Chris, this is a very interesting perspective. As someone that came along during the transition from hand calcs to structural modeling, there is the perception that there are "new" equations, procedures and more accuracy when it things hardly ever deviate from the basics (i.e., Mc/I, P/A, etc.). I remember the argument over whether hand calcs should go beyond 1 or 2 decimal places. There is perceived accuracy in modeling structures with design conditions that we hope will envelope all of the unknowns simply because of the use of the computers. Changes often occur with the information gathered from real world experiences (variable combinations) not previously considered.
In designing a concrete slab for eccentric loading several years back, I was establishing the limits for the placement of a 4 or 6 legged safety platform based on the design thickness. I was using some ACI formula or equation for checking stresses, and needing to understand the basis, I worked backwards to discover that it was simply P/A (+/-) Mc/I. I do the same with structural finite element models to compare shear forces or stresses across a section with some approx. hand calc, especially when new to a structural modeling program. I also visit the disclaimers and user comments and feedback sections to view what may not be functioning quite as expected.
I did a comparison of wind pressures on a existing structure based on the code when built and the code at the time of the alteration or repair. If I recall correctly, the pressures were higher with the older code, but I will have to go back and check. It may have been that the minimum required was higher, I will have to revisit.
Interesting enough, I have found that some are more interested in how closely the model matches the physical appearance than the assumed structural behavior.
Computers have simplified what we do, how fast, and to what level of detail. The question is how accurate is accurate when there are so many unknowns and uncertainty. (i.e., Quality of materials, construction techniques, usage, etc.).
------------------------------
James Williams P.E., M.ASCE
Principal/Owner
POA&M Structural Engineering, PLC
Yorktown, VA
------------------------------
Original Message:
Sent: 10-27-2022 07:58 AM
From: Christopher Evans
Subject: Existing Structures: Old drawings reviews and specifications (Tips)
James, multiple aspects; 1) our understanding of loads is better now, e.g. wind speeds / wind pressures, response of buildings to wind didn't really come into its fore until the 1960s, 2) understanding and modelling of seismic loads, 3) methods of construction using a more hybrid system esp. for lateral force resisting systems, connection types used (as alluded by others, bolts or rivets maybe have been used), 4) materials of construction, in that cast iron gave way to mild steel and the obviously the changes in steel strengths and change from permissible stress to ASD and/or LRFD. Older structures lateral force resisting systems were often not well understood and may have been desiged more for gravity rather than for sway. Often in the past shear walls (probably masonry) would have provided lateral strength & stiffness. Deflections limits (if considered) in the past would likely be more generous that we might assume today and obviously our understanding of 2nd order analysis etc wasn't developed until relatively recently. However, the over-design in old structures often took care of that rather by judgement and experience than calculation. A key note though is to always understand the load-paths, both vertical AND lateral (and indeed if torsional occurs) and model the fixity as well as one can in the analytical model.
------------------------------
Chris Evans
Chief Civil & Structural Engineer
Wood plc
Reading, UK
0118 913 3483
Original Message:
Sent: 10-10-2022 01:31 PM
From: James Williams
Subject: Existing Structures: Old drawings reviews and specifications (Tips)
Chris,
Absolutely and thank you.
When you say "our understanding has changed", do you feel they were more conservative then or simply that our methods of analyses are more sophisticated or allow us to delve into or consider more details?
When it comes to structural modeling software, there are the challenges of over confidence in the pre-, post, and analysis software. I direct individuals performing computer analyses to the user comments, feedback or user forums of the software provider and the fine-printed disclaimers. That is where you discover some of the analysis limitations not specifically spelled out in the promotions along with the message of user responsibility.
The non-engineering variables/unknowns or variability (methods, techniques & quality) associated with the final product still exist (i.e., construction, material supplies, actual usage). I have seen design values (aka factors of safety) change as we learn more.
I always view analyses as converging or getting closer to the actual number (sort of like 1/x gets closer to zero). I have run comparative analyses looking solely for acceptable values of the percentage of stress increases when faced with a lot of unknowns and uncertainty. The one tool we have in our design bag that is ours alone is the ability to increase our factors of safety in proportion with our amount of uncertainty.
------------------------------
James Williams P.E., M.ASCE
Principal/Owner
POA&M Structural Engineering, PLC
Yorktown, VA
Original Message:
Sent: 10-08-2022 06:55 AM
From: Christopher Evans
Subject: Existing Structures: Old drawings reviews and specifications (Tips)
I agree old codes are invaluable as are old engineering books, steel manuals etc. Be careful of ASD vs LRFD and always be wary of material strengths.The AISC for sure have some great reference data for old codes for steel & Ive built up a library of old ASCE loading codes. Be VERY careful and dont use default values in modern design codes or software........Be wary of load paths & lateral force resisting systems as old frames were sometimes only partial moment frames. Be wary of comnection types and assumptions re fixity. Ask around for old drawings as people who may have worked in the building or been involved in the design or construction. Sometimes a custodian might just have a full set of drawings tucked away. Always tread with caution and double check all your assumptions. Be a little bit detective a little bit forensic scientist a little bit magician making things that you cant prove to modern codes but are still standing.......also always be wary of wind loads. Our understanding has changed an awful lot in the last century!
------------------------------
Chris Evans
Chief Civil & Structural Engineer
Wood plc
Reading, UK
0118 913 3483
Original Message:
Sent: 10-06-2022 11:03 PM
From: James Williams
Subject: Existing Structures: Old drawings reviews and specifications (Tips)
What are your tips for assessing existing structures? Whether performing a field survey or structural analyses.
When having to assess any parts of a existing structure that could be defined as ancient, historical or simply built around the time I was born, one of the greatest assets you can have is the drawings/plans. This is rare but if you happen to be at a historical facility, drawings may be readily available. The challenge is that you may not recognize some of the material designations and/or shapes. If your firm has a subscription to a database that provides material specifications, you may access that old spec after a little bit of work. If it does not, you may consider purchasing the old ASTM Specification. I would purchase at my own expense to save time. If my memory serves me, a single spec may range from $25 to $75 dollars.
Here is my 1st tip:
Purchase an old code book for $1.85. [I went crazy and started buying as many old technical references I could find.]
Not only was I able to identify the steel, but it provided insight into calculations. For example, ultimate strength design U = 1.5D + 1.8L (15-1) from (ACI 318-63, Sec 1506)
Brings me to my 2nd tip:
When using structural modeling tools, analyze the existing structure first under what you assume were the original loads, load factors or load combinations. Did I make this mistake many moons ago? Absolutely-Once. I chased over-stressed structural members for structural modifications for days before checking. Later I discovered that under the original loading I assumed, the structure as modelled did not work. One either changes the design assumptions, design criteria, the model or one may choose to use non-linear analyses [Members may not reach that record breaking stress levels and continue to pass along loads]. However, you should start with a structural model whose members are good under the design loadings.
------------------------------
James Williams P.E., M.ASCE
Principal/Owner
POA&M Structural Engineering, PLC
Yorktown, VA
------------------------------