Why PSM Has to Start at FEED — Not Six Months Before Startup

There's a version of process safety management that starts when the facility is nearly complete. The PSM coordinator gets brought in six months before startup. They find that the process safety information is incomplete, the PHA hasn't been done, the operating procedures don't exist, and the PSSR checklist is being drafted by the project team as a close-out task. Then they spend the next six months in crisis mode trying to assemble the compliance package before chemical introduction.

This happens on more capital projects than it should. And it's expensive — not just in cost, but in the quality of the safety program that results.

What Gets Decided During FEED That Affects Process Safety

Front-End Engineering Design is where the fundamental configuration of a facility is established. Not the details — the fundamentals. And fundamentals, by definition, are hard to change later:

Plant layout and separation distances. The spacing between process units, the location of control buildings, the direction of prevailing winds relative to relief device discharge points, the proximity of ignition sources to potential release locations — these are set during FEED. A consequence modeling exercise at FEED takes two weeks and costs almost nothing relative to project scope. The same analysis at construction, when layout is fixed, tells you what you're living with for the next 30 years.

Relief system philosophy. The decision to use a closed flare system vs. open vent, the sizing basis for the main relief header, the location of the flare knockout drum — these are FEED decisions. Changing them at detailed engineering means redesigning the entire relief network. Changing them during construction means physical rework. These decisions need process safety input at FEED.

Control philosophy and SIS architecture. The allocation of safety functions between the basic process control system and the safety instrumented system is fundamentally a FEED-phase decision. Once the control narrative and the SIS design basis are fixed, every downstream decision about sensors, logic, and final elements flows from them. A SIL rationalization at FEED shapes the entire instrumented safety architecture. Done for the first time at detailed engineering, it's often a rework exercise.

Utilities and emergency systems. Emergency cooling water, firewater ring mains, deluge systems, emergency power for safety-critical loads — their sizing and routing are FEED work. Undersized emergency water systems get built into facilities because the consequence modeling that would have informed the decision wasn't done until after the utilities were engineered.

What PSM Integration at FEED Actually Looks Like

FEED-phase process safety integration isn't a PHA. The process design isn't mature enough for a full HAZOP at FEED. What it is:

Inherent Safety Review

A structured examination of whether the process conditions, chemistry, and material selections are as inherently safe as they can be while still meeting the process objective. Minimize — reduce hazardous inventories. Substitute — are there less hazardous alternatives? Moderate — can operating conditions reduce hazard severity? This is a FEED-phase conversation, not a detailed engineering conversation.

Consequence Screening

Before layout is fixed, model the consequence boundaries for credible worst-case releases. Do the separation distances accommodate those boundaries? Are occupied buildings outside the consequence zones for realistic scenarios? A screening-level consequence model at FEED takes hours. Moving a control building after construction takes months and costs millions.

Regulatory Applicability Determination

Does this facility trigger OSHA PSM? Does it trigger EPA RMP? Are there state-level requirements that apply? What are the specific element requirements, and when do they need to be satisfied? These questions need answers at FEED, not at commissioning. The element requirements have lead times — PHA facilitation, procedure development, operator training — that need to be planned from the start of the project.

PSM Program Development Scoping

What PSM elements need to be built from scratch vs. adapted from existing corporate standards? What's the staffing model for PSM during operations? These questions affect facility design — the number and location of lockout/tagout stations, the configuration of the permit-to-work system, the design of the control room for operating crew requirements.

The Cost of Starting Late

The cost of late PSM integration shows up in specific, predictable ways. PHA recommendations that require design changes at 60% IFD. Relief system redesigns that delay IFC. PSSR punch lists with hundreds of items that force commissioning delays. PSM compliance gaps discovered during pre-startup audits that require written corrective action plans and regulatory notifications.

None of these outcomes are catastrophic in isolation. Cumulatively, they're the difference between a project that runs on schedule and one that doesn't — and between a facility that starts with a functioning PSM program and one that spends its first operating year trying to build one under production pressure.