Consequence-Driven Design: Three Questions Before the First P&ID

Consequence-driven design is the second discipline of Cyber-Informed Engineering and the one that most changes a drawing set. Where critical function identification names what the plant must never lose, consequence-driven design decides how the plant is built so it cannot lose it, whatever the digital systems are doing. It reduces to three questions, asked in order, before the first P&ID is issued.

1. What must survive?

The critical functions, and for each one an engineered limit that holds without power, network or software. A fired heater keeps digital firing control, but low-flow and high-skin-temperature trips are hard-wired and relief is sized for the worst case, so a compromised DCS cannot overfire it. A generator step-up transformer is protected by relays acting on measured quantities; SCADA supervises and cannot inhibit a trip. An exothermic reactor has quench and emergency cooling on independent logic with a manual line-up. The pattern is the same each time: the limit lives in physics or in a dedicated protective layer, and the business network has no route to it.

2. What may fail safely?

This is the question teams skip, and skipping it is how optimization software ends up in the protective path. Advisory analytics, optimizers, remote access, reporting and historians may all stop. The design task is to make sure they stop to a known state, that nothing protective waits on them, and that operators can run the unit with them dark. Deciding this explicitly is what lets a plant adopt AI tooling without widening its consequence envelope: the tool can be wrong, offline or compromised, and the worst case is a lost recommendation.

3. What must recover first?

Recovery is designed, not improvised. The sequence is usually power, cooling, protection, control, supervision, with known-good configurations held offline and a restart path the operators have rehearsed. Degraded-mode operating procedures cover the state between fully up and fully down: control network segmented or lost, local control retained, protective functions intact. Written down, those procedures turn a cyber event or a network failure into an entry in the shift log instead of an incident report.

What changes on the drawings

Independent protection layers appear on the P&ID as hard-wired trips and mechanical relief rather than as DCS logic. Segmentation shows up on the network drawing as zones a compromised layer cannot cross into a protective one. The I/O list separates protective signals from supervisory ones. In NEXUS the same tag namespace carries from the P&ID to the integrity calculation to the operator screen, and in NORMEX each engineered limit is traced to the PSM element and RAGAGEP it satisfies. For advanced energy the method is identical: the HELIA balance-of-plant work applies these three questions to cryogenic, vacuum, tritium and power-conversion systems before the heat source is chosen.

Read the full sequence on the Cyber-Informed Engineering page, or start from your audience on the engineering teams and advanced energy pages.

Cyber-Informed Engineering is a framework published by Idaho National Laboratory for the U.S. Department of Energy. Porritt Inc. is an independent company; references describe the framework applied, not endorsement.

Scroll to Top