How to Avoid a Big Bang: Ventilation Concepts for Large Scale Hydrogen Compressor Enclosures.
Matthew Jenkins, Gexcon, UK

Hydrogen compression for large-scale electrolysis introduces competing design challenges between noise mitigation and explosion risk. This study investigates hydrogen dispersion within compressor buildings across a range of natural and mechanical ventilation strategies. CFD simulations are used to assesses how ventilation influences airflow, gas accumulation, and explosion hazard under realistic leak scenarios. A high-level comparison of the benefits and limitations of each concept is presented. The findings support risk-informed ventilation design and aims to help project teams, designers and vendors select appropriate concepts early by highlighting the factors that should be investigated, avoiding costly modifications at advanced stages of design.

IEA TCP Task 43 Subtask D – Hazardous Area Classification Example of Hydrogen Vents in an Electrolyser and Buffer Tank installation
David Torrado, Otto Simon, UK

Air Ingress: Recognition and Mitigation of Flammable Mixtures Forming in Hydrogen Installations.
Nick Ryan, Steer Energy Solution, UK

Air ingress is the exchange of fuel gas and air that occurs when gas systems are isolated and pressure decays towards ambient conditions. While not unique to hydrogen, its low molecular weight, high diffusivity, and wide flammability range mean that flammable mixtures can develop more rapidly than in natural gas systems. Experimental studies demonstrate the drivers for gas exchange, associated ignition consequences, and practical mitigation measures.

LTS Futures: Managing Safety in Repurposed Hydrogen Pipelines
Sophie Bladon, SGN, UK

LTS (Local Transmission System) Futures was a multi-year, innovative hydrogen project, led by SGN. The project objective was to demonstrate how the LTS can be repurposed to transport hydrogen. LTS Futures investigates several different aspects of repurposing including operational changes, material integrity and due to the novel nature of this project, process safety was critical.

A safety dossier was developed to enable the live demonstration, which comprised of different key elements providing evidence and demonstrating that it could be executed safely. The development of these key elements and lessons learnt will be presented.

Regulator Perspective: Why MAH-Barriers Fail in Reality
Fiona Johnson, HSE, UK

Credited, but unverified. Barriers credited in PHAs are frequently more than just the engineered function. Regulators often find that the operational and maintenance requirements that make barriers work on demand are treated as assumed, leaving avoidable uncertainty in barrier capability and weakened ALARP arguments. In barrier management terms, barriers are systems: technical and task elements working together. Crediting the engineering without specifying the task is crediting half the system. Attendees will be shown what regulators expect using familiar examples (SIL proof testing and alarm responses) and will leave with practical understanding of how to deliver systematic, analysed, barrier performance assurance.

Strengthening Barrier Management Through a Stronger Focus on Operational Barrier Elements
Joel Chacon, Eigen, UK

Barrier management in high-hazard industries has long focused on technical systems, yet major accidents repeatedly reveal that human performance is equally critical. This paper presents a structured approach to identifying and monitoring operational barrier elements — safety-critical tasks that directly prevent or limit major accidents — using performance shaping factors such as competence, procedures, and human-machine interface conditions as measurable proxies for task readiness. Drawing on Aker BP's offshore implementation, it describes how these elements are formalised and visualised through a digital Barrier Status Panel, enabling proactive, risk-informed management of both human and technical barriers within a single integrated framework.

From the Hammer to the Hazard: Demonstrating Effective Barrier Management
Martyn Akers, Ithaca Energy, UK

From the Hammer to the Hazard presents a practical approach to strengthening confidence in barrier management by explicitly linking routine work to the hazards it can influence. While MAH frameworks and Safety Cases are well established, incidents continue to show weaknesses at the point of work. Using an example of a worker picking up a hammer, this presentation explores how competence, procedures, work planning, and organisational roles support critical barriers. A method is described for making these connections visible and testable, improving leadership assurance and allowing a demonstration that barriers are being managed as intended during everyday operations.

Why Barriers Keep Failing: Insights from a Decade of Petroleum Incident Investigations in Nigeria
Bello Sayyadi, NMDPRA, Nigeria

Why do known process safety failures recur despite established controls and corrective actions? Using the investigated-incident dataset of Nigeria's midstream and downstream petroleum regulator (397 incidents, 874 fatalities, 2015–2025), this paper applies the Energy Institute barrier-failure framework to a sector-wide population. Findings include that 9.6% of incidents produce 79% of fatalities, that vehicle transport failures drive 57% of all deaths, and that ignition-control and transport barriers recur within 15–17 days median, suggesting corrective actions do not durably restore barrier integrity. Implications for inherent safety prioritisation and corrective-action effectiveness are drawn for international practitioners.

An AI Framework for Structuring Risk-Informed Decision-Making in Process Safety
Masayuki Tanabe, Strategic PSM Initiative Group, Japan

Process safety practices are often reduced to checklist-driven risk assessments, which can overlook critical high-consequence scenarios and fail to support decision-making. This paper proposes a decision-centric framework in which risk is treated as a structured causal system linking design, failure mechanisms, protection layers, and management actions. An AI system (SPSM-AI) is introduced, embedding a process-safety-specific thinking landscape via system prompts to enforce consistent reasoning. Comparative analysis shows that this approach improves causal coherence and decision support across engineering design and PSM implementation. The study highlights the role of AI as a cognitive partner for strengthening structured thinking in process safety.

AI in Safety-Related Industries: The Risk Assessment Gaps Conventional Frameworks Were Not Designed to Close
Colin Chambers, HSE, UK and Bjarne Andre Asheim, Kairos Technology AS, Norway 

Conventional risk assessment frameworks were not built to address the novel challenges AI introduces in safety-related industries. This paper takes a deductive approach, drawing on relevant standards, the UK Government AI Playbook and published literature, to identify seven assessment gaps: bias in model performance; opacity in decision logic; inadequate testing; overreliance on automation; ethical concerns including surveillance creep; insufficient robustness against adversarial inputs; and weak governance of AI system changes. Three gaps, opacity, ethical risks, and adversarial robustness, have no equivalent in conventional frameworks such as HSE's RR151, representing genuinely new blind spots for practitioners.

AI in HAZOP and LOPA: A Powerful Assistant, But Not the Chairman — A Realistic, Experience-Based Perspective
Hamid Kareshki, Fluor, UK

This paper presents a practical, experience-based view of how AI can and cannot be used in HAZOP and LOPA. It challenges the idea that AI could chair or replace multidisciplinary studies, arguing that its real value lies in preparation, reporting, and review—not decision-making. The paper highlights both efficiency gains and hidden risks, particularly AI’s tendency to produce convincing but incorrect outputs. It provides a clear framework for safe adoption, emphasising that while AI is a powerful assistant, ownership of risk and final judgement must remain with competent engineers and operating teams.

Deployment of an AI-Enabled Process Engineering Assistant for Offshore Process Safety and Technical Decision Support
Deema Alluwaimi, ADNOC, UAE

This paper presents an AI-enabled Process Engineering Assistant designed to improve process safety decision-making in offshore operations. By integrating structured engineering knowledge into a controlled AI framework, the solution enhances access to critical information, reduces response time, and improves consistency in engineering decisions. The system supports troubleshooting, MOC reviews, and process evaluations while maintaining a human-in-the-loop approach. The work provides practical insights into the safe adoption of AI in process industries and its role in reducing human error, improving situational awareness, and strengthening process safety performance.

When Everything is Critical, Nothing is Critical: Rethinking SCE Identification for Better Safety Outcomes
Soraya Tamara, Petronas, Malaysia

This paper challenges SCE overclassification, where “everything is critical” undermines effective barrier management. A risk-based, enterprise approach shaped through cross-business review, industry benchmarking, and expert input which redefines SCE identification using three guiding principles. Piloted in two LNG facilities, it achieved a 65–85% reduction in SCEs without compromising safety. The result is a leaner, more focused system that reduces assurance deferrals, optimizes resources, and strengthens prevention of Major Accident Events (MAEs) and Loss of Primary Containment (LOPC).

