Evaluating Existing Steel Truss 13

[VBI]

I need to evaluate an existing steel truss system in order to see if it can support an additional 5 psf roof load. There would be great difficulty in measuring the existing truss as it is a commercial building located high off the ground.

Is it acceptable to approach this by measuring existing deflection. If the existing deflection is minimial than it would be safe to say that the truss system can support the minimal additional load. The building is over 50 years old and determining information or a manufacturer is not really possible.

Thank you,
Vinny

 

[ERV]

It is possible to use existing deflection, however, original camber in the truss is a critical factor in determining deflection. Be careful. If the bottom cord is flat then the truss is probably loaded to its intended design. If there is positive camber in the bottom cord you may have extra loading capacity. Be sure to allow for live load deflections.

 

[sharkksu]

I will have to disagree on using deflection as there are too many unknown variables to account for, and as ERV mentioned, camber being the main factor. In my opinion, you should use the current building code to establish the appropriate loading and determine rather the truss has capacity. If not, it should be strengthened.

 

[krus1972]

VBI,

I have evaluated many different structures and trusses in the same manner as your truss. I have also designed many new trusses as well. It sounds like no one wants to spend the money to have the truss reversed engineered. That is no one wants to spend the money to hire a testing company to go out there and measure up the entire truss and provide you all dimensions, the size of the members, material grade, all connection details, etc,etc,etc.. in order to make a firm and solid decision. The deflection method is not an acceptable method because you do not know what the non loaded deflection (camber) was. Sometimes trusses are designed to have camber and sometimes not. I've seen and designed trusses for your application both ways. If the truss was designed WITH camber you do not know how much camber was specified. Every case is different. Therefore the deflection method is not acceptable to make a justifiable conclusion. Fortunatlly there is an easier way to evaluate the truss given the lack of information you have.

With a lack of information such as in your case you need to evaluate the truss based on a few assumptions and engineering judgement. First you need to ESTIMATE how much load the truss may have been originally designed for. If it's a roof truss (by your post it sounds like it is) you can estimate the original design load very easy. Visit the area and sketch up the geometry of the protion of the sturcture (all of the stuctural components) that directly or indirectly apply load to the truss. From the ground you should be able to come up with the length of the truss (i.e. 40, 45, 50, etc....). Sketch up the geometry of the truss. Take notice of the top-chord and bottom chord panel points and make sure you count the correct number of each. By visual inspection determine if the panel points are equally spaced. If it's hard to look at the truss from the gorund use a pair a binoculars. By doing all of this you should be able to gather enough information to come up with a tributary area for the protion of the roof load the truss supports. Try to estimate the depth of the truss. You can do this by making a comparison with something else thats near by with a known dimension. Another idea would be to measure from the bottom chord to the finish floor and then measure something else in the general area that extends all the way to the underside of the roof. Subtracting the two dimensions will give you a good truss depth to work with.


Take special note if the roof is relativly flat or sloped by visual inspection. If it is flat then only gravity loads need to be estimated. If the roof (and top chord) is sloped or next to a higher roof you may have to take wind and snow drift into account in addtion to gravity load estimation.

Use the current governing building code to estimate the design live load of the roof. Secondly estimate the dead load. In heavy industrial applications using W10-W12 roof purlins with 20ga metal deck, fabricated roof carrying & Jack trussess 20psf dead load is a conservative number. In your commercial application you may want to reduce this number a bit since the roofs are lighter in weight. With your bionoculars you should be able to identify the dead load materials within the existing tributatry area and make an estimate on what dead load to use.

Once you have sketched up the original truss and estimated what the original design loads of the existing truss you can perfrom a simple analysis on the truss. You do not need to to know the material sizes to make a reasonable engineering judgement. Perform the analysis with your estimated design load and then again adding the 5PSF to the estimated design load. For each member of the truss you will have to sets of stresses from the each analysis.

For each member of the truss calculate a PERECENTAGE of ADDITIONAL load each member must carry resulting from the 5psf.

Now this is where past experiance and engineering judgment come into play. I allow each member to carry 10% more load then the original design load. This includes the end connections and the connections to the joints. If any member is more then 10% then I reinforce it or take a closer look at everything leading up to this number for this member. There are many things like live load reductions that can help you reduce and professionally justify reducing the original estimated design load on the truss to avoide reinforcing.


This method gives you a very good idea and "FEEL" of what the impact of the additonal 5PSF is going to do to the existing truss without no information. The higher the existing design load for the truss the less the percantage increases and more likly you will not have to reinforce. If there are existing roof top units contributing to original truss design loads then this will work in your favor.

