Wood Lateral Analysis Software (Structural)

[blondie_3]

I am a structural engineer working on a 5-on-2 wood over concrete podium job. For the wood design, we built a RISAFloor model and then exported to RISA-3D for the lateral analysis. The results we are getting from RISA-3D are essentially junk though (see below) and when I reached out to RISA showing them this, their only response was that FEA is difficult to understand and I should try a smaller mesh size. The recommended mesh sizes were 6" for walls and 36" for floors opposed to the default 24" for walls and 60" for floors. When I reduce the mesh size to the values RISA recommended, every computer I try running the model on crashes.

I am interested in hearing what, if any other software structural engineers recommend for the lateral analysis of wood structures? Wouldn't mind hearing other options for gravity analysis as well.

Thanks,

Example of junk results - the wall panel total shear and unit shear do not agree with each other for the same load combination. The numbers below are on the drastic side of what we are seeing, but this same issue is occurring with nearly every wall in the model.

• Shear wall length: 13.99ft
• Shear wall force:
• o Provided in Wall Panel Forces spreadsheet: V = -4.00 kips
• o Back calculated based on unit shear and wall length: v = 10.506 klf
• Shear wall unit shear:
• o Provided in Wall Panel In-Plane Design: v = 0.751 klf
• o Back calculated based on shear force and wall length: v = 0.286 klf

 

[XR250]

You could probably do this by hand in the time it takes you to mess with this in software.

 

[JoshPlumSE]

Wood structures are tough to do with FEM analysis. Also, it's very difficult to de-bug your model based on the description you've given. At least because we don't have the model.

Some quick thoughts:

a) Mesh size could certainly help. The fact that the model is now crashing is problematic, of course.
b) Is there any chance that you've got some "torsion" in more model?
c) I don't like the way RISA "optimizes" shear wall nailing patterns. Personally, I think it's best to dictate what the nailing and panel thickness are going to be before doing the analysis..... Meaning, I use it as more of an analysis tool rather than a design / optimization tool.
d) You will certainly get differences between the RISA results and a force distribution based purely on wall length. Right? The wall length assumption is based on shear behavior dominating over flexure and hold down deformation. So, we know it's not entirely accurate.
e) Along those same lines, I'm not sure if RISA is smart enough to understand hold down deformation fully. Therefore, I think you might have to use a shear stiffness adjustment factor to account for it in the wall stiffness..... At least if you feel this is going to be a big deal.
f) RISA support used to be quite good. Though most of the staff has moved on to other things and the current group is no where near as experienced as they used to be. Still, I'd suggest that you keep the discussion going with them about this model.
g) I'd personally be happy to look at the model. I love connecting with my old RISA users (not that you're old!). Unfortunately, I don't have a copy of the program anymore. I can open a lot of model in the Demo versions. But, a RISAFloor to RISA-3D model is beyond what is usually allowed in the Demo. Sorry!

Caveat:
I'm a former RISA employee. I spent 16 years in the Tech Support group. Left about 5 years ago when Numetshek took over from the original owner. I didn't enjoy how my time there ended. I don't have great feelings towards their current management either. I feel like their company is dominated more by a "sales" mindset rather than an engineering mind set. Ergo I have some definitive bias in any statements or opinions I give about the current state of the company.

 

[ryaneng]

We do all the lateral design using excel sheets that we made in-house. There is a lot to them, but once you get it set up it is easy to punch in new shearwall lengths, areas & weights and then play around with the sheathing thickness and fastening patterns until you're happy with the outputs.

Here's a link to a PDF published by EGBC:

https://www.egbc.ca/getmedia/eea8ae...d_Practice_Bulletin_on_Mid-Rise_Buildings.pdf

It has some great information on the lateral design and goes through an example in detail which is pretty valuable if you want to build your own calculation sheets. Things of course may vary if you're in Canada or the States.