From Safety Case to Availability Case: Using Process Safety Elements to Improve System Availability
Ian Sutton, Sutton Technical Books, USA

The process safety management discipline is mature. The focus has traditionally been on safety and environmental performance. This paper shows how the discipline can be repurposed to improve system availability and profitability within safety boundaries: “Safety defines the boundaries. Availability defines the outcome.” Availability is treated as an emergent outcome driven by downtime avoidance, uptime sustainment, and restoration speed/quality. The paper maps process safety elements such as hazard studies, mechanical integrity, management of change, procedures/training, incident investigation, and metrics into an availability assurance model. It also discusses how AI decision support can help or hinder system availability.

Execution Can Be Outsourced. Accountability Cannot: Governing the Grey Zone in Process Safety Outsourcing
Mohammed Jimoh, Dr Jimoh Consulting, Germany

As companies increasingly outsource HAZOPs, SIL studies, relief work, audits, and other process safety activities, a critical governance risk can arise: technical execution is transferred, but accountability is not. This presentation examines the “grey zone” between third-party delivery and the operator’s retained responsibility for risk ownership and acceptance. It identifies common failure mechanisms at the client–provider interface and proposes a practical governance framework covering roles, decision rights, participation, verification, traceability, and final technical sign-off. The aim is to help organisations outsource process safety work more effectively without creating false assurance or weakening accountability.

Management of Technical Change: New Energy Institute Guidance on a Structured Framework for Avoiding Common Failure Modes
Peter Webb, Beyond Risk, UK

Management of Technical Change (MoTC) is a recurring causal factor in major accidents, yet failures in its application remain commonplace. This paper presents a practical guidance framework for MoTC in major accident hazard facilities, covering the management system arrangements, process structure, and critically the ways MoTC systems fail in practice. Eight failure modes are examined, including failure to recognise relevant changes, creeping change, inadequate risk assessment, ineffective stage gates, and the absence of MoTC when equipment is retired. Each is illustrated with case studies and supported by practical preventive measures.

Quantitative Risk Assessment of Ammonia-fueled Ship Technologies and the Cost-benefit Analysis of Related Risk Reduction Measures
Rustam Abubakirov, Delft University of Technology, Netherlands

This work provides an overview of safety aspects of ammonia-fueled shipping based on the results of quantitative risk analysis. The work identifies population groups, both onboard and ashore, exposed to elevated risks associated with the introduction of ammonia as fuel. The effectiveness of applicable toxicity risk mitigation measures is then analyzed. The results are supported by cost-benefit analysis. The work is closely aligned with the current regulatory context, and the remaining major uncertainties are discussed.

Effect of Ammonia Cracking Ratio on Hazardous Distance and Dispersion Behavior of Ammonia-Hydrogen Gas Turbine Fuels Using IEC 60079-10-1 and PHAST
Jaeha Kim, Incheon National University, Korea

This study evaluates the hazardous distance, dispersion behavior, and toxic impact range of partially cracked ammonia fuel for gas turbine applications using IEC 60079-10-1 and PHAST. The release scenario and analysis scope are defined with reference to gas turbine-specific hazards in ISO 21789, including fuel systems, enclosures, ventilation, and ignition source control. The ammonia cracking ratio is treated as a key process safety variable because it changes the fractions of residual ammonia, hydrogen, and nitrogen. By comparing IEC-based hazardous distances with PHAST-based dispersion footprints and side-view profiles, this study supports hazardous area classification and safety design.

Propagation Containment in Integrated Solar-BESS-Electrolyser Architectures: Translating McMicken and Moss Landing Learnings to India's Emerging Green Hydrogen
Hidhin Antony, VyroNyx, India

Standalone-BESS propagation containment guidance (NFPA 855, UL 9540A) was not designed for India's emerging integrated solar-BESS-electrolyser hubs — Adani Kutch, NTPC Pudimadaka, SECI Mode-2A ammonia projects — where compact tropical-climate layouts place multiple major hazards in proximity. This work re-extracts McMicken (2019) and Moss Landing (2025) lessons through the integrated-hub lens, applies DNV PHAST/SAFETI consequence modelling to combined-hazard scenarios, and offers a design and ERDMP strengthening checklist before the 2027–2030 commissioning window.

Functional Safety of Battery Management Systems in BESS – The Role of IEC 61508
Simon Burwood, ERM, UK

As Battery Energy Storage Systems (BESS) become integral to grid stability and renewable energy integration, the safety and reliability of their control systems, particularly the Battery Management System (BMS), are paramount. This paper evaluates whether IEC 61508, the international standard for functional safety of electrical/electronic/programmable electronic systems, should be applied to the design and management of BMS in BESS installations.

Safer Chemistry Is Not a Safety Case: A Consequence-Based Framework for Thermal Runaway, Dispersion and Compliance for Battery Energy Storage Systems
James Close, AWN Consulting, UK

Grid-scale battery energy storage systems are being installed rapidly, and recent fires at Moss Landing, Thurrock, and Warwick show that current safety practices fall short. The move to LFP chemistry has improved thermal stability, and updated NFPA 855:2026 and UL 9540A testing characterises thermal runaway behaviour under controlled conditions. But neither captures what happens off-site when a unit fails, and the published science shows LFP off-gas is more toxic than often assumed. The UK framework leaves significant gaps, with COMAH not applying and no clear guidance on consequence assessment. It sets out a framework for consequence-based BESS safety cases.

A Methodology to Assess Battery Energy Storage Systems (BESS) Toxic Hazards and Inform their Location
Murray Urquhart, Cairn Risk, UK

In the UK, lithium-ion Battery Energy Storage Systems (BESS) are being proposed in the built environment near vulnerable receptors. Current statutory consultees may have insufficient process safety knowledge to adequately assess site locations and configurations. The toxic hazard extent tends to be determining for BESS facilities. However, whilst relevant good practice provides clear guidance on managing battery fires and explosions, toxic hazard assessment guidance is lacking. Cairn Risk has developed a methodology to assess BESS toxic hazards, based on existing criteria and guidance, which can guide developers and statutory consultees in evaluating the location and configuration of BESS facilities.

How to Train Your Dragon - Establishing a Process Safety Competency Development & Learning Program within a National Oil Company (NOC)
Michael Godfrey, That Process Safety Guy, UK

This paper details a Middle East National Oil Company’s (NOC) successful transition to a standalone Process Safety discipline. To rapidly upskill a diverse team, ranging from fresh graduates to experienced engineers the organisation implemented an accelerated competency development and learning program. This followed the 70/20/10 learning model but also utilised the "why" and "how" logic to engage the next generation of engineers, prioritising ‘learning’ over ‘training’. We will present a stepwise methodology covering technical foundations, the three pillars of integrity, and PDCA-driven assessments. Attendees will gain insight into a proven approach to build robust Process Safety cultures within operating companies.

Detecting Latent Organizational Deviations through Safety Culture Assessment: Insights from Large-Scale Industrial Data from Japanese Companies
Akira Tose, Niigata University, Japan

This paper presents a large-scale analysis of safety culture assessment data to identify potential early signs of organisational deviation. Using over 200,000 responses from more than 400 sites, a reference profile of typical organisational patterns is derived, enabling identification of deviations at individual sites. These deviations—such as differences across functions and hierarchical levels, and gaps between systems and practice—provide insight into organisational conditions that may influence barrier effectiveness. The approach complements conventional safety monitoring by offering a perception-based perspective and supports more proactive identification of potential risks in high-hazard industries.