I hope this all helps and if you have any questions please ask.

 

[JAE]

krus1972 -
Your overall concept does work, but I'd like to point out a couple of considerations -

First, with your method, you do not have to model anything. You just have to determine the original load on the truss (in psf), both dead, roof live, and snow if applicable.

Then add the 5 psf to that load. Axial forces in the truss are LINEAR - 35 psf in a truss, any truss, will add 16.7% to the shears, moments, and axial forces in any member or connection compared to a 30 psf roof.

Secondly, you cannot estimate existing dead loads with a pair of binoculars. You would see the trusses and the bottom of the roof deck, but not the materials above the roof such as sheathing, insulation or roofing.

So you should investigate the make-up of the roof materials by looking for original building plans or by actual on-site investigation. If unknown, you can simply estimate them from current building practices. Using your method, you would want to UNDER estimate the dead load so that the 5 psf added load is conservatively larger compared to the total.

Finally, the 10% is a little higher than what I know most engineers use (we use 5% as a max.)....but this is a subjective decision.

Deflection can be used to check capacity, but only under a load test methodology. The IBC has section 1713 that outlines an in-situ load test method that depends upon deflections.

 

[KarlT]

One thing to consider is the original snow design loads. Here in Winnipeg, Canada, the old building code resulted in a uniform roof snow load of 36 psf. The newer code has reduced this value down to 32.5 psf, so one could squeeze a few psf out of an existing roof if you can confirm the original snow loads were higher.

Are you able to determine the age of the building, or get your hands on a set of the original structural drawings?

If you can find the original assumed roof dead loads, you may also be able to squeeze a few more psf out if the actual roof dead loads are less. Typically, roof dead loads here are assumed to be around 15 psf for a nice even number, but the actual dead load is usually less than that.

 

[LPPE]

One other stab in the dark - If you can find a joist tag at the end of the joist (almost every bar joist structure I've seen has at least one tag left on a joist somewhere), grab it. If it's a newer structure (circa 1992 or newer), you should be able to call up the manufacturer and give them the job number and they can tell you what size joist it is.

 

[jheidt2543]

JAE is right about checking the existing dead load. Some of those old buildings have more than one built-up membrane roof on them, so that could be 6-10 psf of additional load before the 5 psf you are adding. Also, some of those old roofs leak like a sieve in the rain, which means the roof insulation may be or get saturated which adds additional load.

 

[VBI]

THANK YOU ALL FOR SOME GREAT FEEDBACK!

Although adding 5psf sounds minimal - I understand the need to properly evaluate the structure. Based on your feedback I will FIRST attempt to look at the existing dead loading. I will assume that I will have the difference between 15 psf and the actual existing dead loads to play with.

Another monkey wrench is that I believe our Live Load requirements has increased by 5 psf in our latest building IRC code. If that's the case I have to figure for a 10 psf additional loading capacity when making my evaluation.

Again - your feedback is GREATLY appreciated. I'll let you know how I make out!

 

[jheidt2543]

VBI,

I do believe that when you are evaluating an existing building you can make an analysis of the live loads, just as you do the dead loads, and use that as a basis for checking capacity. I think that, based on your engineering judgement, you can call the building "safe" based on the ACTUAL dead and live loads, even if those loads are different than the current code requirements. BUT, you must document this and I would also have the Owner post the maximum allowable safe loads.

 

[VBI]

Ah ha ... After I SAW the truss I determine the following.

Corrugated Metal secured immediately to the truss system. So I can figure a maximum of 5 psf dead load.

The GOOD news is that if I assumed it was designed for 15psf (or even 10psf) dead load as some of these posts implied - then I'm OK to add another 5psf.

The bad news is that I will DOUBLE the existing dead load and certainly increase the stress'in the members by over 10%. These posts assert that the 10% should be a limit before strengthening.

It would be quite exhaustive to measure this truss system as it would require a man-lift of sorts just to get close to it. (40' in the air)

My gut (engineering) feeling is that the 5 psf is minimal and the structure will be able to support it. However...I did want some logic behind it.

 

[krus1972]

VBI,

You should estimate the live load from the governing building codes as described earlier. You should be able to make a good estimate for a design live load to work with. Secondly you should try to visit the tributary area on TOP of the roof. Take note of the type of roofing is used. Most importantly take note if there are any existing HVAC units or other items that can add to your DEAD LOAD estimate.

