Safety Margin

Cognascents Consulting Group
  • 5.0 (1)
  • Technology

Safety Margin is Cognascents' podcast for engineers and operators responsible for process safety, asset integrity, and operations across the energy, petrochemical, LNG, and specialty chemical industries. From improving mechanical integrity programs and managing process safety risks to understanding where AI can deliver real value, Safety Margin helps listeners learn how their peers are solving problems, applying best practices, and building safer, more reliable facilities. Learn more at https://cognascents.com/

Episodes

  1. 4d ago

    Safety Margin EP2 - EPA's RMP Compliance Gamble and the AI That Faked an RBI Report

    In Episode 2 of Safety Margin, Cognascent's Justin Daarud (Business Unit Director, Asset Integrity, Reliability & AI), John King (Project Specialist, API 510 Inspector), and founder John Perez dig into the Longview, Washington tank implosion, the EPA's RMP/SCCAP amendment rollback, API RP 754's 4th edition, and what happens when you let a large language model write your risk-based inspection (RBI) report. Timestamps: (0:00) - Cold Open: ChatGPT Invents a Risk Assessment (0:53) - Meet the Hosts + What "Safety Margin" Means (3:40) - Episode Rundown: Tanks, RMP, API 754, AI & Griddles (6:13) - Longview, WA Tank Implosion (2026): The Tank Nobody Had to Inspect (8:54) - Reliability vs. Regulation: Why Inspect If It's Not Required? (11:20) - The Molasses Tank Failure & Asking "What Would Happen If?" (12:58) - Gray Areas: Who Owns the Tank & Who Pays to Inspect It? (15:42) - The RMP Compliance Gamble: 2024 SCCAP Amendments vs. 2026 Proposed Rule (22:35) - Shifting RAGAGEP: When "Should" Becomes "Shall" (24:15) - API RP 754 4th Edition: New Tier 1 & Tier 2 Cost Thresholds (26:33) - Inflation, Incident Costs & Outdated Risk Matrices (29:09) - AI in Asset Integrity: Why Chatbots Hallucinate Reports (31:41) - What an "80% Efficiency Gain" From AI Really Means (34:35) - Agentic Workflows: Training an LLM Like an Entry-Level Engineer (37:06) - Fun Segment: Blackstone Griddle Integrity & Quench Cracking (40:25) - Episode 1 Correction: The Sunflower Oil Mistake (41:04) - King's Huzzah: Name That Album Cover & Wrap-Up (45:04) Wrap up Learn more at https://cognascents.com/

  2. Aug 7

    Safety Margin EP1 - Are Process Safety Incidents Actually Getting Worse? Can AI Help?

    Are process safety incidents actually getting worse, or are we just seeing more of them? The debut episode of Safety Margin opens with a fatal January 2026 H2S incident at a Maine pulp mill, then digs into conflicting data on whether the industry is slipping. Hosts John Perez (founder of Cognascents, chemical engineer), Justin Daarud (asset integrity, reliability & AI director, mechanical engineer), and John King (API 510 pressure vessel inspector) also break down what AI really means for engineering—chatbots vs. deterministic workflows, hallucinations, and the "golden thread" of auditability—before closing with a fun take on medieval armor as the original asset integrity program. Timestamps: (00:00) Cold Open: A Refinery "Graveyard" Story (01:08) Are Process Safety Incidents Getting Worse? (02:46) What "Safety Margin" Means + Meet the Hosts (06:39) Why We Started This Podcast & Safety's Social Contract (12:36) Are Incidents Actually Increasing? The Data (14:00) Philadelphia Refinery Explosion (2019) & Piping (16:20) Data Centers, Energy Demand, Cutting Corners (19:23) EPA & CSB Numbers, Biases & Lost Expertise (24:00) AI in Process Safety: Adopters vs. Skeptics (27:19) How LLMs, RAG & Agentic AI Actually Work (32:37) Comparing AI Models & Their Vulnerabilities (37:55) The "Golden Thread": Auditability & Human Oversight (40:29) Fun Segment: Medieval Armor as Asset Integrity (45:05) Shout-Outs, Rapid Fire, Wrap-Up Learn more at https://cognascents.com/

  • 2 Episodes

About

Safety Margin is Cognascents' podcast for engineers and operators responsible for process safety, asset integrity, and operations across the energy, petrochemical, LNG, and specialty chemical industries. From improving mechanical integrity programs and managing process safety risks to understanding where AI can deliver real value, Safety Margin helps listeners learn how their peers are solving problems, applying best practices, and building safer, more reliable facilities. Learn more at https://cognascents.com/

Information

  • Creator
    Cognascents Consulting Group
  • Years Active
    2026
  • Episodes
    2
  • Rating
    Clean
  • Copyright
    © 2026 Cognascents Consulting Group
  • Show Website
    Safety Margin