What Good Looks Like: Process Safety Management Insights and Lessons from ISRS audits of 25 Major Accident Hazards sites
Muhammad Ullah, DNV, UK

Process safety performance varies significantly across major accident hazard sites, even between apparently similar organisations. This paper analyses DNV ISRS audit data — scores and auditor comments across 15 management processes and 123 sub-processes — from 25 MAH sites across Europe, the UK and the US. It identifies the structural relationships between them: where leadership commitment translates into effective risk controls, where it does not, and why. The findings show that high-performing sites are distinguished by coherence across connected management processes rather than excellence in any single one. Delegates will leave with a practical diagnostic they can apply to their own organisations.

From 'Flat World' to Real World: Incorporating Terrain into Buncefield Scenario Modelling
Luke Butcher, ESR Technology, UK

This paper explores an enhancement to the widely used FABIG TN12 approach, using readily available LIDAR (LIght Detection And Ranging) terrain data, to rapidly evaluate the effects of real world terrain and obstacles within a 'Buncefield Scenario' risk assessment. This new approach makes it possible to visualise and quantify vapour cloud behaviour in a way that reflects the real-world complexity of tank storage sites. This can drastically improve upon the understanding of cloud spread and provide meaningful insights to a site’s risk profile not afforded by the current ‘flat world’ assumption.

Confidence in the PDR-CFD Approach for Simulating Vapour Cloud Explosions
Steve Howell, Abercus, UK

The first PDR-CFD code for vapour cloud explosions appeared in 1986. Since then, the approach has evolved from a niche offshore tool into a widely used industry method. However, PDR-CFD remains a coarse, pragmatic engineering approximation, and recent overconfidence in its predictive capability is concerning. The PDR Alliance has been formed to address this through VVUQ-focused work. This paper reviews four Model Evaluation Protocols, finding only one rigorously addresses verification. It presents a joint industry activity examining the scope, ownership, and responsibility for VVUQ, aiming to strengthen confidence in explosion modelling and stimulate discussion within the hazards and safety community.

From Intolerable to Insightful: Lessons from AZ for Improving Semi-quantified Risk Assessment
Julie Bell, Human Factors Expertise, UK

This session will share how AstraZeneca’s (AZ) semi-quantified risk assessment (sQRA) initially indicated intolerable risk levels inconsistent with site experience, implying disproportionate additional controls. A detailed review examined the treatment of human reliability, particularly errors of commission within Process Risk Assessments (PRAs) underpinning batch sheet development. Critical task analyses were undertaken and HEPs re-derived using HEART+ to better reflect task complexity, governance, and competence. The revised model provided a defensible, proportionate representation of risk and highlights key lessons in applying human reliability methods within semi-quantitative MAH assessments.

CCPS-EI New Guidelines - Making Procedures Effective
Andrew Brazier, AB Risk, UK

CCPS-EI book “Making Procedures Effective - Guidelines for Developing and Using Procedures for Process Safety and Operational Excellence,” will be published towards the end of 2026. This marks 30 years since CCPS published its original guidelines for writing effective operating and maintenance procedures.

Does Issuing a Permit Make a Job Safer?  Lessons from a Comprehensive Control of Work Audit at a Major Hazard Facility.
Ashley Hynds, DNV, UK

DNV recently audited the control of work system at a major hazard operator. Despite the company having a mature safety framework, several related incidents had occurred. Using desktop review, interviews, and installation visits, over 100 safety observations were identified. These were grouped into themes aligned with HSE guidance, and a programme was initiated to improve task risk assessments, workpack usability, electronic permit system functionality, high‑risk task controls, and workforce competence. This presentation shares the audit approach, key findings, lessons on effective independent assurance, and insights from implementing a sustainable control of work improvement programme within a major hazard business.

A Practical Interpretation and Application of EI15 to Category III Fluids (Diesel) Systems
Keith Johnson, Sellafield UK

Recent revisions to EI15, incorporating CLP 2015 and HSE Joint Industry Project findings, have led to inconsistent interpretation of diesel area classification requirements under DSEAR, often resulting in conservative zoning. This paper presents a structured, practical methodology for assessing diesel systems, integrating multiple standards into a coherent approach. The method uses logic charts to support consistent decision-making and introduces risk-based threshold criteria to avoid unnecessary classification where no additional risk reduction is achieved. Application across operational facilities demonstrates improved consistency, reduced assessment effort, and increased confidence in design justification.

Controlled Experiment – Major Operator compares Traditional and AI supported HAZOP
Angus Keddie, Process Safety Matters, UK

This paper explores a controlled experiment which compares the preparation, execution and outcomes of a high hazard scope subject to a traditional and AI supported HAZOP study. The rapid rise of AI — particularly large language models such as OpenAI’s ChatGPT — has created new opportunities for automation and efficiency in Process Safety. To move beyond theory, however, a major high-hazard operating company conducted a controlled comparison between a conventional HAZOP and an AI-assisted workflow involving automated modelling and a mini-HAZOP. The paper compares both approaches in terms of cost, time, quality, and residual risk perception.

From Diagram to Deviation: Incident-Backed Recommendations for Auditable HAZOP
Thirakorn Mokkawes, Agency for Science, Technology and Research, Singapore

This contribution presents an auditable workflow for bringing historical incident learning into HAZOP preparation. Process-node context is matched against available reviewed eMARS and IChemE-derived incident datasets to surface relevant causes, consequences, safeguards and lessons learned. The method is not autonomous HAZOP generation; it keeps engineering judgement explicit while improving traceability between incident evidence and recorded recommendations. The paper shows how AI-assisted retrieval can support, rather than replace, human process-safety review.

Surfacing What Gets Missed: An Auditable AI Approach to Identifying, Tracking, and Tracing Safety-Critical Actions in Maintenance Campaigns
Mohammed Alsaabri, ADNOC, UAE

Maintenance campaigns are among the riskiest periods in process operations, where safety-critical actions are scattered across emails and minutes — easily missed, and sometimes deferred to a future campaign and forgotten. Unclosed actions are a known cause of major accidents. This paper presents an auditable, AI-enabled approach that identifies safety-critical actions, tracks them to close-out, and keeps each traceable to its source. AI also surfaces easily-lost items, such as Management of Change conditions or deferred scope, flagging what is needed before or after a decision. The result is earlier risk visibility and easier assurance, with human judgment retained throughout.

Artificial Intelligence in the Control Room: Early Upset Detection and Supporting Operator Decision-Making in Major Hazard Facilities
Bjarne Andre Asheim, Kairos Technology AS, Norway

Control room operators at major hazard facilities face growing process complexity, alarm overload, and sustained pressure on staffing. This paper describes the deployment of an AI-assisted support system using Multilevel Flow Modelling to provide real-time causal diagnosis of process upsets. At an offshore installation, the system detected a developing disturbance and identified its root cause approximately 50 minutes before conventional alarm thresholds were breached. The paper presents this case study in detail and addresses the human factors dimensions critical to deployment: building operator trust, presenting AI reasoning without inducing over-reliance, and structuring audit trails for regulatory scrutiny.

Fire Attack and BLEVE of Liquid Hydrogen Tanks: Preliminary Experimental Observations of Thermal Response, Blast, Fireball and Fragmentation
Steven Betteridge, Shell, UK

While liquid hydrogen (LH₂) handling is established in some contexts, its use in public applications such as refuelling and bunkering is less understood. Double-walled tanks with evacuated, insulated annuli minimise boil-off, but materials like mylar may fail under fire, especially if vacuum integrity is lost. The HySOON project investigated fire exposure of six 680‑litre LH₂ tanks with different insulation (mylar, MLI, perlite) under propane jet fires (100–150 kW/m²). Tests examined thermal response, pressure behaviour, and BLEVE risk when relief valves were blocked, The results will quantify the potential for these vessels to fail under fire fire attack

Does a Lower Flammable Limit of 4% Fairly Represent the Hazard Posed by Low Hydrogen Concentrations? An Empirical Review of Hazard Equivalence for Hydrogen and Methane Gas Close to the Lower Flammability Limit (LFL).
Nick Ryan, Steer Energy Solutions, UK

Hydrogen is increasingly being considered as a low-carbon fuel, but concerns over its wide flammability range, low ignition energy and fast flame speeds have led to conservative hazard assessments. Experimental studies by Steer Energy demonstrate that ignitability does not necessarily equate to hazard severity. Near the lower flammability limit, hydrogen ignitions often result in limited flame propagation, low overpressures and incomplete consumption of the gas inventory, whereas comparable methane mixtures burn more completely and can generate greater consequences. The work highlights the need for realistic, evidence-based approaches to hydrogen hazard assessment and classification.

