I suspect that welding would be easier than bolting a plate to the bottom of the existing beam. That said, here are two suggestions,

Make the plate wider than the bottom flange so that the welds are downhand (versus the much harder overhead welds). Your welder will thank you.

If there is sufficient headroom, consider using an HSS (wider than the bottom flange) instead of a plate. You’ll get more added flexural strength per pound of steel with an HSS, and HSS’s are available in longer lengths than plates (eliminating the need to splice multiple sections of plate together.

Reasons to avoid using bolts: Precise field drilling of bolt holes will be required. Field welding splices between multiple sections of plate will still be required depending on how long your beam is.

But of course you still need to determine the weldability of the existing beam.

Welding (if possible) does seem like the better option...unless the beam will be subjected to cyclic stresses substantial enough to induce fatigue. The fatigue stress limit at the welds will be lower than for a bolted connection.

I'll disagree with cliff234 on one thing, though. Drilling the holes is not that difficult if the cover plate is simply clamped onto beam and the holes in the beam flange are drilled through the holes in the cover plate. This is fairly common for bridge girder retrofits.

Quote (ENGIRL)

My only two options are : bolting or welding a new a plate to the existing beam to increase its capacity.

Why have you concluded that?
You surely have several other options. One way is to use two bolted channels sistered to the existing beam. The approach of sistering isn't the most efficient from an engineering perspective. But it is quick, easy and practical. You have the ease and simplicity of bolting without having the challenges of potentially requiring slip critical bolting. With the right channels you can potentially double the capacity of the beam. (depending on the beam and the type of channels available)

BridgeSmith - You're right! It never occurred to me that first clamping the plate to the bottom flange and then drilling holes through both would be very easy - and would guarantee hole alignment. Thanks!

Thank you so much everyone. I will discuss all of your suggestions with my team before meeting with the client.

I like human909's proposal. That allows drilling the beam web (thinner) horizontally rather than the flange vertically from the bottom. And the bottom flange solution would require enough bolts for composite action, while the web bolting would only need to transfer the shared load.

Agree that channels have a lot going for them. Plating one flange can give disappointing capacity increase in section properties as well and sometimes double whammy from losing little code capacity bonuses that apply to symmetric sections.

One disadvantage to bolting channels to the webs of existing beams is that this does not work if the existing members are girders (i.e., if members are framing to the webs of those existing members).

Fair enough Cliff. Depending on the number of framed connections, channels could still be an option in addition to the bottom flange plate at those framed connections.

We had a little overhead to spare so we welded channels leg-down to the bottom flange.
We selected sizes that gave us 7/16" of steel on each side.
Easy to weld an added a lot of stiffness.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

Quote (ENGIRL)

The building is very old and I am not even sure if the steel is weldable (investigation/destructive testing required?)

On a previous project, welded connections caused major problems because of this. We got very far into the process, and some connections were already welded in place. THEN, someone upstream realized the steel is a few decades old and had uncertain material properties. They started naming the requirements that would let them sleep at night. It took quite a bit of work and money to fix that.

Going forward, the only way I would weld to existing is if it's shown on the EOR's drawings and clear enough that it's on them if someone changes their mind later. Bolted is the default at this point.

Just adding a plate to the bottom does not increase the beam's capacity all that much. You might want to check your calcs.

Quote:

Just adding a plate to the bottom does not increase the beam's capacity all that much.

It depends on how big of a plate is added, and whether the design is limited by bending capacity or deflection.

I just checked a plate girder I had the section properties handy for. Doubling the bottom flange thickness only increased the top flange section modulus by about 10%, but it increased the moment of inertia by 34%.

You may want to ask the contractor for their opinion. Bolting is pretty easy if it can be mag drilled.

How would you make the channels work in the solution of human909? How would you get the load to transfer to those channels? Is it just a matter of getting a lot of pre-tensioned bolts, and connect the bodies by friction?

The bolts just work in shear to transfer the strength deficit from the wide flange beam to the channels. The channels don't need to be full length, as the wide flange will doubtless take the shear without reinforcement, provided the end connections are adequate.

@holbewoner

Friction bolts are not necessary. You are only transfering shear. The biggest differential deflection between the beams is the hole clearance so you are guaranteed to engage the channels unless deflection is tiny. You have deflection compatibility.

While I'm replying I will note that a_urbs's comment about stopping and starting the channel where connections are framed into the beam is highly problematic and cannot readily be addressed by using a bolted plate underneath. You have all sorts of stress flow issues.

EDIT: hokkien beat me to it, I started my reply an hour earlier! As hokie says the channel doesn't need to be full length but I would point out that it does need to be continuous.

So it might be not a good solution if there are intersecting members as has already been pointed out.

How old is very old? If you choose to weld- I prefer bolts- you might want to consider taking a coupon and testing it to ensure it’s weldable.