Lastly by inspecting the surrounding area inside, is there any suspended lighting, pipes, or utilites hanging directly from the truss or around the general area? If there is then an additional 5-10 psf (for hanging utilitites)can be added to your original dead load design load. Utilitles are usually not hung directly from the metal deck but if they are they still contribute to a dead load. Even if there is NO utlites found chances are that a 5-10 psf for utilities load was added to the existing dead load of the roof. In situtations like this many people forget that many trusses are designed for a dead load utility even though there is nothing hanging from it or near it. This will increase your original dead load estimate and bring your percentages down.


The idea here is to make REASONABLE estimates on the TOTAL loading and try to make the original design load as high as possible. That way your percentage increase in loading becomes less. Adding 5-10 psf utility to your roof dead load is OFTEN overlooked and would definatly help you out. I would look at the area again and make an engineering judgment if you can justify that the truss was designed with a hanging utility load.

Good Luck,

Jeffre Krus P.E.

 

[jheidt2543]

VBI,

A few more commnents:

Is the corrugated metal deck the only roofing material? There is no built up roof, no membrane roof only the corrugated deck - like the old Quanset huts? If so, that would be unusual in this day and age, since there would be little insulation, but it is possible.

It would not be prudent to ASSUME that the original design was for 15 psf dead load and since you have "determined" there is only 5 psf dead load that you can add another 5 psf. I wouldn't stake my stamp on an assumption.

I have measured a number of existing buildings in order to check their original capacity. In fact, I'm doing one next week. An old mill type construction with 8"x10" wood roof trusses built in 1907 spanning 68' and it is 45' in the air and I am renting a boom hoist to inspect it. The only way I know how to judge the capacity is measure each member, assess the loads and run the numbers. It takes longer, but I sleep at night, besides it is actually fun to see how these buildings were designed and built before calculators, computers, FEM and CAD. That's when Engineers were ENGINEERS!

 

[krus1972]

VBI,

One thing I forgot to mention. Please be carefull at which code you use when you run your final load estimates and document your final calculations for your permant file. The LRFD says to factor each load with a different factor (1.4 for Dead and 1.7 for Live I am not certain on these) before using in an analysis. The ASD code DOES NOT permit you to factor the design loads for analysis. You would just add the dead and the live load and use the total in the analysis. Since the Live Load & Dead load dactors in LRFD are different numbers AND your Live Load & Dead Load estimates will also be different numbers it's hard to say which code will work in your favor. You need run an analysis on the same truss using the same loads using the two different codes.

After you are confident in your total load estimation I would analyse the truss using BOTH the latest LRFD AND the ASD code as described in earlier posts. BOTH codes are acceptable governing bodies for the analysis and design of structrual steel. Which ever code works in your favor and keeps you under the 10% increase in total load for each memeber is the analysis you keep.

I hope all of my posts have helped. Please ask if you have any more questions.


Jeffrey A. Krus P.E.

 

[CSEllc]

I habe read "most" of the responses to the question with great interset as I have been faces with similar problems over the years. With that in mind, I offer the following (additional and re-iterated) suggestions:

1. MEASURE THE TRUSS.

2. when you are done #1, MEASURE THE TRUSS.

3. Roof dead loads are typcally between 15-20 psf depending on the original roofing systems. The roof must have been replaced since it is over 50 years old, you might be able to reduce the DL to the 15 psf if a single ply roof was installed. Contact the building owner (or property manager) to find out what has beed done to the roof over the years.

4. Research the typical loadings used at the time the structure was built. Many areas did not have "building codes" as we have now. Quite often, the enginers in the area used a conservative number that was common to the area. For example, in my state, CT, the roof LL (snow)used on many older structure (pre-1960's) was about 40 psf. In the 70's to mid 80's it went to about 32 psf. To todays snow load is 30 psf. As you can see buildings over 40 years can probable support the additional 5 psf. Modern structures probably not. Please pay very close attention to drifting and or sliding snow loads on lower roof. Many structures, as recent as the mid-80's, were not designed for drifting and/or sliding snow.

4A. Reasearch the grade of steel common at the time for the shape and size you have. You can find that info in old AISC manuals if you have them......I try to "collect" them for just this purpose. You can also use AISC Deign Guide 15, Rehabilitation of Old Structures.. It lists mosts of the steels and wide flange shapes since the late 1880's.