Experimental Characterisation of Gaseous Jet Fires Impinging on a Pipe
Giordano E Scarponi, University of Bologna, Italy

Hydrogen jet fires represent a key escalation hazard for hydrogen production, storage and distribution facilities. This paper presents large-scale experimental measurements of thermal loads generated by sonic hydrogen jet fires impinging on a steel pipe representative of process equipment. The results demonstrate highly non-uniform heating and peak heat fluxes significantly exceeding typical hydrocarbon jet-fire design values. The findings provide a valuable experimental basis for consequence modelling, CFD validation and equipment vulnerability assessment, while offering practical insights into passive fire protection requirements and fire safety design for emerging hydrogen infrastructure.

Comparison of Hydrogen and Propane Impinging Jet Fires modelling to Experiments Using Different Computational Fluid Dynamics Techniques
Ben Vickers, Thornton Tomesetti, UK

This paper reports on a comparison of two jet fire releases impacting on a target, assessed using the two different CFD modelling techniques, for both hydrogen and propane fuels, to provide an indication of the heat load on a target. The results were further compared to experimental data for the same releases allowing a direct comparison between hydrogen and hydrocarbon jet fires to the measured heat load on a target.

Quantifying Risk and Barrier Independence for AI Systems in Process Safety: Bow-Tie Analysis, LOPA, and the Limits of AI Defence-in-Depth
Ivo Dujmovic, uc2c.ai, USA

When AI systems are deployed in safety-critical PSM workflows, process safety engineers face the question they always face: how bad, and how protected? This paper translates AI-specific risk into the bow-tie and LOPA language used in COMAH safety cases. We propose three anchor analogies — capability evaluations as design basis, adversarial red-teaming as mechanical-integrity analogue, and AI safeguard stacks as suspect IPLs — and present a worked bow-tie of an AI-native PSM platform. The headline finding: defence in depth against AI hazards is real but smaller than naïve barrier counting suggests. Practical templates and a modified LOPA approach are provided.

Assigning Hazard Severity During Hazard Identification Workshops
Jay Grainge, AXIOM, UK

In a hazard study workshop, assigning a consequence severity can be uncertain and time consuming within the boundaries of the meeting room. This paper aims to provide support to hazard study leaders to improve accuracy, efficiency and consistency of hazard severity during hazard identification workshops.

Quality Risk Assessment: A Powerful Tool for Risk-Based Decision Making
Roman Bulgachev, BP, UK

Risk assessment is a powerful technique which allows engineering and operational teams to identify and assess the risks associated with their operations and make informed risk-based decisions on whether the work is safe to execute. This presentation will discuss the initiatives undertaken by the bp wells organization to enhance the quality of risk assessments and risk-based decision-making. This presentation will provide an overview of the most important good practices adopted throughout implementation of the initiatives and describe the key attributes of a good quality risk assessment.

Integrating HAZCON and SIMOPS to Strengthen Risk Management at the Construction–Operations Interface
Mike McKay, TUV Rheinland, UK

Major hazard facilities often conduct construction alongside live operations, but HAZCON and SIMOPS are typically applied separately, creating risk gaps. HAZCON underemphasises operational impacts, while SIMOPS lacks depth in construction hazards, leading to fragmented risk understanding. The paper proposes an integrated HAZCON–SIMOPS framework that captures interactions between construction and operations, improving identification of escalation pathways and compounded hazards.

Bridging the Gap Between Safety Procedures and Operational Reality: A Field-Based Case Study from Oil & Gas Operations
Mohammed Tamim, Rumaila Operating Organisation, Iraq

This paper examines the gap between safety procedures and actual work practices in oil and gas operations. While procedures are designed to control risk, field observations show frequent deviations driven by operational pressures, procedural complexity, and human factors. The study highlights how these gaps can undermine process safety and contribute to major hazard risks. It provides practical insights into aligning procedures with real-world conditions through simplification, improved supervision, and workforce engagement. The findings support a proactive approach to identifying early warning signs and strengthening the effectiveness of safety management systems.

A Method to Assess Subjective Operator Mental Workload Under Different Operational Scenarios
Plato Chan, Keil Centre, UK

This paper introduces a practical method to understand how operator mental workload may vary under different operational scenarios. The method is based on the subjective rating approach in assessing mental workload, and involves combining the data from two of the commonly used workload assessment method: NASA-TLX (Task Load Index) and Multiple Resource Theory. This provides organisations with a method to evaluate the workload impact under different operational scenarios. Case studies will be used to illustrate how the workload data can be collected and how the results can be used to manage tasks and resources to optimise operator performance.

Human Factors Design Principles for Constructability
Claire Pollard, AWE Nuclear Security Technologies, UK

UK nuclear licensed sites have construction programmes for at least the next decade. Designing for constructability is essential to ensure programmes are achievable and safe. Early assurance of constructability minimises errors, delays and cost over runs. The current approach to Human Factors (HF) integrates HF activities through project design development stages, commissioning and into operations, ensuring human capabilities, limitations and behaviours are considered to support process safety, operability and maintainability. This paper provides additional guidance on HF aspects to develop skills and support decision making to deliver effective and reliable construction activities.

Human and Organisational Drivers of Major Accident Risk in the European Oil and Gas Industry: Practical Lessons for Process Safety and Implications for Hydrogen
Bassey Bassey, Coventry University, UK

This paper presents a data‑driven analysis of human and organisational factors underlying major accident risk in the European oil and gas sector. Using the HFACS framework and statistical modelling, it identifies strong relationships between decision errors, supervisory failures, and organisational weaknesses. The findings highlight practical intervention points, including improved supervision, contractor management, and control of work systems. The study is directly relevant to process safety practitioners and extends to hydrogen systems, demonstrating how lessons from hydrocarbon operations can support safer deployment of emerging energy technologies and prevent the transfer of systemic failure pathways.

Reactive Chemistry PHA’s: Case Studies on Critical Elements for Success
Stephen Rowe, Dekra, UK

Reactive chemistry PHA's require additional input data and competence for efficiency and effectiveness. This paper identifies deficiencies in reactive chemistry PHA's as the root cause of almost all major global runaway reaction incidents. A real-world client example is used to demonstrate the effectiveness of PHA’s where reliable reactive chemistry input data is available to the team.

A Novel Cryogenic ARC Testing Method Employing Liquid Nitrogen to Freeze Materials Separately and Evaluate Explosive Hazards during Mixing
Zhanhua Wu, East China University of Science and Technology, China

A Detonation Incident Due to the Redox Reaction in the Solution of Perchloric Acid and Ethanol used in the TEM Electropolishing Step
Weiye Li, East China University of Science and Technology, China

A fatal detonation killing a graduate student during Transmission Electron Microscopy (TEM) sample preparation highlighted the extreme explosion hazards of perchloric acid-ethanol electropolishing mixtures. Addressing quantitative data gaps, this study utilized a self-developed anti-explosion calorimeter, DSC, and ARC to evaluate thermal runaway under abnormal conditions. Results reveal massive decomposition energies up to 4409 J/g (1.05 TNT equivalent). Crucially, stainless steel inside twin-jet electropolishers acts as a fatal catalyst, plunging the decomposition onset from 135 °C to 40 °C and triggering detonation. Additionally, short-circuit sparks instantaneously detonate the system. These findings provide a thermodynamic basis for urgent equipment redesign.

