2.2 IIAR 6 Inspection, Testing, and Maintenance

IIAR 6 Is How the Plant Stays Legal After Startup

Quick Answer: ANSI/IIAR 6 is the standard for inspection, testing, and maintenance of closed-circuit ammonia refrigeration systems. ANSI/IIAR 6-2025 was ANSI-approved on 7 July 2025. Treat current IIAR 6 as the ITM RAGAGEP. Design a new engine room to IIAR 2. Write procedures to IIAR 7. Evaluate an old system’s minimum safety against IIAR 9. Keep the hardware working with IIAR 6.

A machinery room can be designed to IIAR 2-2021 on paper and still be a PSM citation if detectors are past due, relief valves are painted shut, or the emergency fan has not been run since the last shutdown. Inspection, testing, and maintenance (ITM) is the ongoing proof that the designed safeguards still function. That is IIAR 6’s job.

IIAR describes ANSI/IIAR 6 as the standard providing minimum requirements for ITM tasks and record-keeping applicable to closed-circuit ammonia refrigeration systems. The 2025 revision was written to harmonize with IIAR 2, IIAR 4, IIAR 5, and IIAR 7. IIAR also notes that the old Bulletin 109 inspection list, which had lived in an appendix of the prior edition, was removed from the 2025 revision. For CIRO candidates that history has one practical meaning: do not treat IIAR Bulletin 109 as the current ITM document. Bulletin 109 still appears in older RAGAGEP discussions (especially relief-valve culture). The living ITM standard is current IIAR 6.

RAGAGEP and mechanical integrity

OSHA’s Process Safety Management standard (29 CFR 1910.119) requires a mechanical integrity program for process equipment in a covered process. Anhydrous ammonia’s Appendix A threshold quantity is 10,000 lb, so most industrial ammonia plants are covered processes. OSHA expects inspection and testing to follow recognized and generally accepted good engineering practices (RAGAGEP).

For ammonia refrigeration, IIAR 6 is the RAGAGEP most plants write into the MI element. A CIRO supervisor who says “we inspect whatever the mechanic has time for” has no RAGAGEP. A supervisor who says “we follow current IIAR 6, the manufacturer’s instructions, and the written MI procedures, and we keep the records” is answering the exam and surviving an OSHA walkthrough.

IIAR 6 does not replace manufacturer instructions. If a detector vendor requires more frequent calibration than the plant’s generic task list, the more demanding applicable requirement wins. IIAR 6 also does not replace IIAR 2. If a detector is missing, that is a design/gap problem (IIAR 2 for new work, IIAR 9 for existing minimums) as well as an ITM problem. ITM cannot inspect a safeguard that was never installed.

What an IIAR 6 program looks like on the floor

You will not be asked to recite every row of every IIAR 6 task table from memory. You will be asked whether a described activity is ITM, who is qualified to do it, and what a failed test implies.

A functioning IIAR 6 program typically includes:

• Written ITM procedures tied to equipment identity (compressor, vessel, valve, detector, PRV, emergency fan), not a sticky note on the MCC.
• Task frequencies from the standard and the manufacturer, with a documented basis if the plant uses a risk-based interval the standard allows.
• Qualified people. “The night operator is handy with a wrench” is not a qualification system. Visual rounds by operators support the program; specialized inspections (relief devices, vessel integrity, some electrical tests) require people trained for those tasks.
• Acceptance criteria and deficiency handling. A test that fails is a work order and an operability decision, not a log entry that says “still leaking, will watch.”
• Records that identify the equipment, the task, the date, the result, the as-found/as-left condition, and who performed the work. PSM auditors live in those records.

Supervisor scenario: Saturday detector failure

A production freezer is down, and the engine-room panel shows one of the two matched-range detectors in fault. IIAR 2 required two detectors with identical sensing ranges. IIAR 6 is how you prove they still work. The CIRO decision is not “run on one sensor until Monday.” It is: follow the ITM and operating procedures for a failed detector (often meaning increased patrols, a temporary administrative control, or a controlled shutdown of the affected room depending on the written program and the AHJ), replace or repair the device, and function-test the 25 ppm notification, the 150 ppm ventilation start, and the high-concentration compressor/pump trip—without using a live 40,000 ppm cloud as the test method. Simulation and manufacturer procedures exist for a reason.

