don't see any major problems with attaching to the short side? The steel beam can be integrated by making holes in it and then screwing into one beam, through the hole in the beam and into the next beam. Then you attach the remaining wood to the 170-beam that is against the beam.

2x170 over a span of just over 4 meters won't work. It will sway.