Beam sized by chart challenged 5

[EngineerofSteel]

My beam size is challenged by reviewing engineer. "check beam size results," he wrote.

I used EnerCalc to size a beam spanning 39.58 feet. The DL is just 90 PLF and the LLr is just 200 PLF.

Using AISC tables, I get the same result: W12x19

But using tables based upon span, I get a heavier beam. Do I need to satisfy both design methods? I have been designing according to DL+LLr and the moment at the connection to the column. Now I am being asked to meet another design standard.

Does the beam have to meet both design methods, or just one. It doesn't seem correct that I should check two or more design methods and size my beam according to the largest beam resulting.

 

[swivel63]

a 40 foot span? with a 12x19?

whoa.....

 

[RareBugTX]

A few years ago, I designed a steel beam for a 40 foot span, the owner wanted a few columns gone since the space was going to be used for concert hall in a nice heavy timber building.

Came up with a W21X62. A mentor at that time noted the importance of unbraced length and was provided as required.

The project involved checking the 12"x12" existing wood column to carry the additional load.

I think it is nice that you can step back and look at things from a better perspective. The learning curve gets steep in cases like this.

 

[EngineerofSteel]

ChipB,

I am using the unbraced length on the compression flange - this is the purlin spacing, bolted to an angle and the angle welded to the top flange. I also have always used the autocalc for beam weight.


The difference in our results is the "end fixity" option. Your results are from the 'pin-pin' option, correct?

I used the "fixed-fixed" option and designed my connections according to the greater moment forces produced.

 

[ChipB]

Ahh, so we are talking about a roof beam.

Yes, I used pin-pin.

Careful:
1) In a fix-fix scenario, your bottom flange is going to go into compression about 8.5 feet out from the support.

2) Load cases with wind uplift are going to place the bottom flange into compression.

Purlins aren't going to help you in these situations.

 

[hokie66]

To brace the bottom flange, use "fly bracing", or whatever you call inclined braces from the bottom flange up to the purlins.

 

[Lion06]

I would also be sure that you can actually obtain something close to a fix-fix end condition. Just designing the connections isn't good enough if the member it is framing into doesn't have much rotational stiffness.

 

[csd72]

SructuralEIT covered what I was going to say.

Very few situations in reality are truly fixed. If you are relying on a connection to an adjacent beam or column then the flexibility of these members will reduce the flexibility. Any reduction in end fixity will increase your moment and deflection at mid span.

In order to find the real end fixities you will need to carry out a second order analysis.

Please dont get insulted by people trying to help you, we cant tell if you are a guru senior engineer or a student. People will assume you know nothing if that is what your post indicates, if you do not want to be judged wrongly then word your post carefully.

 

[EngineerofSteel]

I appreciate all the input. If anyone has not yet ascertained, my knowledge shortcoming was this: I have been told by two of the six engineers that I have worked with that a column-beam connection is simply what you say it is and how you design it. The idea of performing a second order analysis was what I needed.

Basically, I was told: "You can say this is a pinned connection and design the other members to take the loads. The column beneath the pin is then axial only. Design the connection to standard construction practices. If you design the connection as fixed, then calculate the welds, plates and bolts according to AISC, UBC, etcetera.

So, csd72, can you reference a good article/book on second order analysis to determine fixity?? Appreciated.

 

[csd72]

I wouldnt do it by hand, it would take you a week. Your employer should have a computer package to carry out this anlysis.

 

[EngineerofSteel]

We are using RAM Advanse.

I am reading the second order analysis requirements in the 2005 Steel Construction Manual. We have always checked slenderness effects. Also, I use an omega-O factor from UBC table 16-N when designing connections.

I will study up on Appendix 7 also. I like to use the software, but I want to know very well the principles and assumptions behind what is produced.

The added forces in second order analysis seem quite small (e.g. ".42 percent of gravity dead load added to lateral forces"). The structures I'm doing calcs for are ag buildings without walls. This means a greatly reduced uplift (Cq=0.7 vice 1.3 in UBC). The unbraced penalty for the compression flange in uplift usually is minimal.

 

[EngineerofSteel]

ChipB,

I am interested in your beam calculation of 8.5 feet from the connection will be lower flange compression.

I have been using a "rule of thumb" of 20% of the distance from the haunch to the ridge connection. This 40' out-to-out building has a total haunch to ridge span of 20'. Using the thumb rule, I get approx. 4' in compression.

Is there a more exact method you're using? Thanks. -DD

 

[dbuzz]

Using a rule of thumb span-to-depth ratio of 20 to 25, a 12 metre (39.58 ft) span will require a beam in the order of 480 to 600 mm (18 to 24") deep.

By comparison, your selection of a W12x19 gives a span-to-depth ratio of 39.

 

[apsix]

DairyDesigner
The more exact method is to use structural analysis software. The 8.5 ft is correct for fully fixed end supports. This distance will reduce if the supports are modelled to better simulate the actual structure, ie. by including support columns in the model.

 

[csd72]

If uplift is your critical case then second order analysis may not be necessary. Second order only gives a difference when members are in compression - mainly in portal columns but also in portal beams.

When you have tension in the members such as uplift on a portal frame, the tension tends to straighten out the members reducing the deflections and out of straightness. The net result is that the tension actually reduces the moments (by a very small amount less than 1%) so first order is more conservative for this case.

Check the code, I am sure it says that if compression is less than a certain amount then you can use the first order method. Tension being negative compression is less than any amount of compression.

regards
csd

 

[DaveAtkins]

I do not agree with csd72.

You do NOT need a second order analysis to take into account the relative stiffness between the beam and the supporting columns. You just need to do an indeterminate analysis on RAM Advanse. Model everything correctly, using the correct member sizes, and you will get accurate results.

I still think deflection will be a problem.

DaveAtkins

 

[csd72]

Dave,

I didnt mean for the beam, I meant for the columns. The assumption is that if the beams weren't properly analysed then neither were the columns.

In this day and age I dont see the point of doing a first order analysis for a frame when basically all packages are capable of doing a second order.

But as I mentioned in todays post, if uplift is the critical case then it is not going to make any difference.

csd

 

[ChipB]

DD:
8.5 feet:
Based on fully fixed supports with uniform loading. Look on pg 3-216, 13th Ed of the Steel Manual, top of the page. Moment diagram's inflection point is 0.211(l). (Hey 271828, did you notice I OPENED the new book?)

If you are haunching your beam at the ridge, is it fixed or pinned at the ridge? Defining this condition is going to affect the design of the beam as well as your columns. This would be best to be modeled as others have suggested.

I haven't messed with RAM since 1997 when the company I was working for wanted me to do a side by side comparison with STAAD. I was much more familar with STAAD at the time, and the time for building the model was drastically different with RAMSteel winning out. Haven't used it since.

For a quick 2D analysis, I use a free program called FastFrame. I recommend downloading it from enercalc.com
Chip