Small Scale Industrial Explosions – An IChemE Learned Society Project – Lesson Learnt and Further Work Required
Keith Plumb, Integral Pharma Services, UK

This paper presents an update on an IChemE Learned Society Project covering small-scale industrial explosion focusing on volatile liquids and dusts. This project exposed the significant gaps in understanding of small-scale explosions and indicates the need for more education and training and the scope for further research by both academics and commercial research organisations.

Siting for BESS Developments: Quantifying Risk for Planning Decisions.
Karina Almeida Lenero, Gexcon, UK

Battery Energy Storage Systems (BESS) introduce major hazard scenarios associated with thermal runaway. Key hazards include the dispersion of toxic substance, in particular hydrogen fluoride and the accumulation of flammable gases within enclosures, which could result in a vapour cloud explosions. This study presents a structured methodology for consequence modelling and frequency analysis to quantify the risk to the public, considering the effects of ventilation on flammable cloud build-up and environmental conditions for the external dispersion of toxic gases. The study will provide practical insights for siting BESS installations and can be used for Land-Use-Planning assessments.

Fire, Explosion, and Toxicity Risk Assessment of Li-Ion & Na-ion Battery Energy Storage Systems
Fadi Nweihi, Vysus, Norway

This study presents a comparative safety assessment of lithium-ion and sodium-ion Battery Energy Storage Systems (BESS) for offshore and renewable energy applications. The work combines a review of battery technologies with HAZID, empirical analysis, and CFD consequence modelling to evaluate fire, explosion, toxicity, and gas dispersion hazards. Two representative case studies were assessed: an offshore energy recovery installation and a utility-scale renewable energy storage facility. The study compares hazard profiles, mitigation measures, fire protection and detection strategies, and emergency preparedness requirements, providing recommendations for risk-informed deployment and future standards development for large-scale BESS applications

Functional Safety in Battery Storage: Lessons from Design, Integration and Assurance
Tom Ridout, ERM, UK

Battery systems are central to the energy transition but introduce significant safety challenges, including thermal runaway and complex fire and explosion risks. Managing these hazards requires coordination across design, control, and operations throughout the lifecycle. Drawing on the UK Battery Industrialisation Centre's (UKBIC) prototype battery pack project, this presentation explores practical lessons from applying the functional safety lifecycle, supported by ERM. It highlights key challenges and solutions in design, system integration, and assurance, demonstrating how established techniques can be adapted for battery systems and other emerging energy technologies where complexity and evolving standards make consistent application of functional safety challenging.

Improving Management of Change Performance Through Continuous Improvement (Lean Six Sigma)
Maryam Alburaidi, OQ Gas Network, Oman

This CI project demonstrates how Lean Six Sigma methodology was applied to enhance the Management of Change (MoC) process at OQ Gas Networks. The initiative addressed long lead times, and process inefficiencies affecting operational performance and risk management. Using the DMAIC approach and analytical tools such as SIPOC, process mapping, and root cause analysis, several improvements were implemented, including process re-engineering, system enhancements, accountability and sustainability plan. As a result, MoC lead time was reduced from 110 days to 41 days, while rejected quality decreased from 81% to 21%, significantly improving compliance, workflow efficiency, and process safety governance across organization.

Continuous Improvement in Major Accident Hazard management on an Upper Tier COMAH Site Using Cyclic Process Hazard Review
Nick Wise, SSE Thermal, UK

SSE Thermal has undertaken five-yearly Process Hazard Reviews (PHRs) for its Aldbrough Underground Gas Storage facility since 2012, supported by TÜV Rheinland Industrial Services. This paper explores how SSE Thermal has developed the PHR process, with incorporation of related safety studies to improve consistency through using the same study team. It explores the benefits of holding the PHR a year in advance of the COMAH Safety Report submission date to integrate the PHR into the Safety Report update process, and considers how the PHR process can address requirements in the new CDOIF Major Human Harm Hazard Risk Assessment guidance.

The Evolving Landscape of Best Practice in ALARP for COMAH Sites, Following MHHHRA changes.
Owen Llanwarne, OTECSA Consulting, UK

Our presentation explains how new HSEx guidance, based on CDOIF’s MHHHRA methodology, is reshaping COMAH safety reporting by replacing the older SRAM approach. MHHHRA is becoming standard practice and requires more comprehensive assessment of major accident hazards, with stronger ALARP demonstrations and a more QRA-based approach. This significantly increases the complexity and workload for COMAH operators, especially new entrants and Lower Tier sites moving to Upper Tier due to hazardous material reclassification. Drawing on first-hand consultancy experience, the presentation aims to simplify these new expectations, clarify technical implications, and explore how they may affect future safety report updates and reviews.

Process Safety Design for Dense Phase CO₂ in CCS Plant Facility
Kei Matsuzaki, JGC Corporation, Japan

The growing adoption of CCS in the oil and gas industry introduces new process safety challenges associated with high-pressure dense phase CO₂. This paper presents a practical design approach based on project experience, starting with inherently safer design and extending protection layers and risk reduction measures. Design measures for pre-loss of containment such as BPCS, ESD, and emergency depressurization systems are discussed alongside post-LOC mitigation, including CO₂ detection, cryogenic protection, and emergency response. The integration of these elements provides a structured framework for managing both the likelihood and consequences of CO₂ release incidents in CCS applications.

Quantitative Risk Assessment for Multi-mode CO₂ Onshore Transport
Anindityo Patmonoaji, University College London, UK

Expanding CCUS infrastructure requires moving massive volumes of high-pressure CO2, introducing lethal toxicity risks if containment fails near populated areas. Conventional QRAs lack exposure-time metrics and often rely on oil and gas data surrogates due to limited CO2 history. This work introduces an enhanced QRA framework. It models dynamic releases and applies probit functions to map precise lethality zones, while deriving accurate likelihoods from three decades of pipeline, road, and rail data from the UK HSE, EGIG, and US PHMSA. The resulting predictive framework enables risk-informed modality and route selection, embedding inherent safety into decarbonization supply chains.

Advancing Safe CO₂ Vent Design: Reducing Uncertainty in Depressurisation Behaviour and Impurity‑Driven Hazards
Henry Tran, Wood Group, UK

The ability to vent inventory is an requirement for the design and operation of process facilities. This presentation examines process safety challenges associated with CO₂ venting in CCS systems; focusing on low-temperature behaviour during depressurisation and impurity-driven toxicity risks. This will be achieved by sharing experience, modelling insights, and lessons learned from a collaborative project between Wood, National Gas Transmission (NGT), SGN, and Ofgem. Drawing on scenario modelling, CFD validation, and joint industry research, the presentation provides guidance to improve design conservatism, refine risk assessments, and develop safer vent systems. Delegates will gain transferable lessons applicable to CCS.

What Determines Acceptability of Residual Risk - Experience from Different Jurisdictions
Pat Swords, PM Group, Ireland

The principle of As Low As Reasonably Practicable (ALARP) is enshrined in UK law and supported by the Health and Safety Executive's publication R2P2 providing quantification of its boundaries. However, other jurisdictions take a different approach depending on their legal structure and as to how they adopt the use of technical standards. The paper therefore addresses the regulatory context in a variety of international jurisdictions, including Europe, USA, China and India, for the determination of the acceptability of residual risk and will be of interest to engineers, who are increasingly required to work in an international environment.

