Value Stream Mapping for Structural Steel Fabrication: From Shop Drawing Release to Site-Erected Steel Without the Weld Rework Loop

Why map the full steel value stream?

In structural steel fabrication, a member’s journey involves far more than cutting and welding. Information must move with the steel: approved drawings, current revisions, material identification, weld requirements, inspection results and erection sequence all affect whether the right assembly reaches site on time.

Value stream mapping (VSM) makes both flows visible. It captures the steps that transform material and the information and queues that determine when work can proceed. Mapping from approved shop-drawing release through site erection helps teams see where lead time accumulates, why weld repairs recur and how production priorities connect to the construction schedule.

This worked example follows a hypothetical fabricator producing welded plate girders for a construction project. The figures below illustrate how to build and interpret a map; they are not industry benchmarks. Replace them with observed data from your own shop.

1. Select a focused scope and collect real data

Choose a product family with sufficiently similar routing, such as plate girders or portal-frame columns. Avoid combining assemblies with very different sizes, weld requirements or routes: an average across unlike products can conceal the actual constraint.

For this example, the boundaries are:

  • Start: approved shop drawings issued as a controlled fabrication package.
  • Finish: steel transported to site, erected in the planned location and bolted.
  • Included steps: drawing-package release, material ordering and receipt, CNC profile cutting, drilling, fit-up and welding, NDT inspection, blast and paint, load-out, transport and site erection.

Before drawing, agree how each measure will be defined. Record work time, queue time, WIP tonnes, weld repairs and schedule performance using consistent units. Walk the process with detailers, planners, operators, quality staff, logistics and site representatives; compare the documented route with what work actually follows.

For drawing and approval control, agree on how revisions and review status will be identified and tracked. The AISC model review and approval guide offers relevant guidance for project teams coordinating structural-steel approval documents. A useful project-scoping starting point is this guide to Define-phase project scope.

2. Read the current-state map

The table represents one typical assembly moving through the selected stream. Cycle time is observed active work time per assembly; queue is elapsed waiting before the next operation. The work-time total includes an average allowance for repair effort across the family, rather than only a best-case, defect-free assembly.

Process step Active time per assembly Queue before next step
Issue and control approved detail package 18 min 1.5 working days
Order and receive material 30 min 5.0 days
CNC profile cutting 24 min 1.5 days
Drilling 21 min 0.5 day
Fit-up 66 min 2.0 days
Welding 120 min 2.5 days
NDT inspection and disposition 27 min 2.0 days
Blast and paint 54 min 3.0 days
Load-out 18 min 1.0 day
Transport 30 min 1.0 day
Site erection and bolting 12 min 1.6 days
Total 420 min (7.0 hours) 23.1 days

The queues total 23.1 working days. Adding seven hours of process time gives approximately 24 working days from release to erected steel. Assuming an eight-hour working day, total elapsed lead time is 11,520 minutes. Process cycle efficiency is:

PCE = value-added time ÷ total lead time × 100
PCE = 420 ÷ 11,520 × 100 = 3.65%

For this illustration, the seven hours are treated as direct transformation time. In a live study, define carefully whether inspection, transport or other necessary steps count as value-added; use the same rule before and after improvement. The Process Cycle Efficiency Calculator can help structure that calculation.

Compare the work pace with demand

Suppose the project requires 12 assemblies per week and the cell has 40 net production hours available. Takt time is:

2,400 available minutes ÷ 12 assemblies = 200 minutes per assembly

Takt is the demand pace, not the sum of all work content. The welding cycle of 120 minutes is below takt, but that alone does not prove the stream can reliably deliver 12 assemblies each week. Queues, repair work, equipment availability, product mix and shared resources affect actual output. Check capacity across the complete route and align fabrication release with the site’s required erection sequence.

Current state in a structural steel fabrication shop: identify queues and handoffs

3. Find the causes behind the weld-rework loop

The map should show repair as a feedback loop, not bury it within the welding box. A typical loop is: weld completed → inspection identifies a non-conformance → disposition and repair plan → defect removed and joint prepared → re-weld → repeat inspection → release. Each loop consumes capacity and may create extra handling, delay coating and disrupt the planned load-out.

In this hypothetical baseline, 12% of inspected welds require repair. Track the rate by welds inspected, and also record repair hours: a small number of complex repairs can consume more capacity than a larger number of brief corrections. Review defects by type, joint, detail revision, fit-up condition, WPS and repair location. The purpose is to find patterns and prevent recurrence, not assign blame.

