Issue #4 - March 24, 2026

1. Code & Standards Watch

Stay informed of newly released codes and standards; keep current on updates and errata; and let your voice be heard through working sessions, public comments, and balloting on standards under development.

2. Research Snapshot

Interpretable machine learning for predicting the bearing capacity of double shear-bolted connections: a data-driven evaluation
If you saw the words “machine learning” in the title above and you instantly went on guard, I’m right there with you.  However, the first word of the title is really the focus of this paper, and it’s worth the read.  “Interpretable” models focus on transparency; they let us peel open the black box and see what’s inside.  

In this study, researchers took datasets from ten different studies of double-shear bolted connections in heavy steel plates, trained ten different machine learning models with them, and then validated, ranked, and most importantly, examined them critically.  Many of these models did better at predicting actual strengths of assemblies than our simplified code metrics (AISC and Eurocodes evaluated), which shouldn’t shock anyone.  Most of our code equations hark from the days of the nomogram and the slide rule, so they had to be simple, and they usually focus on just one or two factors at once.  Machine learning is effectively a big, complex, multi-factor weighted average, which can go non-linear if it includes a few steps.  This added complexity is better able to capture real-world nonlinearities, like the effects of edge distance, and the interactions between discrete variables.

Key Takeaways: The same variables we’re familiar with (edge distance, plate thickness, bolt diameter, etc) dominated the models; they just played together in more complex ways than we’re used to.  More importantly, this increased transparency in machine learning sets strong precedents for more work in these fields, advancing practice while still letting us see the mechanics behind it all and ask “why?” when things work differently from our existing equations.

3. Tools & Workflow

SAP2000 27.1.0 Released

4. Case Study of the Week

The Hard Rock Hotel Collapse - New Orleans
In October 2019, several stories of the Hard Rock Hotel collapsed during construction, killing three construction workers.  A few undersized members, some poorly designed connections, and construction loads that were not well-coordinated / well-considered in design all contributed, but the biggest danger was brought on by the progressive collapse from lack of redundancy in the system.  Caught on video, it’s clear to see several stories pancaked, with the failure rippling inward, and the facade peeled off to fall to the street below.

Key takeaways: Construction loads are a critical consideration, and redundancy is always important.

5. Upcoming Free Live PDH

6. Quick Hits

👋 From the Editor

I’m Eric, the engineer behind StructEd Brief. I dig through stacks of journal articles and software patch notes to find useful information for practicing engineers and keep an eye on the scattered code updates & errata for you. I’m just getting started, so if you find this useful, the best way to support the newsletter is to share it with a colleague or post it on LinkedIn. It helps more than you’d think!

📬 Subscribe & Share

If someone forwarded this to you, you can subscribe here:
👉 https://structedbulletin.com

Know someone who’d appreciate this?
Forward this email or share the link above.