Regulatory Gaps for UK Onshore Carbon Capture and Storage (CCS):  A Practitioner Framework for Temporary CO₂ Storage, Port Terminals, Ship Loading and Export Interfaces
Ashish Kelkar, Aker Solutions, UK

This paper focuses on the interface scenarios within UK-CCS developments that give rise to these inconsistencies, with particular emphasis on temporary storage and transfer nodes. The framework introduces a structured basis for hazard identification, barrier governance, and CO₂‑specific hazardous atmosphere management, together with clear expectations for emergency response at system interfaces.The framework is intended to provide a practical approach for designers and operators, particularly during early project phases where regulatory expectations are least defined. By improving consistency in safety justification and interface management, it supports more robust and auditable decision‑making as CCS infrastructure scales-up across the UK.

Are Your Suppliers Ready for the Cybersecurity Resilience Act (CRA)?
Richard Butler, Exida, Ireland

The European Cyber Resilience Act (CRA) comes into force this year, with full compliance required by December 2027 for products with digital connectivity. While impacts on suppliers are clear, implications for asset owners are less well understood. Industrial users depend on CRA-affected equipment such as PLCs, DCSs, safety systems, and network devices. Post-2027, these products will likely be more secure but more costly, with some solutions becoming unavailable due to the onerous nature of CRA compliance. This paper outlines the risks to installed systems and supply chains and provides guidance to assess exposure, and future-proof for the CRA.

Invited speaker: Boiler Blowback Incident
Matthew Branny, Astra Zeneca, UK

A case study of a boiler blowback incident during commissioning, examining how a seemingly minor design modification led to equipment damage and exposed gaps in hazard review and testing processes.

Marsh - 100 Largest Losses - Learning from Root Cause Analysis Patterns
Natali Walton Chacin , Marsh, UK

In today’s rapidly evolving energy and power sector, companies face complex operational challenges, regulatory pressures, and increasing exposure to a wide range of risks. Understanding the nature and scale of these risks is critical—not only to protect your assets but also to ensure long-term resilience and sustainable growth. The presentation will examine the most frequent immediate causes and barrier failures for major losses, highlight trends across upstream, downstream, and midstream losses, and reflect on what these patterns mean for risk leaders today.

Behavioural Analysis of an Incident - A Case Study
Lorraine Braben, Braben Consulting, UK

The paper describes a behavioural analysis conducted on a loss of containment incident, which had already been investigated through a standard investigation technique. The organisation specifically wanted to explore the potential human factors which had contributed to the release. The Kestrel method is based on psychological research to identify barriers which impact the desired behaviour, and then to suggest suitable interventions associated with the barrier type to address the problem. Applying psychological principles to investigate an incident was a novel approach and is another useful tool for enhancing risk management.

Chernobyl Beyond the Control Room: Systemic Failure and Organisational Drift in a High-Hazard System
Zsuzsanna Gyenes, Global Industrial Safety Solutions, UK

Chernobyl did not begin with an explosion. This paper examines the accident as a systemic failure shaped by pressure, drift and movement outside the safe operating envelope. Through the turbine rundown test on Unit 4, it explores how technical vulnerability, operational pressure, hierarchy, fatigue and repeated deviations combined to reduce resilience while the system still appeared manageable. The lessons remain relevant across high-hazard industries where pressure to continue and complete the task can make stopping increasingly difficult. The paper also reflects on the wider consequences of process safety failure beyond the plant boundary.

Applying Functional Safety Principles to Electrical Protection Systems: Updates from the Second Edition of the Energy Institute Guidance
Hugo Siddle, WSP, UK

This paper presents the second edition of the Energy Institute’s ‘Guidance on Assessing the Safety Integrity of Electrical Supply Protection’. It presents a structured, risk-based methodology for determining and demonstrating the safety integrity requirements of electrical protection systems, aligned with the functional safety principles of IEC 61508 and IEC 61511. The guidance is intended for scenarios where electrical schemes differ from standard practices due to novel technology or have the potential for high-consequence hazards. The paper outlines the practical application of the functional safety lifecycle to electrical protection systems, supported by a worked example.

Digital Twins for Functional Safety: Emerging Opportunities, Future Prospects and Technical Challenges
Nirav Chokshi, KBR, UK

Digital twins can provide powerful tools for enhancing safety assurance in complex processes and systems. They offer new opportunities for continuous safety validation, real‑time risk assessment, and adaptive safety functions. However, major challenges remain that include ensuring model fidelity, managing real‑time data quality and synchronisation, and addressing computational demands of high fidelity simulations. Regulatory acceptance will also be a key barrier, as current standards rely on physical testing, and cybersecurity risks grow with increased data connectivity. Despite such challenges, digital twins present significant opportunities to reduce development costs, enable virtual testing, and support predictive, data‑driven safety assurance.

Surrogate Models for Predicting Hydrogen Dispersion During Liquid Hydrogen Releases
Jack Stevenson, University of Aberdeen, UK

This presentation discusses the use of graph neural network surrogate models to predict hydrogen dispersion following liquid hydrogen releases. High-fidelity CFD simulations were used to train a graph-based model capable of reproducing key dispersion features on an unstructured mesh. The work demonstrates how AI-based surrogate models could accelerate consequence analysis, support scenario screening and contribute to future hydrogen safety digital twins. The presentation will describe the modelling workflow, initial prediction results and key limitations identified from this case study.

A Multi Domain Modelling Framework for Integrated Hazard, Resilience, and Energy System Assessment.
Daniel Cooper, Amentum, UK

This paper presents a multi-domain digital twin framework for assessing hazard, resilience, and energy system behaviour in nuclear-backed AI data centre colocation scenarios. By integrating models of nuclear generation, data centre loads, cooling systems, grid constraints, and environmental stressors, the approach captures cross-domain interactions and emergent risks not addressed by traditional methods. The work demonstrates how uncertainty-driven, scenario-based analysis can identify critical vulnerabilities and inform early-stage design decisions, offering Hazards delegates a practical methodology for extending process safety approaches to increasingly complex, interconnected infrastructure systems.

RAM Does Not Eliminate Uncertainty – It Exposes It. Reframing RAM for New Technologies: From Prediction to Governance Under Uncertainty
Hortensja Malek, Fluor, UK

RAM analysis is often questioned when applied to new or first of a kind technologies where failure data is limited. This paper recognises that RAM cannot provide reliable availability predictions for immature systems. Its value lies elsewhere. When used appropriately, RAM helps to make uncertainty visible, examine assumptions, and explore how system performance may vary as designs and operating arrangements change. Drawing on practical project experience, the paper shows how a transparent use of RAM can support informed decision making, governance, and assurance for new technologies, without overstating confidence or overstretching the evidence.

Good Engineering Practices for Storing and Handling of Highly Toxic Chemicals
Asith Anantharama, Chemical & Speciality Chemicals, India

AHF is a critical industrial chemical used as a primary fluorine source for manufacturing fluorochemicals. However, it is highly toxic IDLH: 30 ppm and can be fatal if inhaled or through skin contact. Given these risks, it is essential that companies handling AHF take all necessary precautions to prevent Loss of Primary Containment by RAGAGEP, such as those outlined in API 752 and Eurofluor guidelines.To ensure safety and optimize resources, a systematic, risk-based approach should be used to evaluate the potential for accidental release during storage and handling. This allows for risk-based intelligent investment and prevents over- or under-engineering

Risk Based Development of Minimum Safe Distances for Gas Pipelines in High Occupancy Areas (HCAs)
Afida Sikandar, Petronas, Malaysia

Rapid urbanisation has increased pipeline proximity to high occupancy areas (HCAs), exposing gaps in existing standards that lack explicit, risk based separation criteria. This paper presents a methodology using consequence modelling to define minimum safe distances, supported by benchmarking and technical requirement enhancement. Incorporating lessons from recent industry incidents, the work emphasises conservative design, evolving land use risks and public vulnerability. The proposed framework strengthens pipeline routing decisions, improves regulatory alignment and enhances public safety in densely populated environments.

