Proactive Risk Mitigation and Asset Integrity using Physics-Informed AI: A Framework for Hydrogen Production and Petrochemical Applications
Fuente:
Zenodo
Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Recurso digital |
| Pubblicazione: |
Zenodo
2026
|
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866901802120642560 |
|---|---|
| author | Shibah, Sami Rashid Mohammed |
| author_facet | Shibah, Sami Rashid Mohammed |
| contents | <p>Process safety in high-consequence industries such as petrochemicals and hydrogen production faces significant challenges from complex, dynamic systems where traditional reactive safety approaches often fail to predict emerging risks. This paper presents a novel framework for integrating Responsible Artificial Intelligence (RAI) and physics-informed modeling to transition from reactive monitoring to proactive risk mitigation. Our core methodological contribution introduces a hybrid AI architecture based on Physics-Informed Neural Networks (PINNs) coupled with Bayesian Uncertainty Quantification (UQ). This framework is specifically designed for predictive risk assessment in inherently complex systems. We demonstrate its application through a high-fidelity quantitative simulation of a hydrogen production unit, focusing on predicting catalyst degradation and potential runaway reactions. The methodology rigorously integrates first principles (reaction kinetics and heat transfer equations) with operational data to generate predictions characterized by quantified uncertainty (epistemic and aleatoric). Quantitative results, derived from a fully reproducible Python simulation, demonstrate a significant improvement in anomaly detection sensitivity (ROC AUC of 0.98 compared to 0.84) and a quantifiable reduction in detection latency (4 hours compared to 18 hours) compared to traditional Statistical Process Control (SPC) methods. This reduction in false positive rates is crucial for preventing operator fatigue in safety-critical applications. By embedding physical laws, the model offers verifiable trustworthiness and enhanced explainability for human-in-the-loop decision-making. The framework is directly applicable to key petrochemical processes including fluid catalytic cracking (FCC) units for coke deposition monitoring and Carbon Capture, Utilization, and Storage (CCUS) systems for amine solvent degradation and CO₂ injection risk assessment.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19563668 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Proactive Risk Mitigation and Asset Integrity using Physics-Informed AI: A Framework for Hydrogen Production and Petrochemical Applications Shibah, Sami Rashid Mohammed <p>Process safety in high-consequence industries such as petrochemicals and hydrogen production faces significant challenges from complex, dynamic systems where traditional reactive safety approaches often fail to predict emerging risks. This paper presents a novel framework for integrating Responsible Artificial Intelligence (RAI) and physics-informed modeling to transition from reactive monitoring to proactive risk mitigation. Our core methodological contribution introduces a hybrid AI architecture based on Physics-Informed Neural Networks (PINNs) coupled with Bayesian Uncertainty Quantification (UQ). This framework is specifically designed for predictive risk assessment in inherently complex systems. We demonstrate its application through a high-fidelity quantitative simulation of a hydrogen production unit, focusing on predicting catalyst degradation and potential runaway reactions. The methodology rigorously integrates first principles (reaction kinetics and heat transfer equations) with operational data to generate predictions characterized by quantified uncertainty (epistemic and aleatoric). Quantitative results, derived from a fully reproducible Python simulation, demonstrate a significant improvement in anomaly detection sensitivity (ROC AUC of 0.98 compared to 0.84) and a quantifiable reduction in detection latency (4 hours compared to 18 hours) compared to traditional Statistical Process Control (SPC) methods. This reduction in false positive rates is crucial for preventing operator fatigue in safety-critical applications. By embedding physical laws, the model offers verifiable trustworthiness and enhanced explainability for human-in-the-loop decision-making. The framework is directly applicable to key petrochemical processes including fluid catalytic cracking (FCC) units for coke deposition monitoring and Carbon Capture, Utilization, and Storage (CCUS) systems for amine solvent degradation and CO₂ injection risk assessment.</p> |
| title | Proactive Risk Mitigation and Asset Integrity using Physics-Informed AI: A Framework for Hydrogen Production and Petrochemical Applications |
| url | https://doi.org/10.5281/zenodo.19563668 |