Everything you need to understand about commissioning a chemical plant, refinery, or industrial facility — from FEED through steady-state operations. Written by engineers who have done it on billion-dollar projects.
Commissioning is the process of verifying that a newly built or significantly modified facility is ready to operate safely and as designed — and then starting it up. It begins before construction ends and finishes when the facility reaches stable, normal operations.
The word "commissioning" is used in two ways in the industry, and the distinction matters. Broadly, commissioning refers to the entire process from end-of-construction through stable startup — everything it takes to go from a built facility to a running one. More narrowly, "commissioning" refers specifically to the hot phase: the introduction of process chemicals and the initial operation of the unit under live conditions. This guide uses both meanings and is explicit about which one applies when.
What commissioning is not: commissioning is not simply turning things on. It is a disciplined, documented verification process that confirms every system is built to design, every safety device works, every procedure is written, every operator is trained, and every regulatory requirement under 29 CFR 1910.119 is satisfied — before any hazardous material arrives.
More incidents occur during commissioning and startup than during any other equivalent period of facility operation. The reasons are consistent: equipment that has never run at design conditions, operators following procedures for the first time, schedule pressure to "make up time," and unexpected interactions between systems that looked fine on paper. Commissioning is not a phase to rush through. It is a phase to manage.
A capital project commissioning effort involves the owner/operator, the EPC contractor (engineering, procurement, construction), equipment vendors, subcontractors, and — increasingly — a process safety consultant responsible for PSM compliance. Each party has defined handover responsibilities. Understanding who owns what at each stage is the foundation of a well-managed commissioning effort.
The owner/operator is responsible for PSSR and for the regulatory compliance of the facility. That obligation cannot be contracted away — the EPC delivers the facility; the owner operates it and is accountable to OSHA for PSM compliance on day one of startup.
A capital project moves through six phases from initial concept to steady-state operations. PSM compliance work spans all six. The single most common PSM failure on capital projects is treating compliance as a phase-6 activity rather than building it in from phase 1.
Conceptual design is finalized. The project scope is defined. Process flow diagrams and initial P&IDs are developed. This is where the PSM program development plan should be written — defining what compliance work is needed, who will do it, and when it must be complete. The PHA methodology is selected and scoped. Covered process determination is documented. Most projects that run out of time at PSSR can trace the problem back to FEED: the PSM scope was never scoped, scheduled, or resourced.
P&IDs are developed to IFD (Issued for Design) and ultimately IFC (Issued for Construction) status. The HAZOP is conducted — ideally at 90% IFD P&ID maturity. The PSI package is developed in parallel with the engineering deliverables that populate it. Operating procedures are being drafted against the developing P&IDs. MOC for design changes must be in place by this phase — HAZOP deviations corrected in the field without re-review are the most common source of unreviewed hazards on new builds.
The facility is built. Field changes are common — construction rarely matches IFC exactly, and the gap between what engineering designed and what construction built needs to be tracked, reviewed, and incorporated into the as-built P&IDs. Construction-phase MOC must be active and functioning. The procedure library, training program, and PSI package continue to develop. PSSR planning begins during construction — not after mechanical completion. Blinds, spectacle plates, and temporary modifications must be tracked from the day they are installed.
No hazardous materials are present. Systems are mechanically complete. Equipment is cleaned, flushed, tested, and verified system by system. Instrument loops are checked. Safety-critical instruments are tested against their design function. Electrical systems are energized. Blinds and spectacle plates are tracked. Punch lists are managed — A-items must be cleared before chemical introduction authorization, B-items can be cleared within an agreed period after startup. This phase ends when the facility has received mechanical completion sign-off and PSSR has been completed.
Required under 29 CFR 1910.119(i). The PSSR is the regulatory verification that the facility is ready for hazardous material introduction. It confirms: construction matches design, procedures are in place, PHA recommendations are resolved, and employee training is complete. PSSR is a multi-discipline team effort with formal sign-off authority. It is not a rubber stamp — open PSSR items block startup. When PSSR takes a week instead of a day, the problem started six months earlier.