 

[dold]

I use RISA floor > 3d for a lot of wood lateral design. It's very easy to end up with a 'messy' model if you're new to the software and the results can be confusing and/or nonsensical. It sounds like you're also using semirigid diaphragms which can add to the confusion when comparing to hand calcs and flexible diaphragm assumptions. I do a good bit of troubleshooting for colleagues who use risa and this is a common problem to have shearwalls picking up extra load somewhere and/or dumping load into other members/nodes that are connected to the shearwalls without the user realizing what's going on. Another thing I see a lot is that people add boundary conditions (for some reason....who knows) at random locations throughout the 3d model, ostensibly to stabilize the structure. This of course will give unrealistic (and completely incorrect) results.

That said, I've had great results using risa floor>3d for lateral design. Similar to what Josh said, I use it mainly as an analysis tool, then sometimes boil down the results and design certain items using custom spreadsheets. The optimization routines can be simplified by modifying the redesign lists (e.g., force the program to only optimize nailing patterns and use only 15/32" RS with 10d nails, as opposed to selecting any of the panel/nailing groups).

Would you mind sharing your model? Feel free to scrub any project information. At this point I'm just curious and maybe I can give some pointers.

 

[JStructsteel]

I use the Woodworks software for shear walls. For beams, joists etc. Either the BC software, Forte, or enercalc.

 

[blondie_3]

I appreciate all the replies and information. I believe I will be pushing for Excel spreadsheets on my projects moving forward.

JoshPlumSE: I follow what you are saying about the tension and compression forces not correlating exactly with the shear values but my main concern is the unit and total shears for the same wall for the same load combinations not correlating. Is RISA providing a maximum unit shear rather than the average unit shear over the length of the wall? The program uses the unit shear when checking the shear capacity so if it is using a peak value rather than an average, it would lead to excessive design by my estimation. I have increased the mesh size for walls to 12" and finally have the program running without crashing a computer. It has been running for over 3 hours now and still has some time to go. Even if it produces better results, this does not feel like an efficient or effective way to do our analysis or design.

dold: I am not an expert in RISAFloor as the previous firm I worked at did not use it and instead used spreadsheets. The engineers who built the model for us familiar with the software and I have spent a lot of time learning it in the last couple months. It is possible our model could use some fine tuning but it does not report any errors and we certainly haven't added boundary conditions to resolve instabilities.

 

[JoshPlumSE]

Is RISA providing a maximum unit shear rather than the average unit shear over the length of the wall? The program uses the unit shear when checking the shear capacity so if it is using a peak value rather than an average

If it's just a single wall with no openings then it should just be the average for the whole wall.

If there are openings, then it depends on your design method for the wall. I believe there are 3:
a) Force Transfer Around Openings: Which looks at each region above, below, right and left of openings and reports the max shear for those regions.
b) Perforated wall: Which uses a prescriptive method based on the NDS' SDPWS code.
c) Pier method?: I might have this name wrong. But, it basically puts hold downs (or straps) for each "active" pier and then softens the rest of the wall so that it doesn't participate at all. This method was meant to mimic traditional hand calc methods. It's not guaranteed to match because of the flexural vs shear contribution to stiffness.

Regarding your excessive solution time:
Three hours is certainly a long time. How many Load Combinations are you solving? Often times, this is one of the main causes of excessive solution time. This is a personal frustration of mine as some code committees seem to require very weird combinations of all kinds of things that leads to hundreds of Load Combinations that aren't really necessary.

Another thing to look for is any relatively small offsets between semi rigid diaphragm and shear walls. Or, beams that frame into a shear wall near the end, or almost in-line with an opening or such. You should be able to turn on your mesh after a solution to see where the program may be meshing excessively.

 

[dold]

It has been running for over 3 hours now and still has some time to go. Even if it produces better results, this does not feel like an efficient or effective way to do our analysis or design.

Good heavens. How big is your model and how many load combos? Increasing semirigid diaph mesh to 48-60" generally gives fairly accurate results, and smaller diaph meshes really blow up solution times. I'm usually comfortable with using a 24" wall mesh unless there are a lot of odd conditions. You can also turn off the option to 'optimize shearwalls', which I turn off sometimes if I'm just trying to rough in a model. If you leave that on it will solve the model > design the shearwalls > re-run the model using the new stiffness matrix (due to the change in stiffness of the walls which is dependent upon sheathing, nailing, etc) > design the shearwalls > re-run the model > design the shearwalls. You can also adjust the number of iterations if you choose to keep this on. I also usually turn off p-delta for walls and generally just turn off p-delta in the load combos. It's a wood building after all, and there's no way anyone would do this using spreadsheets/hand calcs.