Bringing Historic British Best Design Practice into the 21st century
Malcolm Toft, Cavendish Integrity Services, UK

Many companies have inherited extensive suites of best practice documentation, and we are seeing more requests from clients interested in exploring if these documents remain proportionate to risks, and if optimisation is possible. This is a vital component of retaining global competitiveness in our chemical industry. We discuss some of our recent experience in assisting clients answering these questions, considering their evolving risk profiles, factors specific to the energy transition, changes in relevant codes, and organisational evolution. In some cases, we have been able to help achieve significant streamlining by careful attention to the key aspects of clients’ plant standards.

Barrier Model Improvements - An Operators Perspective
Matt Hastings, Ithaca Energy, UK

A new Barrier Model software tool was introduced, using standard KPIs and traffic‑light indicators to highlight deviations and risks. After six months, several enhancements were identified. This presentation explains why changes were needed and what was updated from an end‑user perspective. The model aims to visually represent asset risks, enable cross‑asset comparison, prioritise work, track trends, and identify cumulative risk. Following a company merger, a unified “best of both” model was created with revised KPIs, thresholds, and weightings to avoid double counting and better reflect risk. Testing and stakeholder review confirmed the updated model’s improved usefulness.

From Static Mapping to Dynamic Barrier Health: AI-Assisted Assessment of Fire and Gas Detection Coverage Degradation
João Antônio Pereira de Paula, PRIO, Brazil

Fire and gas detection coverage is commonly assessed through mapping studies, but actual barrier performance may degrade during operation due to failed detectors, inhibited channels, overdue calibrations, bypassed logic, temporary obstructions or maintenance backlog. This work proposes an artificial intelligence framework that links Fire and Gas Mapping assumptions with operational data from CMMS records, bypass logs, test results and impairment registers. The approach evaluates whether detector unavailability affects coverage, redundancy, voting logic and major accident scenario detection, supporting dynamic barrier health monitoring, maintenance prioritisation, impairment management and ongoing assurance of Fire and Gas detection performance.

AI-Assisted Restart Readiness: Using Hazard Intelligence to Surface Hidden Risk Before Startup
Ivo Dujmovic, uc2c.ai, USA

AI-Assisted Restart Readiness describes how hazard intelligence can strengthen startup safety after shutdowns, turnarounds, or degraded operation. By linking evidence across management of change records, procedures, incidents, maintenance findings, and barrier status, the approach creates a transparent and reviewable picture of restart risk. Rather than replacing PSSR or engineering judgment, it supports them by improving completeness, traceability, and preparation speed. The result is earlier visibility into interacting changes, impaired safeguards, stale assumptions, and the highest-priority checks required before startup.

Beyond Algorithms: Human Trust, Governance, and Cultural Perspectives on AI in Process Safety
Maria Kamran, Bureau Veritas, UK

Artificial Intelligence (AI) is increasingly applied in process safety, supporting tasks such as HAZOP by analysing large datasets to identify hazards and credible scenarios. While it enhances risk-based decision-making, challenges include system complexity, governance, accountability, and trust. Achieving safe and effective integration requires calibrated trust, human oversight, and robust governance frameworks.

Estimating Blast and Fragment Hazards for Accidental Explosions in Liquid-Propellant Fuelled Launch Vehicles
Zhixin Hu, UK Civil Aviation Authority, UK

Safety cases are submitted for UK spaceport and launch license applications, in which different models are used in flight and ground safety analyses. This paper aims to provide a critical review of commonly used models for characterising liquid-propellant launch vehicle explosions, evaluating their strengths and weakness in terms of predicting blast effects, fragmentation behaviour, and debris catalogues in support of launch safety assessment and risk management.

Prediction of Dust Explosion Properties of Homogeneously Combustible Dust
Asma Abousrafa, Texas A&M University, USA

Dust explosions pose significant risks in powder processing facilities. Industry evaluates these risks by quantifying dust explosion properties. However, existing dust explosion models remain limited specifically on homogeneously combustible dusts which vaporize to form flammable mixtures that ignite once met with an ignition source. This study focuses on the development of a phenomenological model to predict the minimum explosible concentration and minimum ignition energy. The model describes uniformly dispersed particles exposed to a transient spark. The properties are determined when Ignition occurs as the vaporized fuel exceeds the lower flammability limit, and when flame propagation occurs across dust layers.

Introducing the Modal QRA – Enabling Real-time Risk Calculations
Angel Casal, Vysus Group, Spain

An overview of a new method to assess risk in real time will be presented. This method has been implemented in a tool which aids asset managers to plan operation and maintenance activities in major hazard facilities taking into account the actual level of risk and not an average. The level of risk will change real-time depending on external and other factors, status of barriers, manning levels and the type of operations being carried out at each moment.

On the Importance of NOx Oxidisation in Dispersing Exhaust Plumes for EH40 Compliance: with Particular Reference to Decommissioning Offshore Assets
Steve Howell, Abercus

This paper raises a concern regarding NOx exposure from exhausts on offshore platforms at low wind speeds, particularly as the role of diesel generators has expanded on platforms entering the decommissioning phase where they may be run continuously or near continuously. Specifically, the composition of NOx in a dispersing exhaust plume will vary as NO is oxidised and becomes NO2 over time, and this can have a significant impact upon the toxicity of the dispersing plume. The case study presented demonstrates that assuming a fixed NO2:NOx ratio representative of the exhaust gases leaving the exhaust outlet is under‑conservative.

Computational Fluid Dynamics (CFD) Carbon Dioxide Dispersion Modelling Experimental Comparison and Validation
Tim Jones, Tetra Tech, UK

This paper validates CO2 release modelling using the KFX CFD package against experimental data from the CO2PIPETRANS and COSHER programmes. It examines both horizontal above-ground releases and buried full-bore rupture scenarios, focusing on how source term assumptions such as inlet temperature, mass flow rate, and velocity affect predicted concentrations. The study highlights the importance of accurately representing CO2 thermodynamics, including dry ice formation and sublimation, to support safety cases, pipeline design, emergency planning, and separation distance assessments. The work informs good practice for CCUS projects and ongoing validation for the Skylark CO2 experimental programme.

Evaluating the Reliability of Consequence Modelling for Hydrogen and Hydrogen‑Blend Gas Releases: Insights from Full‑Scale Testing
Oluwakemi Eniolorunda, Wood Group, UK

As gas transmission networks transition toward hydrogen and hydrogen‑blend operation, predicting the consequences of gas releases becomes critical to safety and design. This study evaluates the reliability of existing consequence modelling tools using full‑scale venting and flaring demonstrations of hydrogen blends and pure hydrogen. Measured parameters such as noise, thermal radiation, and overpressure, were used to inform an integrated modelling framework. By comparing multiple tools, the work highlights differences in predictive behaviour and underlying assumptions. The study demonstrates the importance of understanding modelling limitations to support robust, cost effective, and defensible engineering decisions for hydrogen infrastructure.

Quantitative Risk Assessment of Hydrogen and Ammonia Facilities with Carbon Capture: Modelling Challenges and Practical Guidance
Faiq Raedaya, MES International, UK

Blue hydrogen, ammonia, and CCS facilities are being designed against QRA workflows that were calibrated for hydrocarbons and not validated for NH₃, H₂, or CO₂. Drawing on a Pre-FEED QRA for a large-scale blue H₂/NH₃ facility with CCS, this paper sets out where standard practice fails and what to do instead — covering consequence model selection, frequency assessment, inventory treatment, and sensitivity design for each substance. Delegates take away a substance-by-substance modelling checklist, a candid map of the remaining gaps in tools and data, and transferable lessons on managing early-stage QRAs.

