A steel package arrives at the fabricator with a general arrangement, a member schedule and a set of end reactions. Three weeks later there is an RFI asking whether the reactions are factored or service loads, and the steel has moved a month to the right. Nothing was designed wrong. The handover was incomplete, and the handover is a specified deliverable, not a courtesy.
First decide who is designing the connections
The Code of Standard Practice makes this an explicit choice that has to be made per connection, and published in the structural design documents. It is not a default and it is not an assumption.
| Option | Who designs the connection | What the design documents must then carry |
|---|---|---|
| 1 | The engineer. The complete connection design is shown in the structural design documents. | The connection itself, fully detailed. |
| 2 | An experienced steel detailer selects or completes the connection from reference information provided. | Design criteria: see below. |
| 3 | A licensed engineer working for the fabricator designs it, and submits substantiating connection information. | Design criteria, plus a review of representative samples before the main submittal. |
What the calculation actually has to cover
A connection calculation that checks the bolts and stops is not a connection calculation. Chapter J of the Specification sets out the limit states, and a complete note walks each one that applies.
- J2 Welds: J2.2 for fillet weld limitations, J2.4 for weld strength and the minimum fillet size table.
- J3 Bolts: J3.1 pretension requirements, J3.3 minimum spacing, J3.4 minimum edge distance, J3.6 tensile and shear strength, J3.7 combined tension and shear in bearing-type connections, J3.10 bearing and tearout at bolt holes.
- J4 Affected elements: J4.1 tension, J4.2 shear, J4.3 block shear. Block shear is the check most often skipped and the one that most often governs a coped beam.
- J10 Concentrated forces: J10.2 web local yielding and J10.3 web local crippling, wherever a beam flange lands on an unstiffened web.
Which code, on a Saudi project
SBC 306 is the Saudi steel code. Its own preface states that ANSI/AISC 360-10 is the base code used in its development, under an agreement with AISC, with modifications for local materials and regulatory requirements.
When the job runs on Eurocode
EN 1993-1-8 covers joints. Section 3 is bolted, riveted and pinned connections; Section 4 is welded connections; Section 6 is the component method for structural joints connecting H or I sections: the method that decomposes a joint into springs and assembles their resistance and stiffness. EN 1993-1-8:2024 has been published as the successor to the long-standing 2005 edition; national adoption of a new Eurocode part typically lags publication, so confirm which version the project's national annex actually calls up.
Finite element checks, and what they do not replace
Component-based finite element analysis (the method behind IDEA StatiCa) replaces the analytical component models of the classical component method with a finite element model of the plates, welds and bolts, while keeping code-based resistance checks on the individual components. It is genuinely useful on a joint the hand methods do not cover: a skewed brace, a multi-member gusset, a haunch with an irregular bolt group.
What it does not do is decide the load path. The software checks the joint you drew against the forces you entered, under the code you selected. A wrong assumption modelled precisely is still a wrong assumption, and it now carries a colour plot. The assumption register is therefore part of the deliverable, not an appendix to it.
The handover package
There is no single clause that lists the deliverable set: the Code of Standard Practice specifies the substantiating information and the approval process, and the rest is practice. This is the set we issue, and the set we ask for when we are the reviewer.
- A calculation note, per connection type, naming the clause beside every check.
- An assumptions register: load path, whether forces are factored or service, LRFD or ASD, weld access assumptions, erection sequence assumptions, and every value taken from the design documents rather than derived.
- Bolt and weld schedules that a shop can buy from, with grade, diameter, length, hole type and pretension requirement stated.
- Shop and erection drawings, cross-referenced to the calculations so a reviewer can get from a detail to its basis in one step.
- NC / DSTV files, matching the drawings exactly.
- An erection sequence reviewed with the steel contractor, because a connection that cannot be reached is a connection that cannot be made.
- A marked-up copy of the design documents showing which forces were used, and where they came from.
Questions this raises
Does the engineer's approval of shop drawings transfer responsibility for the connection?
No, and the Code of Standard Practice is explicit that approval does not relieve the fabricator of responsibility for the accuracy of the detailed dimensions or for the general fit-up of parts in the field. Approval confirms that the contract documents were correctly interpreted and that fabrication may begin. Under Option 3 the fabricator's licensed engineer additionally has to confirm in writing that the approval documents properly incorporate the connection designs.
The drawings give end reactions but do not say whether they are factored. What now?
Raise it as an RFI rather than assuming, and cite AISC 303-22 §3.2.3, which requires the design documents to state whether the data is given at service load or factored load, and whether the design is LRFD or ASD. Assuming factored loads when they were service loads under-designs the connection by the load factor; assuming the reverse wastes steel on every joint in the building.
Is a finite element check enough on its own?
It is enough for the resistance of the joint you modelled. It is not evidence that the load path you assumed is the load path the frame delivers, that the joint can be erected, or that the welds can be reached. Issue the model alongside the assumption register and the erection sequence, and a reviewer can check all three.
What changes when the frame is a seismic force resisting system?
The connection has to be qualified, not merely checked. AISC 341 sets the seismic provisions, and AISC 358 publishes prequalified moment connections that satisfy them without project-specific testing. Both were reissued in 2022. If a connection is outside the prequalification limits, qualification means testing, and that is a programme item, not a calculation.