Hazardous materials are introduced. The process is operated for the first time at design conditions. Initial startup is the most hazardous phase of any project — equipment has never run at design conditions, procedures are being followed for the first time, and unexpected conditions are routine. This phase ends when the facility reaches stable, sustained production at design capacity and is formally handed over from the project team to the operations organization.
Mechanical completion is the formal declaration that construction is complete and the facility is ready to begin pre-commissioning activities. It is not a declaration that the facility is ready to start up — it is the starting line for the pre-commissioning process.
Mechanical completion is achieved system by system, not all at once on a single date. As each process system completes construction, it is walked down by a multi-discipline team — process, mechanical, instrumentation, electrical, operations, and safety. Deficiencies found during the walkdown are recorded on the punch list.
The punch list distinguishes between two categories of deficiency, and the distinction is critical:
Projects under startup schedule pressure downgrade A-items to B-items to clear the mechanical completion gate. This is how safety-critical deficiencies end up being "resolved" by reclassification rather than correction. The authorization level for downgrading A-items to B-items must be defined and enforced — it should require sign-off from the operations manager and the process safety lead, at minimum.
Every blind, spectacle plate, and isolation device installed during construction, cleaning, hydrotesting, or pre-commissioning must be tracked from the day it is installed. A blind left in at startup has caused fatalities. The blind register is not optional — it is a life-safety document. It should be controlled, sequentially numbered, and signed off for removal by a qualified person before PSSR authorization. Every blind on the register must be accounted for — removed or accepted — before startup is authorized.
Pre-commissioning is the verification of each system under non-hazardous conditions. It confirms that equipment was installed correctly, instruments function as designed, safety devices respond as intended, and utilities operate reliably — before any hazardous material arrives.
Pre-commissioning activities are organized by system — a defined envelope of equipment and piping that can be reasonably isolated and tested as a unit. For each system, pre-commissioning follows a defined sequence: inspection, cleaning, testing, verification, and sign-off.
New piping contains construction debris — weld slag, scale, dirt, and occasionally tools. Cleaning and flushing removes debris before equipment is exposed to process fluid. The specific cleaning requirement depends on the service: hydrocarbon systems are oil-flushed; water systems are water-flushed; high-purity services (oxygen, hydrogen) require chemical cleaning. Every cleaning activity is documented, including the inspection of flush outlets to confirm acceptable cleanliness.
Hydrostatic testing confirms piping and vessel pressure integrity. Each piping system and vessel is pressurized to 1.5× MAWP (for ASME B31.3 process piping) with a compatible test medium — water for most services, pneumatic alternatives only where water would cause damage or create a hazard. Test pressure, hold time, and results are documented for every test. Relief devices are isolated or removed during hydrostatic testing.
Every instrument loop is checked from the field device to the control system and back. The loop check confirms: transmitter calibration and range, signal integrity at the control system, alarm setpoint response, and for safety-critical functions, shutdown initiation. Safety instrumented system (SIS) functions are tested against their design specification — the same test that must be repeated on the proof test interval during operations.
Loop check documentation for safety-critical instruments is a PSI deliverable. It establishes the baseline for the mechanical integrity program — the initial verification that the safety function was installed and tested. It must be retained and accessible as part of the PSM records.
Electrical systems are energized in a defined sequence. Protection relay testing is completed before energization. Motor rotation is verified for all rotating equipment — wrong rotation direction on a centrifugal pump or compressor can cause immediate equipment damage. Each energization is documented, including the authority who authorized it and the field verification that was completed.
Centrifugal pumps, compressors, fans, and other rotating equipment have specific pre-commissioning requirements: alignment verification, bearing lubrication, seal system preparation, coupling installation, and vibration baseline measurement during initial uncoupled motor run. Positive displacement machines have their own requirements — gear inspection, discharge relief verification, and no-flow protection confirmation.