Use DOWNTIME to examine waste across the whole stream:

  • Defects: dimensional non-conformance or weld repairs discovered after the assembly has moved downstream.
  • Overproduction: cutting or welding ahead of confirmed drawing release or erection need.
  • Waiting: work held for drawing clarification, inspection disposition or paint-booth availability.
  • Non-utilised talent: operators repeatedly correcting issues without a clear route to feed causes back to detailing and planning.
  • Transportation: unnecessary movement, including oversized members travelling between separated work areas.
  • Inventory: steel plate or partially completed assemblies held while approvals or downstream capacity are pending.
  • Motion: avoidable searching, walking or repositioning during fit-up, welding and inspection.
  • Extra processing: repeated checks, handling or cleaning caused by incomplete information or poor first-pass quality.

Some waste overlaps: approval delay can produce both waiting and excess inventory. That is useful to note because the same root cause can affect several performance measures.

4. Design a practical future-state map

The future state should connect controlled information release to a stable material flow. The goal is not to maximise utilisation at every station; it is to deliver complete, conforming assemblies in the sequence and timing the site needs.

First-time-right fit-up, weld sequencing and verification reduce repair loops

Build the improved flow around these practices:

  1. Create a ready-for-fabrication gate. Release a package only when the drawing revision and approval status are controlled, required clarifications are resolved, material is available and quality requirements are clear.
  2. Kit plate and parts by assembly. Verify identity and revision before cutting; provide a complete, sequence-based kit to the work area.
  3. Level-load fabrication cells. Group compatible work, balance skills and equipment, and limit WIP with visible FIFO lanes.
  4. Make fit-up first-time-right. Verify key dimensions and joint preparation before welding a batch. Standardise fit-up checks, weld sequencing and applicable procedure references.
  5. Pull work from the site erection sequence. Use erection zones and required dates to guide release, fabrication, coating and load-out priorities.
  6. Plan inspection and coating capacity. Coordinate NDT availability and paint-booth slots with the production schedule so assemblies do not accumulate in hidden queues.
  7. Standardise repair learning. Record defect causes, confirm repair and reinspection requirements, then feed recurring patterns back to detailing, fit-up and welding teams.

5. Set targets and sequence kaizen over 90 days

The future-state figures below are proposed targets for this hypothetical stream, not guaranteed outcomes. Validate them through trials and revise them against actual capacity and project constraints.

Measure Current state 90-day target
Lead time, drawing release to erected steel 24 working days 12 working days
Average process time per assembly 420 min 390 min
Process cycle efficiency 3.65% 6.77%
First-pass yield 84% 95%
Weld repair rate 12% 4%
WIP in selected stream 420 tonnes 220 tonnes
On-time delivery to site 78% 95%
Paint-booth utilisation 92%, with uneven queues 85%, level-loaded

A lower paint-booth utilisation target is intentional in this example: 92% with irregular arrivals may indicate an overloaded, inflexible schedule. A steadier 85% target creates planned capacity for mix, changeovers and urgent work. Measure utilisation alongside throughput, queues and on-time performance, not in isolation.

Wave Owner(s) Key actions and measurable targets
Days 1–30: Stabilise and measure VSM lead, detailing lead, quality manager Confirm baseline definitions; map actual queues; create a controlled drawing-release checklist; code weld repair causes. Target: 100% of selected packages have visible revision status and a recorded release date.
Days 31–60: Pilot flow changes Production supervisor, fit-up and welding leads, planner Pilot one family cell; kit material by assembly; introduce fit-up first-piece checks, weld sequencing standards and FIFO/WIP limits. Target: reduce repair rate from 12% to 8% and cut pilot WIP by 20%.
Days 61–90: Connect to site pull and sustain Operations manager, logistics lead, site coordinator Schedule load-out by erection sequence; reserve NDT and paint capacity; review performance weekly and correct drift. Target: reach 95% first-pass yield, 95% on-time delivery and the proposed 12-day lead-time target, or document the remaining constraint and next action.

Future-state pull flow: level-load, kit and deliver to the erection sequence

Turn the map into a capability

A current-state VSM is a measured picture of how work and information flow today. A future-state map is a testable design, supported by owners, capacity checks and follow-up measures. Review the key indicators weekly, verify that gains hold, and keep safety and applicable quality requirements central to every change.

Build the skills to lead evidence-based improvement: explore Lean 6 Sigma Hub’s CSSC-accredited, self-paced Lean Six Sigma certification courses, from White Belt through Master Black Belt, and apply the methods to your own value stream.

Kaizen. Kai-Care. Kai-Done. Lean Six Sigma

Related Posts