Column Embedment into Base Plate

[ClarksonEng]

A legacy practice of my engineering group has been to cut a hole in our base plate and embed an 8x8x1/4 HSS column 5/8" into the plate. The new manager wants detailed calculations on each structure, including base plate design. It is my understanding that the individual that detailed our base plates went with this design due to the moment on the structure. Unfortunately, this is a unique design and I cannot find any similar calculations to analyze this plate. The TIA, AISC, and ASCE 113 don't answer this question. To continue with my analysis of the base plate, how do I determine if this plate is adequate? The plate is A36 and column is A500 grade B, design loads are as follows:
Bearing Case: Shear = 4kips Axial = 9 kips Moment = 55kip*ft
Uplift Case: Shear = 5kips Uplift = 6kips Moment = 49kip*ft

Any direction would be appreciated. Also, Section 2-2 isn't shown correctly. The hole is completely through the base plate and the column is being held entirely by the 5/16 weld on top and 3/16 weld on the bottom.

 

[transmissiontowers]

We use this connection all the time. ASCE 113 has a section on the design of base plates (I am the Chair of the revision committee) and this method is covered.

The general method is to find the anchor bolt loads (I assume you have the base on leveling nuts but it will work for plate on concrete or grout), then you find the plate bending moment on the assumed bending plane (we usually limit the length to 12 times the thickness) from the highest loaded anchor bolt. In your case the bending plane will be from the corner of the HSS to the plate edge and a second plane will be across the corner at a 45° angle. Once you have the bending moment on the plane, you see what the bending stress is and solve for thickness.

You can mesh up the plate in a FEM and load the bolts in axial and shear and look at the stress in the plate elements, but be careful of the stress concentration at the joints. If you make a constraint line along your assumed bend planes so you get joints along the lines, you can find the moment along the plane and see what the average stress is along the plane. I use GTStrudl and they have a base plate wizard that will do the meshing. If you just take the maximum stress from the FEM, you will get stresses much higher than yield at one or 2 nodes but it falls off quickly. If you want to be ultra conservative, you can make the plate thick enough so that no node is over yield, but the average stress over the bending plane has worked for 40 years before we had computers.

In general we put the base plate on leveling nuts without grout to allow the water to drain. If the gap between the top of concrete and bottom of base plate is less than 2 times the bolt diameter, you can neglect bolt bending. This is allowed in ASCE 48 and AASHTO as well as ASCE 113 in the current version. The new version being edited now may recommend that the bending not be neglected, but there is no evidence that the bolts fail during extreme events and we have been neglecting bolt bending for over 40 years.

Sorry for the long rambling answer, but I ate too much Thanksgiving dinner and needed to get my brain working.

_____________________________________
I have been called "A storehouse of worthless information" many times.

 

[ClarksonEng]

Yeah we use this design on leveling nuts to allow water to drain and also for ease of galvanizing. Using the bend line method the plate is of adequate thickness, but I'm still not 100% confident about this design. The bend lines don't take into account that there is a significant reduction of area with that hole in the center of the plate. I only have Staad.Pro and RAM Connection, and neither help with analyzing this design. I don't think I could justify to management buying software or design standards for this one application. How do I calculate the stress in the plate without that area included? Since it is close to the N.A. it may not be required but I want to be confident with that before approving the design as a "standard." Also, how do I evaluate if the 5/16 weld on top and 3/16 weld on bottom are adequate? Assuming a solid plate the 5/16 weld on top is plenty, but again I'm not sure how to check with the column suspended in the center of the plate relying on both welds. I was thinking of creating a moment couple then adding the strength of each weld in shear.

Thanks for the response, hope you enjoyed your Thanksgiving.

 

[transmissiontowers]

The bend planes do not include the hole. For the highest loaded bolt, the bend line goes from the corner of the hole to the edge of the plate so that the edge takes all the bending stress. For the diagonal bend plane that crosses the corner, you sum moments about the bend plane for the highest loaded bolt. There is no canned program to do base plate design. You could do a spreadsheet (I have a couple of them) to document the process but it all boils down to following the very basics of Engineering and drawing a free body diagram of the bending plane and sum shears and moments about the plane to find the stress Mc/I. I don't do Mathcad or any of the engineering programs, but I assume you can document the process and equations.

I've done hundreds if not thousands of base plates over the last 40 years and you still follow the basic principles you learned in school. I don't do any building type structures and don't look much at the AISC manual, but in substation design, there are no standard connections from beams to columns. You have to follow your basics.

For your welds, you do about the same process. If you can find the old Lincoln electric book by Omer Blodgett on design of welded structures, it has weld design procedures for many welds (IIRC we bought ours for $7.50 back in 1975). The basic process is to find the moment at the base plate and use the section modulus of the weld pattern (there is an equation for 2 squares separated by the plate thickness (1/2 of it anyway)) You solve for the fillet weld thickness needed to give the weld the correct stress allowable.

https://www.amazon.com/Design-Welded-Structures-Omer-Blodgett/dp/9998474922

You can document the process and do a few spreadsheets if you have more than a couple to do.

As far as convincing your management that the plate and welds are sufficient, you will have to show them the calculations you did and get them to buy off on the process. Some experienced engineer at for firm must have done some hand calculations before the computer came in and made life so easy.

If you have the time, you could make up some loads and design a small column and baseplate, have them fabricated, and go out and test the connection to failure by putting a big moment on the connection to show everyone the concept works.


_____________________________________
I have been called "A storehouse of worthless information" many times.

 

[transmissiontowers]

Doing a quick hand calculation on your plate with 55 ft-kip moment, the bolt load is 27.5 kips. The moment arm is 2" to the bend line with a width of 4". The bending moment is then 55"k on a 4" wide strip of base plate. If you solve for the thickness required, I get 1.51 inches to limit the stress to under 36 ksi. I would use a plate 1.625" thick of A36.

_____________________________________
I have been called "A storehouse of worthless information" many times.