OWSJ needs strengthening, how do I model it without shop drawings? 5

[ANE91]

First time poster

We, a private consultant, performed a load survey for an owner on their large warehouse building. There were precipitating load events prompting our survey. In one area, we discovered that the 30K7 roof joists supported four new 6-inch SCH10 fire suppression lines. The lines are attached not-quite-concentrically between panel points. New Millenium says the joists don't calc out (but will not provide their calcs or models). I am tasked with modeling each joist to include the four "new" point loads. I have access to RISA-3D and VisualAnalysis.

How do I model this joist? I do not know the precise double-angle section comprising the chords nor the crimped web channel properties. I looked into "Virtual Joists," but they seemed intended for initial design prior to requesting SJI sizes from the manufacturer. This is just a joist and not a joist girder. I do not have shop drawings. The as-builts from the city do not contain joist details. In the past, I have recreated such joists in a model when provided the as-built joist detail containing specific dimensions, section properties, joint assumptions, etc.

SJI provides strengthening details and calc sheets. In addition to any flexural strengthening (e.g., with tensile rod reinforcement, etc.), I anticipate needing struts between the point loads and the bottom panel points. I cannot know exactly where and how to reinforce the joist without first modeling it precisely to see where the overloads (if any) are.

There are no alternatives. The lines cannot be moved. I do not want to specify excessive supplemental steel "hoping" it works without first calcing it out.

 

[271828]

I think you will have to measure the chords and webs, develop your model, and evaluate strength using the SJI Spec.

 

[jerseyshore]

I've done this dozens of times. You have to go up there and measure each of the truss members for modeling purposes. Thicknesses are very important and make sure you measure multiple web members as it is common that they vary from middle to end.

I've only used Ram Elements for this and that program does have some preset sections that can be used for modeling a reinforced section (example is angle with a piece of rebar).

 

[JStructsteel]

Gotta love the joist industry....we checked your design, it doesnt work, but by god we are not going to help you beyond that.

You will need to measure and model.

 

[ANE91]

I think you will have to measure the chords and webs, develop your model, and evaluate strength using the SJI Spec.

This is what I was afraid of. Thank you, I believe your answer (and the others like it) are the best I am going to get.

 

[dylansdad]

I understand that you already know what your joist size(s) is/are (30K7). Plot your allowable shear and moment diagram against your proposed shear and moment diagram. There may be load reversal in some of the web members due to concentrated loading. Reinforce the chords and webs (SJI has some tools to help you with this) where necessary. Usually the addition of cold rolled bars in the fillets of the top and bottom chords and angle members to reinforce the web members will get you the capacity you need. Also check the joist seat which may need reinforcing.

 

[dold]

+1 for dylansdad's method. That's how we've always done this sort of analysis. I've used Alex Tomanovitch's "JOIST.xls" spreadsheet to compare the "capacity" M&V diagrams to the "actual" or "applied" M&V diagrams, then provide additional reinforcing for the additional required moment and shear capacity.

It's a pretty simplistic approach, but if you end up trying to model each joist in FEM you'll struggle to get the joist to work even without the additional loads, just from a modeling perspective and assumptions used. Plus, SJI spec uses all sorts of funky K-values for their members which are a pain to track down and apply in the model. Kind of like how it's damn near impossible to get a PEMB frame to check out if you model those members. They have all sorts of sneaky tricks and exceptions they like to use to economize the design.

 

[KootK]

+2 for dylansdad's method. Elaborated:

1) Assume that any local issues associated with the load being applied mid panel can, and will, be solved by adding in some webbing.

2) Check the envelopes to guage the scale of the overstresses, if any.

3) Proceed to the more onerous modelling work only if you have to.

 

[jayrod12]

Call me a third vote for the reinforce only where necessary based on the original capacity envelope. Localized chord bending is solved with additional web verticals from point of load application to opposite chord panel point as per KootK's point 1.

 

[dik]

In the days of the sliderule, it was a matter of checking the ql^2/8 BMD and the As*fy capacity of the chords and reinforcing for the difference, determining the extent of the reinforcing based on the length required to produce a moment equal to the difference and extending this reinforcing beyond the panel points. Diagonals were then checked for shear... For large trusses it was similar except cremona diagrams were used. This can still be used today...

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[ANE91]

We tried to field-measure every section, today, but were missing several thicknesses and lengths due to access issues. We put the idea of building a model to bed.

I ended up hand-calculating global shear and moment diagrams. I backed out axial chord forces from the maximum moment and joist depth. I had to assume the distribution of shear stress between web verticals/diagonals. I used SJI's free calc sheet (located here) to check supplemental web struts and chord reinforcement. Welder is on site tomorrow and eager to get started.

Thank you dylansdad et al. I hope to be of similar service to you in the future.

 

[JLNJ]

The thing which makes me uneasy is the welds. I've been to the shop and seen joists being assembled. They spend MAYBE two or three seconds per weld per panel point. It's crazy fast. I have no idea how you would check such a thing. Any web reinforcement should probably include panel point connection reinforcement.

 

[jerseyshore]

Good tips for that method and spreadsheet from SJI. Certainly useful when access is not there to accurately measure each member. I do prefer modeling since it doesn't take me long, but it can be difficult to get past a lot of the existing members not working with no increased loading. Always useful to have multiple methods.

The thing which makes me uneasy is the welds. I've been to the shop and seen joists being assembled. They spend MAYBE two or three seconds per weld per panel point. It's crazy fast. I have no idea how you would check such a thing. Any web reinforcement should probably include panel point connection reinforcement.

For me, I find myself adding angles to the webs as reinforcement majority of the time welding them directly to the chords, bypassing whatever crappy weld the existing webs have.