Functional safety is making sure that a system which controls something dangerous, such as a car, a train, a machine or a medical device, cannot hurt people when it fails. The aim is not zero failures but a known, acceptably low risk of dangerous ones. Sector standards set the process, and the evidence must be written down.
Key concepts
- Hazard and risk analysis: find what could go wrong, how bad it would be and how likely it is.
- Safety requirements and integrity levels: the stricter the risk, the stricter the rigour demanded of design and testing.
- Safe state: what the system does when it detects a fault, such as stopping, holding or switching to a backup.
- Diagnostics and redundancy: self-tests, watchdogs, monitors and independent channels detect and survive faults.
- Safety case: a structured argument, backed by evidence, that the system is safe enough.
Vocabulary
- Hazard: a potential source of harm.
- SIL, ASIL: integrity levels from sector standards, rising with the severity of the risk.
- Systematic and random faults: design mistakes, which process fights, and hardware failures, which probability describes.
- Fail-safe: failing in a way that leaves the system safe.
What to study, in order
- Concepts of risk: severity, exposure and controllability.
- Do a hazard analysis for a simple device, such as a lift or a pump controller.
- Derive safety requirements, define the safe state and design the monitoring that reaches it.
- Read the general structure of the standard for your sector, and the process it asks for.
- Write a small safety case with claims, arguments and evidence, and have it reviewed by someone else.