Historical process safety incidents in the Chemical Process Industries constitute a well-documented record of failure modes, escalation pathways, and systemic vulnerabilities that extend well beyond their original mechanical or procedural causes. This study examines 122 notable industrial incidents from 1980 to 2025 across the refinery, petrochemical, gas processing, and bio-process industries and demonstrates how these historical events can inform the development of cyberattack scenarios in modern digitally integrated facilities. The scenario mapping methodology is grounded in the US Department of Energy's Cyber-Informed Engineering Framework Principle 8 calls on engineers to systematically hypothesize how digital functions could be abused or misused to cause harm, and to use those exploitation pathways to drive both digital and engineered mitigations. By applying this principle to the historical process safety record, selected incidents are mapped to realistic cyber-physical attack scenarios involving sensor spoofing, alarm suppression, safety instrumented system bypass, and audit trail manipulation. Although these incidents were not originally caused by cyberattacks, their initiating conditions and escalation mechanisms present credible exploitation pathways in today's threat landscape. This work advances cyber awareness among process engineers, safety professionals, and plant operators, and reinforces the need to integrate cybersecurity considerations into established hazard analysis and risk management frameworks.
Process Safety Management (PSM) programs are essential for preventing catastrophic incidents in high-risk business sectors, such as oil and gas, petrochemicals, liquefied natural gas, ammonia refrigeration, and agricultural chemical industries. However, logical fallacies and cognitive biases undermine these programs, leading to flawed decision-making and increased safety risks. This paper examines key logical fallacies, such as post hoc ergo propter hoc (inferring causation from mere temporal sequence), false dichotomy (oversimplifying choices), and appeal to authority (relying on unverified expertise), alongside cognitive biases, including confirmation bias (favoring data that supports existing beliefs), overconfidence bias (overestimating judgment accuracy), and anchoring (over-relying on initial information). These human-driven errors impact critical PSM elements-hazard identification, risk assessment, training, and incident investigation-often resulting in systemic failures. For example, overconfidence may lead to underestimating hazards in risk assessments, while confirmation bias can skew incident investigations by focusing on familiar causes and ignoring critical evidence. Geographic and cultural factors may amplify these vulnerabilities. In the Global South (e.g., Africa and Latin America), resource constraints may foster optimism bias, where limited safety investments are sometimes deemed sufficient by business decision-makers. In contrast, industrialized nations like the United States may exhibit overconfidence due to robust regulations, assuming compliance ensures safety. Cultural norms, such as hierarchical structures in Japan or collectivist tendencies in parts of Asia, can reinforce Groupthink or appeal to authority, suppressing dissenting safety concerns. These regional differences highlight the need for tailored PSM strategies. To mitigate these risks, the paper proposes practical solutions: structured decision-making tools, like risk matrices, reduce biases such as anchoring; bias-awareness training encourages critical thinking to counter fallacies like post hoc reasoning; data-driven analytics challenge subjective assumptions; and generative artificial intelligence and large language models enable broader knowledge sharing and cross-organizational learning. These measures are adaptable across global contexts, from resource scarcity to advanced economies, ensuring robust PSM without ideological influence.
Hydrogen infrastructure is expanding across industrial production, compression, storage, transport, distribution, and vehicle-refueling applications. This expansion increases the need for incident-based safety engineering methods that connect hydrogen hazards to degraded barriers, control losses, and emergency intervention requirements. This study develops and analyzes a structured incident dataset covering hydrogen-related events across production, storage, transport, distribution/refueling, industrial use, and maintenance contexts. The dataset codes each event according to infrastructure segment, operational phase, hydrogen form, initiating event, failure mode, loss of containment, ignition, explosion, fire, detection issues, isolation issues, ventilation issues, barrier degradation, and safety control requirements. The analysis shows that hydrogen accident prevention cannot be reduced to component integrity alone. Recurring scenarios involve interactions among containment, detection, isolation, ventilation, maintenance control, emergency shutdown, and response readiness. The paper contributes a system safety-oriented coding framework for hydrogen infrastructure and derives practical control requirements for process safety, loss prevention, and emergency intervention.
Abstract The 1976 Seveso disaster prompted a major shift in European industrial safety, leading to the Seveso Directives to prevent major chemical accidents. This review outlines the evolution of the Seveso framework, specifically focusing on the transition to the Seveso III Directive. While this framework successfully aims to harmonize process safety, significant regulatory divergence still remains among member states regarding land‐use planning, inspection regimes, and safety management systems. To address these implementation gaps, advanced risk assessment methodologies like the collaborative ARAMIS framework, probabilistic Bayesian networks, and ontological models are increasingly utilized. These advanced approaches help operators manage complex, emerging risks, including cascading domino effects, natural hazard triggering technological disasters (NaTech), and aging infrastructure. Furthermore, the digital transformation of Industry 4.0 highlights the critical need to integrate cybersecurity into process safety. Ultimately, resilient industrial risk management requires a robust socio‐technical approach that combines technical safety, integrated safety‐security frameworks, and active community risk communication.
Abstract This paper presents a comparative analysis of three HAZOP studies conducted on a small generic solvent purifier system: one using Generative AI (GenAI), one employing an Expert System tool (HAZOP Assistant), and one conducted as a traditional human workshop. The GenAI approach, leveraging probabilistic models, identified hazards effortlessly but introduced inconsistencies as well as proved non‐compliant with the European Union (EU) AI Act's (Regulation (EU) 2024/1689) transparency requirements for high‐risk applications. In contrast, the HAZOP Assistant, with its algorithm‐based logic, achieved a faster analysis than the human workshop. It only generated deviations related to flow, pressure, temperature and level, but managed to fill in all the results from the traditional workshop and highlighted some missed scenarios from it. This makes it possible to ensure that the outputs are aligned with standards and are auditable. The human workshop, while thorough, required more time and personnel, and missed some of the hazards identified by the HAZOP Assistant. This study demonstrates that even though GenAI demonstrates great promise, older, deterministic types of computing have the ability to enhance efficiency and quality in safety studies, offering a scalable solution for process safety while meeting stringent regulatory demands.
