Diagnostic Coverage
How exactly is DC used in SIL / PL calculations for machine safety?
DC feeds into PFHd calculation: undetected dangerous failure rate λdu = λd × (1 - DC). Only λdu contributes to long-term dangerous failure probability, while detected failures are assumed repaired within defined time. In ISO 13849-1, DC ranges help determine Category and PL, combined with MTTFd and common cause measures.
How does DC differ from Safe Failure Fraction (SFF) and why does it matter?
DC deals only with dangerous failures, while SFF includes all failures not dangerous (safe and dangerous detected). SFF ≥ DC by definition. Using SFF alone can hide limited DC; DC gives clearer picture of fault detection for risk-relevant failures.
How is DC determined in practice for a real subsystem?
Two methods: 1) Manufacturer-provided data (λd, λdd, λdu, DC percentage). 2) Engineering analysis using IEC 60812 FMEA: identify failure modes, classify as safe/dangerous and detected/undetected, sum failure rates, compute DC = λdd/λd × 100%.
Diagnostic coverage (DC) is a quantitative measure defined in IEC 61508 and ISO 13849 as the ratio of detected dangerous failures to total dangerous failures, expressed as a percentage. It evaluates how effectively a system's diagnostics detect failures that could cause loss of safety function, with typical ranges: low (<60%), medium (≥60% to <90%), high (≥90% to <99%), and very high (≥99%).
On the shop floor, diagnostic coverage is applied in safety-related control systems for machinery, such as emergency stops, safety interlock doors, light curtains, and safety PLCs. It quantifies how built-in checks (e.g., cross-fault monitoring, cyclic tests, plausibility checks) detect faults that could cause loss of safety function. For example, in a guard door with a safety relay, diagnostics monitor for short circuits, cross-faults, and discrepancy between redundant contacts. In a light curtain protecting a press, self-tests and output verification detect dangerous failures. DC values from manufacturer data or FMEA feed into SIL/PL calculations to verify risk reduction, as required by standards like IEC 62061 and ISO 13849-1.
Assuming high DC without evidence or manufacturer data, leading to overstated SIL/PL and underestimation of PFHd.
Diagnostics present but faults not annunciated or acted upon, so detected failures do not contribute to effective DC in practice.
Incomplete coverage of dangerous failure modes, such as missing systematic software errors or mechanical failures, resulting in lower DC than assumed.