Pre-commissioning of process systems depends on utility systems being operational first. Cooling water, instrument air, nitrogen, steam, electrical power, and fire water systems must be commissioned, tested, and available before process system pre-commissioning can begin. Instrument air quality (dew point, oil content, particle size) must meet the specification required for pneumatic instruments and control valves. A single undersized instrument air header has delayed startups by weeks.
PSSR is the regulatory gate between pre-commissioning and chemical introduction. Under 29 CFR 1910.119(i), it is required for all new PSM-covered facilities and for modified facilities where the modification is significant enough to require a change in the process safety information.
"The employer shall perform a pre-startup safety review for new facilities and for modified facilities when the modification is significant enough to require a change in the process safety information."
The PSSR must confirm: (1) construction and equipment is in accordance with design specifications; (2) safety, operating, maintenance, and emergency procedures are in place and are adequate; (3) for new facilities, a process hazard analysis has been performed and recommendations have been resolved or implemented before startup; and (4) training of each employee involved in operating a process has been completed.
PSSR is not a punch list walkdown. It is a formal, multi-discipline verification of PSM program readiness — the confirmation that all 14 elements of 29 CFR 1910.119 are in place for this facility, not just that construction is complete. The four regulatory requirements stated in §1910.119(i) span six PSM elements: Process Safety Information (d), Process Hazard Analysis (e), Operating Procedures (f), Training (g), Mechanical Integrity (j) — at minimum — and Emergency Planning (n).
The PSSR team should include — at minimum — a PSSR team leader, a process engineer, a mechanical engineer, an instrumentation/electrical engineer, an operations supervisor, and a process safety representative. For large or complex facilities, additional discipline representation is warranted. The EPC contractor may participate in construction-verification aspects, but the sign-off authority rests with the owner/operator — the entity that will be accountable to OSHA for the facility's compliance.
PSSR is consistently cited as a project startup risk because it is consistently started too late. The PSSR checklist, team composition, and scope should be defined at FEED — not at mechanical completion. Procedures and training take months; they cannot be compressed into two weeks. A project that begins PSSR preparation at mechanical completion is not going to have a clean PSSR sign-off.
When PSSR finds deficiencies late in the project, the pressure to authorize startup anyway — with "open item lists" instead of closed items — is intense. OSHA's PSSR requirement has no provision for conditional authorization. Either the PSSR is complete or it is not. Incomplete PSSR is a §1910.119(i) violation. Startup without completed PSSR is a willful violation if an incident occurs.
Hot commissioning begins the moment hazardous process material enters the boundary of the facility. It is the phase with the highest concentration of incident risk on any project. Everything before this phase was preparation — this is the phase that preparation was for.
Chemical introduction must be formally authorized after PSSR sign-off is complete. Authorization is not implied by PSSR completion — it is a separate, documented act. The authorization process confirms: the PSSR sign-off package is complete, all A-items on the punch list are cleared, the blind register is reconciled, emergency systems are operational, communications are tested, key personnel are on-site, and the first chemical introduction sequence is understood by all involved parties.
The sequence for introducing the first hazardous material matters enormously. Purging and inerting before hydrocarbon introduction, for example, must be verified complete — not assumed. Initial fill quantities may be limited to confirm system integrity before full inventory is introduced. The chemical introduction procedure is the document that governs this sequence, and it must be followed step by step with sign-offs at each hold point.
Introducing a process to its design temperature and pressure for the first time reveals conditions that no amount of design review or pre-commissioning can fully predict: thermal expansion that exceeds expectations, mechanical seals that leak at temperature, instruments that drift under process conditions, and process chemistry that behaves differently at scale than in the lab. The hot commissioning team must have the authority and the standing orders to stop the process immediately if anything deviates from expected behavior — without being overridden by schedule pressure from above.
