A functional design specification explains how an automation system will satisfy approved requirements without becoming an unreviewable mixture of wishes, code details and vendor manuals. It defines functions, states, interfaces, behavior, constraints and acceptance in a traceable structure.
The short answer
Start from stable requirement IDs and an agreed system boundary. Define operating modes, sequences, permissives, interlocks, alarms, controls, data, interfaces, user roles, time behavior, failure response, cybersecurity constraints and recovery. Separate required behavior from implementation detail unless the detail is itself constrained. Give every function measurable acceptance criteria and trace it upward to need and downward to design, software and tests. Review assumptions and open decisions before approving the baseline.
- Describe observable behavior and constraints, not unexplained implementation.
- Use stable identifiers for functions, interfaces and requirements.
- Define abnormal, degraded and recovery behavior alongside normal operation.
- Make every acceptance statement testable.
Decision guide
| Section | Question answered | Output |
|---|---|---|
| Scope and context | What is inside, outside and interfaced? | Boundary and responsibility map |
| Functional behavior | What must the system do in every relevant state? | Modes, sequences, logic and timing |
| Non-functional constraints | How well and under which limits? | Performance, security, availability and usability |
| Verification | How will compliance be shown? | Traceability and acceptance criteria |
A practical workflow
- 01
Approve requirement set, glossary, system boundary and document owner.
- 02
Structure functions, modes, interfaces and operating scenarios with stable IDs.
- 03
Specify normal, abnormal, failure, recovery and access behavior.
- 04
Review traceability, ambiguity, feasibility, risk and testability across disciplines.
- 05
Baseline the FDS and control every subsequent change and impact.
Evidence and acceptance
Retain document ID and revision, source requirements, glossary, scope and exclusions, assumptions, system and interface diagrams, functions and modes, sequence and state definitions, permissives/interlocks, alarm philosophy links, data and security rules, failure and recovery behavior, open items, review comments, approvals, trace matrix and acceptance criteria.
Common mistakes
- Copying a vendor manual instead of specifying project behavior.
- Mixing current and future scope without labels.
- Using words such as fast or user-friendly without measurable criteria.
- Specifying the happy path but not failure and recovery.
- Approving an FDS with unresolved ownership and interfaces.
Responsibility boundary
Primary sources
Technical claims and product boundaries were checked against these primary sources. Fact-checked
Frequently asked questions
Is an FDS the same as a control narrative?
No. A narrative explains process behavior; the FDS organizes the complete functional system definition and can reference narratives.
Should PLC code structure appear in the FDS?
Only where implementation constraints are required; otherwise specify observable behavior and leave detailed design to the SDS.
Who approves the FDS?
The project must name competent technical and stakeholder authorities for its scope.
Does AutoNexit generate an FDS?
No. It keeps the authored specification, reviews and evidence connected.