5. Once you have your DL and LL, you can perform a a rather quick check, if the truss is flat, to determine if you are in the range or totally out of luck. To perform the check, you would determine the maximum moment of the truss, divide the moment by the distance between the top and bottom chords neutral axis, that will give you a tension/compression couple. You can compare the value with the aloowable for your shape and size. I would not use these values for yuou final analysis, but they are adquate as I said for a quick in the range check. If the truss seems to workor is close, then I would spend the time inputting the configuration into a anaysis program to get all the actual member forces and perform my member check. If the truss is not even close, discuss reinforcing options with you client and a local steel fabricator who has reinforced trusses before. It might be less expensive to remove a portion of the roof and metal deck and install additional joists (oopen web or wide flange) rather than reinforcing the existing. You might also be able to install interior beams and columns.

Generally, if the truss works using ASD method it also works using LRFD. However, if the truss is marginal to begin with LRFD will probably allow the additional load while ASD would not. I believe the load foctors for LRFD are 1.2DL & 1.6LL. You can check these values in either the LRFD manual or ASCE 7. Actually, if you analyze the truss using a commercially available software program you should be able to toggle between ASD and LRFD after it has done the checks for one method.

Good luck with it..

 

[connect2]

Seems to me the main point being made here is that you must perform at the least a visual inspection and survey of the truss(es) in their as-built condition. This is a good idea as additionally the member connections are equally as important as the members themselves. Roofs can leak and this can have a detrimental affect on the trusses structural capacity over time. All current loading conditions, proposed loadings and those that existed at the time of construction should be analaysed. A history of the roof should be compiled. Look for old penetrations through the roof for HVAC, and openings that may have been added at a latter date with subsequent new loads and new locations etc. Additionally since you are being asked to evaluate a structural roof system you will need to review all the trusses. Bracing, bridging, uplift, diaphragm capacity etc., Their could be trusses in the roof system that fufill more than gravity only load carrying functions. Finally since this is a roof, you should take your load to the ground where it will end up and do at the least a preliminary review of the structure. When you are saying that you can apply additional loads to an existing roof you are also by implication saying that you can apply additional loads to the whole structure below it. Sounds like a lot of work for assumed 5 psf additional load but we have seen to many cases were the project drawings and design have been changed at the time of construction for what ever reasons. If it is possible and if they do exist get a hold of the original design and the original structural engineer this info could help you tremendously. The analysis is simple enough. Don't let budget get in your way. Bite the bullet, get to work, start looking and crunching numbers. Next time around you'll include the cost of a thorough visual inspection, historical review and then detailed structural analysis. There are Guidelines published for Evaluatuion of Existing Structures, they are the minimum requirements, you'll at the least want to meet them. Nuff said, first principles.

 

[VBI]

WOW - Great advice. It is a wonder that our National Building (IRC) code allows 2 roof layers to go on on conventional roof systems without an engineering analysis.

 

[EIT2]

More comments:

Yes, most joists will have a thin gage metal tag at one end which should tell you what size the joist is, which may allow you to refer to the SJI 60-year Manual. But PLEASE! If you are an engineer who is expecting the joist tag to be there to assist in your investigation, PLEASE do not just "grab" it. Either write the information down or remove it, copy it, then PUT IT BACK! They are put there to assist all, no just one!

Assumptions must be made. But, never assume there is any unused pressure or load within the original design, allowing you to reduce or neglect those which you are investigating and intending to newly impose, unless you have existing structural construction documents with evidence!

Never assume 15 psf is max. Consider what has been said previously about multiple layers. Consider 10 psf ballast. Get on top of the roof!

Never assume a 5-10 psf utility, miscellaneous or any load that has never been employed. Budget, value engineering, human error, aiming to please (another human error), whatever the case, you do not know the attitude or motives of the original engineer.

If you are going to reinforce the existing, do not be limited to, but, consider the following:
1. Reinforce not only the top and bottom chords, but, also the web members.
2. Consider also, reinforcing the end bearing points to resist the increased reaction.
3. As was stated previously, consider the structure between the joist and the ground.

Thank you ...

 

[CSEllc]

EIT2,

The tags you are refering to are usually indicate the manufacturer and their job number and "piece" number. They very rarely (I have never had any luck anyway)determining what joist it is in the SJI 60 yr catalogue with this info.

Depending on the age of the joist, you might try contacting the manufacturer. (Most fabricators, such as Vulcraft, keep this info on file for a few (10 to 15 max)years due to the amount of joists they fabricate.) They might be able to help you determine what size it is; but you will also need the approximately top and bottom chord sizes and sometimes the first few (from each end) diagonal web sizes.

Basically, if the joist is over 10-15 years old. You are S.O.L......You need to analyze the joist.