Hot commissioning rarely happens in a clean, single-sequence manner. Construction punch-out continues in some areas while others are live. Testing, sampling, and commissioning activities overlap. Contractors and operators are working in adjacent areas simultaneously. SIMOPS management — formal coordination of concurrent activities in the same area or on connected systems — is essential during hot commissioning. Each day of hot commissioning activity should begin with a SIMOPS review that identifies who is working where, what systems are live, and what coordination is required.
Hot commissioning always produces deviations from the design intent — process conditions that don't match predictions, equipment that requires adjustment, procedures that need revision. Every one of these deviations is a change that must go through MOC if it involves a modification to the process, the equipment, or the procedures. "We'll fix it in the procedure later" is how uncommissioned process hazards end up in permanent operations. The commissioning team must have a streamlined MOC process that can turn around a simple change in hours — not days.
During hot commissioning, the default answer to any deviation from the planned sequence is "stop and review" — not "keep going and document it." Momentum is the enemy of judgment during this phase. The team must be empowered to stop, and they must know that stopping will not be treated as a failure.
Startup is the ramp from initial chemical introduction to sustained design-rate production. Transition to operations is the handover from the project team — which built and commissioned the facility — to the operations organization that will run it.
The startup window — the period from first chemical introduction to steady-state operation — has a risk profile unlike any other operating period. Equipment has no operational history. Operators have procedure knowledge but no muscle memory. Unexpected equipment behavior is normal, not exceptional. Incident data consistently shows this period has a higher incident rate per operating hour than any other phase of a facility's life.
Managing the startup window means accepting that the process will not behave exactly as expected, building extra margin into the initial operating targets (temperature, pressure, production rate), maintaining a safety professional on-site during initial operations, and treating every deviation as a learning event rather than a schedule threat.
Production rate during the startup window should be ramped progressively — not pushed directly to design capacity. Initial operations at 50% design rate confirm basic process stability before full throughput is attempted. Each rate increase should be planned, authorized, and monitored. Equipment vibration, temperature profiles, pressure balances, and utility consumption at each operating point should be compared to the design predictions. Unexpected deviations at 50% rate are far safer to manage than at 100%.
The formal handover from the project team to the operations organization transfers responsibility for the PSM program, the document library, and the operational readiness of the facility. The handover package must confirm that the PSI is final and current (as-built), the procedure library is complete, the training program is in place for ongoing training and qualification, the mechanical integrity program is activated, and the PSSR records are retained. This is also when the MOC program transitions from construction-phase to operations-phase administration.
Every startup produces information that should be captured and acted on — deviations from the design, unexpected equipment behavior, procedure steps that required field modification, communication gaps, and near-misses. A structured post-startup debrief, conducted within two weeks of reaching stable operations, captures this information before it disperses with the project team. The lessons from the startup become the baseline for the operations organization and feed back into the PSM documentation.
Every element of 29 CFR 1910.119 has work that must begin well before PSSR — many begin at FEED. The following table maps key PSM deliverables to the project phase in which they must be initiated, not completed, to make PSSR on schedule.
| PSM Element | Initiate At | Complete By | Common Timing Failure |
|---|---|---|---|
| Process Safety Information §(d) | FEED | PSSR | PSI document register not started until construction; as-built P&IDs not updated before PSSR |
| Process Hazard Analysis §(e) | FEED (scope); DE (execution) | Before IFC P&IDs finalized | HAZOP started too late on immature P&IDs; action items not closed before startup |
| Operating Procedures §(f) | Detailed Engineering | PSSR | Drafting begins at mechanical completion; 60+ procedures can't be written in 4 weeks |
| Training §(g) | 3–6 months before startup | PSSR | Training planned after procedures are finalized; procedures aren't finalized until PSSR week |
| Contractors §(h) | Pre-award (construction) | Ongoing through commissioning | Contractor safety qualification not documented; hot commissioning contractors not trained on process hazards |
| PSSR §(i) | Planning at FEED; execution at pre-commissioning | Before chemical introduction | PSSR checklist written the week before startup; team not assembled until the last month |
| Mechanical Integrity §(j) | Detailed Engineering (baseline documentation) | PSSR | MI program not defined until operations; baseline inspection records don't exist |
| MOC §(l) | FEED (design changes); Construction (field changes) | Active throughout | MOC not implemented until operations; HAZOP deviations corrected in field without re-review |
| Emergency Planning §(n) | 6 months before startup | PSSR | Emergency response plan written for a hypothetical facility; not updated to reflect actual layout and inventories |
If startup is 18 months away, the PSM program development plan should already be written. If startup is 12 months away, procedures should be in draft. If startup is 6 months away, training should be underway. If startup is 3 months away, PSSR preparation should be well in hand. If startup is 6 weeks away and none of the above has happened — call us.
