This page provides copies of presentations, other documents and software related some of our projects and areas of interest. Therse include a link to a summary of an ONR R&D report written by Jacobsen Analytics on multi-unit Level 3 PSA. You can also download a copy of a simple Python tool and example fire model developed to show how a relatively simple code can be used for fire modelling in a meaningful way.

Underpinning the UK Nuclear DB Criterion for Naturally Occurring External Hazards

The UK Nuclear Regulator (ONR) recently published a report on Underpinning the UK Nuclear Design Basis Criterion for Naturally Occurring External Hazards. The regulator's main objective was to look at the relation between risk, cliff edge effects and design basis criteria and where the bulk of the risk profile is expected to be compared to the DB hazard magnitude in each case. Naturally, PSA appeared to be the tool of choice as it gives valuable insights on the adequacy of designing to certain hazard magnitudes with certain margins or degrees of conservatism. Jacobsen Analytics contributed to the project by performing the PSA modelling and quantification (Task 2) which covers lightning strike, external flood and seismic hazards. For each hazard we created a model based on publicly available information (e.g. fragilities, hazard curves etc.) which was quantified using the damage state approach.

Research Report on Multi-Unit Level 3 PSA – Executive Summary

The ONR commissioned Jacobsen Analytics to investigate the effects of a multi-source release on Level 3 PSA consequence results to inform future updates of ONR guidance. The work was conducted using state-of-the-art Level 3 PSA software PACE and investigated how multi-source release consequences scale compared to the contributing single-source release consequences, whilst accounting for factors such as location and time lag between releases. The executive summary of the work is published here.

Design and Development of an Outage Planning and Risk Model

PDF version of a Powerpoint presentation on the design and development of an outage planning and risk model for Sizewell B. This Defence-In-Depth model is being developed in RiskWatcher and will track Technical Specification compliance using different coloured end states. This will be used by Sizewell B to aid outage planning and management to ensure a sufficient safety margin for all key safety functions throughout the outage.

Fire PRA Uncertainty

PDF version of a Powerpoint presentation on uncertainty analysis application within a Fire PRA

Database application for Fire PRA

PDF of a Powerpoint presentation on the development and use of a database application for Fire PRA

Python code to calculate probability for damage to fire targets on a control panel

quant_eqns.py is Python code that carries out a calculation of the overall probability that a fire scenario within a main control board results in a fire that is able to damage targets at a given distance and is not suppressed before damage occurs. It was Jacobsen Analytics contribution to an upcoming White Paper that updates the original implementation of the NUREG/CR-6850 Appendix L method. The software carries out a numerical integration over the range of heat release rates from the fire and then calculates an average over possible positions of the fire within a rectangle on the surface of the main control board. The code is released under the Apache 2.0 license, which is an open source license with few restrictions (the main restrictions relate to attribution - see license for full details).

Simple Fire Model Example with Python

Taking the fire modelling slightly outside the typical frame of "spreadsheet/Zone Model/ CFD", we decided to put together a very basic fire modelling example coded entirely in Python. The main purpose of this case study is to demonstrate that a very simple code can actually be applied as a very powerful fire modelling tool. It was created entirely in Google's Colaboratory platform which means you are free to get a copy of the code and run it yourself. Please read, test, forward to friends and let me know what you think.