J-Bolts in tension, revisited

[Once20036]

I`m looking into some modification to a small existing industrial facility that necessitates modifying the existing lateral system/braced frames.
These braced frames are along perimeter walls only, and attached to a pier via (4) 1" dia hooked anchor bolts with 8" embed to the top of the hook.
This is a low seismic, high wind location. Design loads on approximately 20k tension.

Popular wisdom says that these bolts are terrible in tension.
There's this thread,

http://www.eng-tips.com/viewthread.cfm?qid=320100,

and many others
There's AISC 13th p14-10, "hooked anchor rods should be used only for axially loaded members subject to compression..."
and then there's ACI appendix D (318-05, or 08, or 11)

Per ACI eq`n; I`m calculating a steel strength of 142k, concrete breakout of 29k (including provisions for 3 free edges) or 12.9k ignoring those provisions, and 34k for anchor pullout per eq`n D-16. Common wisdom says the pullout failure mechanism is crushing of the concrete at the inside corner of the radius, however, the ACI equation utilizes the entire horizontal leg to calculate capacity (limited to 4.5diam). Ultimately, the equations are saying that pullout is not an issue, and it's breakout that controls the design.

This is an existing facility so we're stuck with the cards we're dealt and trying not to call for work that isn't necessary, but I`m feeling a little conflicted by this contradictory information. Has anyone dug into this below and signed off on J-hooked bolts for tension loads?
Why is common knowledge not to use these when ACI equations clearly show significant capacity?

As always, Thanks!

 

[hawkaz]

Even in the 1988 UBC, these bolts only had 3200# allowable tension capacity (service loads) per bolt with a 12" bolt spacing and 6" edge distance. (Double if SSI was used)

I would find a way to supplement the connection

 

[JedClampett]

Why don't you test one or the whole assembly up to your design load plus some margin? If they have sufficient capacity, you're off the hook (hook; get it?). If they don't work, it's there for all to see.
I personally don't think they're Satan, but it's been pounded in my head so many times, I just repeat the information.

 

[Once20036]

I picked up a research paper, "Pull-Out Strength of L-Bolt Anchors"

https://www.concrete.org/publications/internationalconcreteabstractsportal/m/details/id/51686875

It was done in 2014 by a bunch of guys at WJE. It's an interesting read, and concludes, "The existing pullout failure load model for a hooked anchor is very conservative and can be increased by almost a factor of two."

Like Jed, I've heard numerous times that hooked anchors cannot develop any significant capacity and they should never be used, however, I can't find any technical basis to reject them in an existing installation. For new installs, I`ll continue using headed studs/bolts as they're much more efficient in resisting pullout.