Otherwise you might consider just using flexible diaphragms and making sure you've got lateral members around the perimeter. And at any drags/collectors you're trying to account for. You can view the attributed loads in R3D after solving. It's a little easier to visualize where loads are going rather than digging through spreadsheets.

Try turning on your reactions (X, Y, Z) to just look around and see how the model is behaving. Ditto for simply looking at the deflected shape. If that all seems normal, you can turn on force/stress contour maps for the walls and diaphragms to look for abnormal stress concentrations and the like.

Guess we're not going to get your model to play around with...

c) Pier method?: I might have this name wrong.

"segmented"

 

[XR250]

It's a wood building after all, and there's no way anyone would do this using spreadsheets/hand calcs.

I am certain there are plenty of folks doing that.

 

[CConrad]

ENERCALC has some tools that may be useful for a reality check/second opinion. (It is by no means a full building solution.)

The Torsional Analysis of Rigid Diaphragm module might be helpful for checking the force distribution to the shear walls. It only has the ability to assume that the diaphragm is totally rigid, but it may serve as a back of the envelope check in this case.

If you can get comfortable with the shear distribution to the shear walls, the Wood Shear Wall module may be useful. At this time, it only implements the Individual Full-Height Wall Segments method.

Director of Technical Services
ENERCALC, Inc.
Web:

http://www.enercalc.com

 

[dold]

It's a wood building after all, and there's no way anyone would do this using spreadsheets/hand calcs.

I am certain there are plenty of folks doing that.

But...why? I've never come across anybody doing this. I've never even heard it discussed in regards to wood (apartment, etc) buildings, except when discussing risa.

 

[phamENG]

dold - in areas with low seismic activity (strength and serviceability design driven by wind) and layouts of simple to even moderate complexity, shear wall design is pretty straightforward. It takes a little bit of book keeping in some cases, but for many apartment buildings it's quite feasible to be 95% confident in the framing and lateral scheme in the time it would take somebody to build a RISA model. No debugging required.

Not the case for everything, of course - RISA is quite useful for complex layouts, layouts that don't stack, etc. But for the more straightforward designs (that save developers money), RISA can be a bit much.

 

[XR250]

What did folks do before RISA? Usa a slide-rule?

 

[dold]

My most recent comment was regarding p-delta when designing these buildings. Of course we've all designed these these buildings with spreadsheets, but i've never seen p-delta even discussed re: apartment buildings. Sorry if there was confusion. Or are you guys contending that p-delta is something everyone does...with spreadsheets? Surely not?

 

[Aesur]

We typically use a combination of tools, most commonly excel to keep track of loading/shear walls and EnerCalc for shear walls. When we need force transfer or other shear wall types, we use hand calcs or excel files. We also have an excel file that does semi-rigid analysis and distribution of forces based on stiffness for the larger projects (apartments, etc.).

When we have a large straight forward apartment or similar building with shear walls at 90 degree angles to each other we have been using Woodworks Shear Walls as it saves significant time once you learn the ins and outs of the program and the "tricks" to make model the structure faster.

I haven't used RISA for shear walls successfully to date, the few tests I ran gave results that didn't come close to matching hand or excel spreadsheets that have been around for years, including engineering-international force transfer spreadsheets. Currently I stick to RISA for 3D frames only at the moment.

 

[phamENG]

Ah. I see. Not as much, no. I do try to consider drift induced moments in key columns, but keeping track of it can be challenging. Those are rarely a significant concern in light frame wood, though, as it's almost exclusively bearing walls. And interstory drifts aren't usually enough to matter. I usually design wood walls with a d/6 eccentricity to begin with, so unless it's a side loaded column that's usually enough "cushion."

I'd rather consider it specifically...but apartment developers don't exactly pay for that level of detail...