Playbook
Shift review in a steel plant: replay the facts, rank the actions, record the decisions.
Most shift reviews run on the shift log and memory. The argument is about what happened, not what to do next.
The problem
- Each shop reports its own number; the converter met its target while the caster broke sequence.
- Holds, delays and energy overruns are reconstructed from logs and phone calls.
- Decisions taken in the shift are not recorded with the reason.
The playbook
- Replay the shift: heats, sequences, holds, delays and energy, with each event’s source and time.
- Rank by cost: look first at the items that cost the most or put orders at risk, not at whatever was loudest.
- Look across shops: check how each shop’s decision affected the next, such as converter timing against caster sequence.
- Check decisions: what was decided on each hold or alert, by whom and why.
- Assign actions: each action with an owner and a due date.
- Close the loop: start the next review with last review’s actions.
A 20-minute agenda
| Minutes | Item |
|---|---|
| 0–5 | Replay: throughput, sequence breaks, holds |
| 5–10 | Top items from the action queue by cost |
| 10–15 | Energy against norm by unit |
| 15–20 | Actions, owners, dates |
How to measure it
- Holds decided within the shift.
- Repeat causes week on week.
- Actions closed by their due date.
How Steel Twin does it
Steel Twin is a digital-twin control room for integrated steel plants. It joins PLC and SCADA, the historian, MES, LIMS, ERP, EAM and energy data on heat_id, so every heat is traced from charge to coil with its grade, sales order, route and chemistry. Every event keeps its source timestamp and a quality flag, so stale data is never hidden. Holds arrive with re-treat or regrade options priced against the order they affect and similar past heats; holds, sequence risks and energy variances against norm are pushed to the shift controller, and one tap logs the decision to the genealogy. A single action queue is ranked by cost and urgency, and any shift can be replayed with its sources. Delivery starts with steelmaking and casting on one shop, then adds energy, asset health, iron-making, mills and dispatch on the same record.
Questions
How should a steel plant run a shift review?
Replay the shift with sourced events, rank issues by cost, look at effects across shops, check recorded decisions and assign actions with owners and dates.
Why do shift reviews in steel plants go wrong?
Each shop reports its own number and events are reconstructed from memory, so the meeting argues about what happened instead of what to do.
See it on your own data. Steel Twin — Every heat, traced to the coil. Book a 30-minute working session with an engineer.