Work from system boundary to evidence
| Step | Filling line example |
|---|---|
| Identify the system and boundaries | Include the controller, recipe HMI, service gateway and connections to the vendor portal and production planning. Identify the party responsible at each boundary. |
| Describe the unwanted state | An unreleased recipe becomes active. Product quality and batch release are affected. Assess separately whether this machine could also present a hazard to people. |
| Consider conditions and consequences | Who can change recipes? Which version does the shift see? How many batches could be involved before detection? What consequences are credible without the proposed controls? |
| Justify the objective | Changes need authorisation and a link to a work order. Staff must be able to compare the recipe with the approved version before release. The operator sets its risk acceptance criteria. |
| Agree measures and evidence | Assign diagnostic and change rights separately. Maintain a reference configuration. Demonstrate at acceptance that staff can attribute an authorised change and identify the approved state. |
| Assign ownership of residual risk | What if the comparison service is down? Who may hold a batch or revert to the reference state? Name an owner and a trigger for reassessment. |
This illustrates an application of IEC 62443-3-2. Risk grading and acceptable targets depend on the process, the consequences and the operator’s criteria. A number copied from a sample risk matrix carries little meaning without its assumptions.
The outcome: a defined system, credible scenarios, agreed requirements and evidence suited to each significant control. Where risk remains, record the operating decision and when it will be reconsidered.
Related reading
- Understanding IEC 62443: what needs protection on this particular line?
- Who can reach the machine during remote service?
- What should we agree before buying a connected machine?