These are not hypothetical risks. They are the specific failures we see on capital projects repeatedly — different facilities, same patterns. Understanding them is the first step to preventing them.
The most fundamental failure. The PSM program is assigned to a compliance team that starts work six months before startup, discovers that procedures, training, and PSSR can't be completed in that timeframe, and creates a startup delay measured in months. PSM compliance on a new build is an 18–24 month workstream, not a 6-month sprint. It must be resourced and scheduled from FEED.
The HAZOP is rushed to meet a project schedule milestone before the P&IDs are mature enough to support a meaningful review. The result is a HAZOP that misses hazards because the design was still changing, generates action items that are immediately obsolete, and requires a re-study that the project schedule cannot accommodate. HAZOP should be conducted at 90% IFD P&ID maturity — not before.
Engineering changes made after the HAZOP is complete — to address cost overruns, constructability issues, or vendor deviations — are the most common source of unreviewed process hazards on new builds. If the change is not replacement-in-kind, it must go through MOC and may require a HAZOP re-review of the affected node. Many projects do not have functioning MOC during detailed engineering, which means this failure happens invisibly.
A blind or spectacle plate installed during construction, hydrotesting, or flushing and never removed before chemical introduction. The result ranges from a process upset (blocked flow path) to a catastrophic overpressure event (blocked relief path) to a fatality (workers entering a "safe" space that is not isolated from the process). The only control is a numbered, controlled blind register that must be reconciled — every blind accounted for, removed or accepted — before startup authorization.
The punch list A-item count is blocking startup. Rather than fixing the items, the project team reclassifies them as B-items and proceeds. Safety-critical items — non-functional safety instruments, uncalibrated relief devices, incomplete electrical protection — end up in the B-item list with a 30-day post-startup close-out commitment that is never met. The authorization threshold for A-to-B reclassification must be defined in writing and enforced by someone who will not be rewarded for meeting the startup date at the expense of safety.
Operating procedures are drafted during detailed engineering against the IFD P&IDs. Construction deviates from the IFC design. As-built P&IDs are not complete at startup. Operators follow procedures that reference equipment, instruments, and valve positions that don't exist or are in a different location than the procedure states. This is a training deficiency and a PSI deficiency simultaneously — and it is extremely common.
During hot commissioning, construction punch-out continues in areas adjacent to live process. An uncoordinated activity — a contractor opening a flange to fix a punch item on a line adjacent to a pressurized system, a welder working near a live hydrocarbon vent — can produce a serious incident. SIMOPS management during hot commissioning requires a daily coordination meeting, a live work permit system, and a clearly defined authority to stop all work in an area if a conflict is identified.
An operator or process safety professional identifies a deviation during hot commissioning and recommends stopping to investigate. A project manager or executive overrides the recommendation in the interest of schedule. This is the causal factor in a substantial proportion of commissioning and startup incidents. The authority and obligation to stop operations during hot commissioning must be defined, documented, and protected from commercial pressure. It is not a suggestion.
SafeGuard Projects has done this. We can do it alongside your team — from FEED through first startup — so commissioning goes the way it was planned.
Talk to a Commissioning Expert Document Library →