Aurorisk FR

Chapter 1 · Lesson 1.7

From LOPA to SIL

LOPA indicates by how much the risk must still be reduced. When the other layers are not enough, this reduction is entrusted to a safety instrumented function, from which the target integrity level (SIL) is deduced.

Two ways to choose a SILRisk graph: Consequence, exposure, avoidance, demand; Qualitative reading; IEC 61508-5 — LOPA: Frequency with existing barriers; Compared with a target; Reduction factor: SILRisk graphConsequence, exposure,avoidance, demandQualitative readingIEC 61508-5LOPAFrequency with barriersexistantesCompared with a targetReduction factor: SIL
The risk graph is quick; LOPA is quantified.

The SIS and the SIF in brief

domino effect safety instrumented system (SIS) is an automatic system that reacts to a feared event to bring the installation back to a safe state: sensors, a logic solver and actuators (valves, pump trip). It implements one or more safety instrumented functions (SIF), each linking sensor, logic and actuator; it is separate from the control system. In LOPA, it is one of the protection layers, the last one added when the others are not enough.

The four SIL levels in low demand mode

The complete safety integrity level (SIL, from 1 to 4) sets the PFD rangeavg that a function called upon at most once a year must achieve. It reads like the confidence level of a barrier, with the risk reduction factor RR = 1 / PFD.

SILPFDavgRisk reduction factor
410−5 to < 10−410 000 à 100 000
310−4 to < 10−31 000 à 10 000
210−3 to < 10−2100 à 1 000
110−2 to < 10−110 à 100

Values from IEC 61508-1 (table 2), taken up by IEC 61511. For a function in high demand mode, the measure is the dangerous failure frequency per hour (PFH): see the course on SIS and SIL.

The LOPA approach

The LOPA method (lessons 1-2 to 1-6) calculates the frequency of the final event with the existing barriers, compares it with a risk target and deduces the reduction factor, hence the SIL, that the SIS must provide.

Rule of thumb

If the frequency of the scenario with the existing barriers is F0 and the target Fobj, the SIS must have PFDSIS ≤ Fobj / F0. The SIL is then chosen whose PFD rangeavg is below this value, with a margin.

Example recalculated. F0 = 10−4 per year (without SIS), target 10−6 per year: PFDSIS ≤ 10−6 / 10−4 = 10−2, hence SIL 2 (if a safety margin is sought). For a target of 10−8 : PFD ≤ 10−4, which corresponds to the SIL 3 / SIL 4 boundary, or to two independent SIL 2 functions.

The risk graph (decision tree)

The qualitative method follows a tree whose parameters are:

ParameterLevels
C : consequencesCA minor injuries; CB irreversible injuries to one or several people or death of one person; CC death of several people; CD catastrophic consequences, large number of deaths
F : frequency and exposure timeFA rare to frequent; FB frequent to continuous
P : possibility of avoidancePA possible under certain conditions; PB hardly likely
W : desired probability of occurrence of the consequencesW1 very low; W2 low; W3 relatively high

Reading the tree (C, then F, then P, then W) leads to a cell: no safety requirement, no special requirement, SIL 1 to 4, or “a single safety function is insufficient” (requirement beyond one SIS). The complete tree and its cells are to be read in IEC 61508-5 (annex on risk graphs); its exact transcription is not reproduced here.

To go further

The course on SIS and SIL covers what follows the choice of the SIL: low and high demand, SIL levels (lesson 1-3), allocation to layers and target SIL (lesson 2-2), then, in chapter 4, the architectures, the PFD calculationavg and the verification of a safety function.