---

Pressure-Relief Valves: The ITM Item That Fails Audits

Relief valves are the last mechanical barrier between a blocked-in vessel and a rupture. IIAR 2 tells the designer how relief is arranged. IIAR 6 tells the plant how that relief stays real.

Follow the manufacturer and current IIAR 6 — do not invent a paragraph number

Current IIAR 6 requires periodic inspection, testing, and replacement of pressure-relief valves in accordance with the standard and the manufacturer. Many ammonia plants replace PRVs on a five-year cycle. That five-year culture is long-established RAGAGEP in IIAR Bulletin No. 109 and No. 110 and in manufacturer literature (replacement from the date of installation is the usual clock). Teach the five-year cycle as common plant practice and historical IIAR guidance, not as a fake “IIAR 6-2025 section 12.4.7” number you cannot open in the book.

Some IIAR 6 editions have also described alternate bench-test programs that can extend service life when a documented sampling plan passes. Those programs are not “skip relief valves for ten years because we are busy.” They are engineered alternatives with sample sizes, pass/fail criteria, and a still-finite maximum life. Unless the exam stem gives you a documented alternate program, the expected CIRO answer is: replace or recertify on the plant’s IIAR 6 / manufacturer interval; five years from installation is the cycle most ammonia plants still run.

Operators do not field-adjust sealed setpoints

A pressure-relief valve is a sealed, set, and named device. The setpoint is on the tag. The spring is not a hand-expansion valve. Operators do not field-adjust sealed PRV setpoints with a wrench on the bonnet because the condenser “looks a little high.” Changing a setpoint is a relief-system design change: manufacturer or authorized assembly, new nameplate or recertification, and usually management of change. If a valve weeps, you isolate it only if a dual assembly lets you keep relief in service, then replace it. You do not “tighten it a bit.”

Dual relief with a three-way valve

IIAR 2 design practice for vessels that cannot be taken out of service uses two relief valves on a three-way dual-stop (three-way) valve so that one valve remains in service while the other is isolated for replacement. The three-way plug cannot blank both valves at once if it is the correct device, correctly piped, and locked/documented in a position that opens one PRV to the vessel.

CIRO implications:

• Never leave the three-way in a mid-position that starves both valves.
• Each valve has its own service clock. Installing a dual assembly does not mean the idle valve can sit for twenty years because “it wasn’t seeing pressure.” Industry commentary on dual assemblies is that both valves are installed devices exposed to weather, vibration, and time; plants still replace or bench-test both on the ITM interval.
• Changeover is a written procedure (IIAR 7 territory) with isolation confirmation. The mechanic who spins the three-way “real quick” during production without documenting the in-service valve has created an unverified relief path.
• Discharge piping, rain caps, and supports are part of the relief system. A perfect PRV that relieves into a crimped, oil-filled, or bird-nested stack has not relieved.

What IIAR 6 is not

IIAR 6 still inspects all of those things after they exist. The exam trick is the primary standard for the decision in the stem.

Supervisor scenario: “Just bump the relief”

A second-shift operator reports a dual-relief assembly hissing at the stack. The condenser pressure is below the nameplate setpoint. A contractor offers to “back off the spring until it seats.” The CIRO answer is no. Isolate to the sister valve on the three-way so one PRV stays in service, replace or shop-test the leaking valve under the IIAR 6 / manufacturer program, log the work, and treat any setpoint or capacity change as MOC plus IIAR 2 design, not as operator craft. If the plant has no dual assembly, the vessel may have to come down to a safe state before the only relief valve is removed—because you do not isolate a vessel from all relief while it is in ammonia service.

That is CIRO supervision: which standard, which valve stays in service, and which wrench never touches a sealed spring.