Industrial¶
Manufacturing¶
Rules found in this domain: P4, P39.
Illustrative, not evidence.
Scenario. On an assembly line, workers sometimes fit a bracket the wrong way round. Postmortems show the same error each month. The team redesigns the bracket so it fits only one way. A jig that refuses to close on a wrong part was the fallback, but it would only have made the error visible. Any worker who sees a defect can press a stop button on the line. The stop works in every mode, resetting it does not restart the line, and every stop is reviewed without blame.
- P39: A person makes the same error repeatedly, so the team makes the error impossible, with a bracket that fits one way. The jig, which only makes the error visible, is the next level down and is not needed.
- P4: Anyone who finds a problem that could cause harm can stop the line without permission or blame. The stop works in every operating mode and does not restart on reset. Each use is reviewed without blame.
Quality engineering¶
Rules found in this domain: P81, P27.
Illustrative, not evidence.
Scenario. A plant makes plastic housings, and the count of cracked parts rises. The team plans a change to the mould temperature, runs it on one shift, checks the crack rate, and acts by adopting or dropping the change. They also test the process with varied inputs, such as resin batches and room temperature, so that the result changes little. Open actions from earlier cycles are tracked until they are completed.
Occupational and industrial safety¶
Rules found in this domain: P38, P66.
Illustrative, not evidence.
Scenario. A worker can be hurt by a press that crushes parts. The team asks first whether the hazard can be removed, for example by changing the process so no hand enters the press. The process still needs hand loading, and a record states why. The team next adds a light curtain, an engineering control, and only then training and gloves. Before cleaning or repair, a technician isolates the press from power and locks the isolation with a personal lock, so nobody can restart it. The maintenance procedure starts with an isolation step and ends with a release step.
- P38: A risk is found, so the team uses the highest possible control level, starting with elimination, and acts first at the source. A record states why a higher level was not possible.
- P66: People work on the press, so it is isolated from all inputs and energy, and the isolation is locked so that nobody can restart it unexpectedly.
Fire protection¶
Rules found in this domain: P73.
Illustrative, not evidence.
Scenario. A warehouse stores packed goods that burn. The team estimates how fast a fire can grow to a size that blocks the exits. It then adds up the time for a detector to sense smoke, for the alarm to sound and for staff to react and get out. A drill measures the alarm-to-exit part, and detector tests give the detection time. The total is shorter than the time until the exits are blocked, with a margin.
- P73: Harm develops over time after a fault starts. The time to detect, alert, respond and recover is measured and is shorter than the time until harm, with a margin.
Industrial engineering and operations research¶
Rules found in this domain: P50, P44.
Illustrative, not evidence.
Scenario. A distribution centre wants to ship more orders a day. A model shows that packing stations are the constraint, and the team measures and names it before it buys anything. It adds packing capacity first. The packing queue is kept well below full utilization, and the new capacity is sized above the expected peak order load. A peak-day trial shows the stations keep up with the planned margin.
- P50: The team wants more throughput, so it finds and names the constraint, the packing stations, and improves that first.
- P44: The team designs capacity above the expected load and keeps the queue well below full utilization. A peak-day trial shows capacity at least the design load times the chosen factor.
Supply chain engineering¶
Rules found in this domain: P51.
Illustrative, not evidence.
Scenario. A retailer, a wholesaler and a factory each order from the stage below them. Each stage only sees the orders of its direct customer. Small changes in shop sales grow into large swings in factory orders. The firms agree to share the real shop sales data with every stage, so each stage can see the end demand.
- P51: Several stages pass demand signals along a chain, so the real end demand is shared with every stage instead of each stage reacting only to the stage below it.