PRAMEA Probabilistic risk assessment method development and - - PowerPoint PPT Presentation

PRAMEA – Probabilistic risk assessment method development and applications

Presentation at the SAFIR2018 research programme final seminar, Hanasaari 21.3.2019

25.3.2019 VTT – beyond the obvious

1

HRA and advanced control rooms

§ Accepted review manuscript about the effect of digitalization of NPP control rooms, in

RESS (Reliability Engineering and System Safety)

• Digital systems provide operator with more information but also with secondary tasks, compared

with analogue systems

• With digital UI, operator is not self-evidently aware of activities of other operators
• Digital systems are more often updated, causing pressure to plant design and training

§ User study among Finnish hybrid main control room operators about possible human

errors in an emergency; comparison of errors

• Operator interviews related to PRISE procedures in a hybrid control room (analogue for

controlling, digital for monitoring the system) about possible errors

• In practice, less possible errors in digital UI (user interface)
• In principle, in most cases the same error could take place in both types of UIs

§ The cause behind digitally originating/analogue originating error may be different

may affect error probability BUT the mere digitality/analogueness does not does not dictate the amount or level of human errors

25.3.2019 VTT – beyond the obvious

2

Site risk analysis

• Current PRAs are unit-specific
• Inter-unit dependencies taken into account

in a limited manner

• Operational experience world-wide shows that

multi-unit disturbances are not uncommon

• No international consensus yet with regard to
• Applicable risk criteria for a site
• Method

3

Site PRA developments and findings: NKS project SITRON, report NKS-419

• Site risk analysis has two purposes
• Complement single-unit PSAs
• Provide multi-unit risk metrics
• Safety goals are still connected to units, not at site level

4

• SITRON method provides a practical and cost-effective way of

performing a site risk analysis based on unit-specific PRAs

• Multi-unit initiating events can be limited to external hazards
• Loss-of-offsite power, Loss-of-ultimate heat sink

FinPSA level 2

§ Tool based on dynamic containment event trees

(CETs) developed further

§ Risk integrator combines results of different

CETs and calculates total results

§ Tight integration of PRA levels 1 and 2

implemented

• Contributions of basic events and initiating events in

level 2 results

• Level 1 results can be utilised in level 2 modelling

25.3.2019 VTT – beyond the obvious

5

Level 2 PRA modelling

§ A simplified BWR plant PRA model has

been developed

• Five containment event trees
• Levels 1 and 2 tightly integrated
• Model can be utilised in further research,

demonstrations, trainings and FinPSA testing

§ Uncertainty analysis

• Problems and good practises identified and

demonstrated

• In dynamic modelling, proper handling of

epistemic and aleatoric uncertainties is important

25.3.2019 VTT – beyond the obvious

6

Time-dependent analysis in FinPSA

§ Can be used for maintenance

planning, risk follow-up and analysis

• f seasonal variations

§ Time-dependent task file

• User specifies events to a time line
• e.g. maintenance activities, tests,

parameter changes, configuration changes

• FinPSA calculates time-dependent risk

curve, using existing PRA results

25.3.2019 VTT – beyond the obvious

7

Level 3 PRA

Improvements to an analysis method

• Developed earlier by VTT
• Systematic analysis of scenarios with event trees
• Improvements: no discretization of continuous

factors (e.g. wind speed) needed, source term uncertainty taken into account

Nordic guidance document on level 3 PRA

• Uses, benefits, regulations, analysis steps etc.

Seasonal and contextual factors in level 3

• Groundwork for systematic incorporation in

analyses

25.3.2019 VTT – beyond the obvious

8

0,00E+00 2,00E+02 4,00E+02 6,00E+02 8,00E+02 10 20 30 40

Population dose (M anSv) Rain intensity (mm/ hour)

1,00E-06 1,00E-05 1,00E-04 1,00E-03 1,00E-02 1,00E-01 1,00E+00 50 100 150 200

The probability to exceed the cancer deaths The number of cancer deaths

Portfolio optimization in PRAMEA

§ Pipe inspections (c.f. RI-ISI)

• Problem: Which pipes should be inspected based on incomplete

information about failure probabilities and failure impacts?

• Large-scale optimization of inspections of the sewerage network in

Espoo

§ Defence-in-depth

• Problem: Which combinations of safety measures are cost-efficient in

a system with event dependencies and multi-state failures?

• Optimization of safety measures for the airlock system of a CANDU

Nuclear Power Plant

25.3.2019 VTT – beyond the obvious

9

Portfolio optimization in PRAMEA

§ Dynamic systems

• Problem: Which combinations of safety measures are cost-efficient in

preventing time-dependent accident scenarios?

• Real case study on the vapour cloud ignition of an industrial system

§ Cybersecurity

• Problem: Which mitigation strategies minimize the risks of cyber

threats to cyber physical systems?

• Case study on the Advanced Metering Infrastructure (AMI) of an

electric power system

§ Four papers in leading international journals

(2 published in RESS, 1 revised for RESS, 1 submitted)

25.3.2019 VTT – beyond the obvious

10