Beyond the Risk Contour: Leveraging Multiple Risk Metrics for Ammonia Safeguard Decisions
Thirakorn Mokkawes, Agency for Science, Technology and Research, Singapore

This contribution presents an anonymised ammonia QRA case study showing why safeguards should not be selected by maximum contour reduction alone. Safeti/Phast outputs were translated through an impactQ/Terra co-pilot workflow into comparable metrics for PLL, injury proxy, land footprint and land-use opportunity cost. Although each metric derives from the same contour results, they aggregate it differently, weighting by frequency, population or area, so the rankings disagree: water curtain performed most consistently, while double-wall pipe gave the highest fatality PLL reduction yet no injury-metric change. The work offers Hazards delegates a transferable approach for ALARP discussions when metrics disagree.

Reframing Semiconductor Manufacturing as a Process Safety Discipline: Identifying Process Safety Management Gaps in High-Hazard Semiconductor Operations
Edison Sripaul, Texas A&M University, USA

Semiconductor fabrication facilities handle significant inventories of acutely toxic, pyrophoric, flammable, and reactive chemicals, yet have historically been governed through occupational health and contamination control frameworks rather than process safety management (PSM) principles. This paper argues that semiconductor fabs should be recognized as high-hazard chemical processing systems and examines the applicability of established PSM frameworks to this sector. Critical gaps are identified across process hazard analysis, management of change, asset integrity, contractor management, and safety culture. The paper proposes a pathway toward integrating risk-based process safety principles into semiconductor operations, and invites dialogue between the semiconductor and process safety communities.

Repurposing Facilities for Automotive R&D: The Financial Case for Early-Phase Hazardous Materials Reports (HMR)
Edward Naranjo, Kenexis Consulting Corporation, USA

As the automotive industry shifts toward hydrogen fuel cells and volatile fuels, R&D facilities face complex compliance challenges under the IBC and NFPA 2. Integrating a Hazardous Materials Report (HMR) early in design prevents the "H-Occupancy Trap"—avoiding costly high-hazard classifications. Using a Midwestern Tier-1 supplier case study, the text illustrates how strategic HMR use achieves significant capital savings, operational flexibility for new fuels, and faster permitting. By treating the HMR as a front-end design driver rather than a final checkbox, firms reduce project risk and save millions in construction costs.

The Other Side of Hydrogen Production - Decarbonising the water industry
Lisa Griffin, Anglian Water, UK

There is a lot of talk about hydrogen production, and how hydrogen will support the decarbonisation of our energy sector – but there is a further positive side. Hydrogen production could support the decarbonisation of the waste water sector. This paper explores this alternative side of hydrogen production, and shares the results of a current pilot trial.

Assessing Flammable Gas Production in a Food Waste Treatment Plant - A Model Approach
Christophe Gilson, WSP, UK

WSP have recently been appointed to support a client in the commissioning of an Acid Phase Digestion (APD) plant upstream of a mesophilic anaerobic digestion process. As part of this work, WSP have taken a model based approach to characterise biogas quality and to identify potential risks associated with flammable and hazardous gas formation. A model is currently under development and calibration, using three years of operational data, to provide a dynamic platform enabling operations, suppliers, and contractors to visualise how the food waste AD plant performs under different operating conditions, supporting safer and more efficient plant operation.

Operational Resilience of a High Hazard Acid Transfer System
Jaananni Gandhi , Rolls-Royce, UK

This presentation reviews the operational resilience of an automated hydrofluoric (HF) acid transfer system designed to eliminate manual pouring risks in a legacy facility at Rolls-Royce. Developed via rigorous hazard studies (HAZID to Hazard Study 5), the system has enhanced operator safety over three years of service. However, the chemical’s extreme corrosiveness has resulted in loss of containment and technical issues, explored through three case studies. Key findings highlight the tension between complex engineered safety systems and maintenance accessibility in high-hazard environments, offering lessons learnt on sustaining automated systems for hazardous chemical handling.

Resilience: a Powerful Tool in the Safety Toolbox
Sander Andela, NRG PALLAS, Netherlands

Resilience is usually defined as the ability of a system to recover quickly from both known and unknown external events. NRG PALLAS has performed a literature study into the concept of resilience and potential evaluation methods of the resilience of a system. It was discovered that a resilience-based strategy of protection can contribute significantly to the safety of a system. We highlight the advantages, and disadvantages, of a resilience-based strategy of protection and give examples of resilience-based engineering. Also, the concept of measuring the resilience of a system will be discussed, including specific methods like CIRAF and GTST-DMLD.

Bridging Safety and Security: Assessing Natural Hazard Effects on Chemical Site Protection
Federica Ricci, University of Bologna, Italy

Natural hazards may impair the effectiveness of Physical Protection Systems in chemical and process facilities, increasing vulnerability to intentional attacks during and after extreme events. This study investigates the impact of seven natural hazards on seventeen security elements through an expert elicitation involving 61 multidisciplinary experts. Quantitative performance modification factors were derived to estimate the probability of damage of security systems under hazardous conditions. Results indicate that floods, earthquakes, and wildfires may result in significant degradation of security systems, with impairment probabilities reaching 0.95 for specific elements. The findings support the integration of natural-hazard effects into Security Risk Assessment methodologies.

Simplifying DSEAR Compliance for Food Sector - Focus on Combustible Dust Hazards
Oyinda Gunn, AXIOM, UK

This paper provides a step-by-step guide to help workplaces in the food sector achieve DSEAR compliance by translating regulatory requirements into a clear, structured, and practical methodology specifically tailored for food manufacturing environments. It highlights a critical blind spot whereby common, everyday materials such as flour, sugar, cinnamon, and other powder spices are routinely perceived as inherently safe. Familiarity with these materials can lead to an underestimation of their hazardous potential and inconsistency in the assigned risk level even though they prevent high explosion hazards under certain processing or handling conditions.

Leakage Without Harm - Operationalising 'No Significant Risk' in Geological CO2 Storage Projects
Matt Baggaley, Risktec, UK

This presentation examines what the EU CCS Directive’s requirement for 'no significant risk' should mean in practice for geological CO2 storage. It argues that leakage risk is not the same as harm-based risk, and that treating the two as equivalent can obscure key questions of storage integrity. The presentation proposes a practical and defensible framework for interpreting significance, drawing on major hazard risk practice while remaining consistent with the Directive. The aim is to support clearer regulatory decision-making and give operators a more workable basis for demonstrating compliance.

Carbon Dioxide Vessel Failures: a Review of Relevant Incidents, Models and Experiments
Simon Gant, HSE, UK

This work reviews CO₂ vessel failure incidents, consequence models and the experiments that can be used for model validation.

Experimental Study of Pipe Wall and Jet Temperatures During Gaseous-phase CO2 Release Through Orifices
Raphael Oberti, NaTran, France

NaTran is adapting its infrastructure for the energy transition, including CO2 transport. In 2024, NaTran R&I partnered with INERIS to conduct controlled leakage tests through 1 and 2 mm orifices on a CO2-filled pipe (17–41 bar, 8–16 °C). Instrumentation captured pressure and temperature along the line, at the pipe wall, and in the jet. Results showed significant wall cooling, stronger than for natural gas or hydrogen, increasing with pressure and orifice size. This highlights potential embrittlement risks for CO2 pipelines. Future work will address dense-phase conditions and operational constraints.

CO₂ Consequence Modelling: Insights from Two Decades of Validation Against Experiments
Jan Stene , DNV, UK

Executing accurate, cost-effective quantitative risk assessments for expanding CCS infrastructure requires an evidence-led modelling strategy. This paper synthesises two decades of work on CO2 hazards across large-scale Spadeadam experiments, Phast integral modelling and KFX 3D CFD simulations. Drawing on validation data from the BP, Shell, CO2PIPETRANS, and COSHER campaigns, we highlight key advancements in multiphase thermodynamics, pipeline decompression, crater formation, and dispersion modelling. The paper explores where efficient integral tools are appropriate and where scenario complexity may justify escalation to 3D CFD while considering model confidence levels for different scenarios.