A shutdown or turnaround (STO) is a planned, full or partial stop of a plant to do the maintenance, inspection, and capital work that cannot happen while equipment is running. It is scoped months ahead, run against a fixed milestone schedule, and judged on whether it restarts on time, on budget, and safe.
Every plant lives on one of these events. A vessel that has to be opened for internal inspection, a main drive that can only be replaced with the line cold, a code-required boiler tear-down, none of it happens on a normal maintenance day. So the plant picks a window, stops, and does a year of postponed work in a few compressed days. The difference between a turnaround that ends on schedule and one that runs a week long is almost never the wrench work. It is the planning that happened before anyone picked up a wrench.
What is a plant shutdown or turnaround?
A turnaround is a scheduled, plant-wide (or unit-wide) maintenance event where equipment is taken offline so work that requires a cold, de-energized, or opened asset can be completed. The industry groups these as STO events, shutdowns, turnarounds, and outages, and treats them as their own discipline because the cost of a lost day is measured in lost production, not just labor.
The work falls into four buckets: inspection and testing that regulations or codes require on a fixed cycle; replacement of components that are worn but not yet failed; capital or improvement projects that need the asset stopped; and the deferred corrective work that has been piling up in the maintenance backlog waiting for a window. A turnaround is where a proactive plant pays down the reliability debt it has been tracking all year.
Why do turnarounds run over budget and schedule?
They run over because scope grows after it should have been frozen, and because work that should have been packaged in advance gets discovered in the field. Both are planning failures, not execution failures. When the schedule is built from a worklist that keeps changing, every added job ripples through crew loading, parts, and the critical path.
Scope creep is the single most cited cause. Late additions arrive with no pre-staged parts, no procedure, and no slot on the schedule, so they either extend the outage or push out other work. The second killer is discovery, opening a vessel and finding twice the corrosion you planned for. You cannot eliminate discovery, but you can shrink it with good pre-shutdown predictive maintenance and inspection data so fewer surprises wait inside the equipment.
How do you scope a turnaround?
You build one master worklist, challenge every item against a clear justification, and then lock it. Scoping is the act of deciding what goes in the window and, just as importantly, what stays out. A job earns its place only if it truly needs the asset stopped and cannot wait for the next cycle.
Each worklist item should carry a reason code: regulatory or code-required, condition-based (driven by inspection or condition monitoring data), reliability-critical, or improvement. Items without a code are candidates to defer. This is also where equipment reliability history earns its keep, the assets with the worst failure records get the deepest attention, and the ones running clean get a lighter touch. Rank the list by criticality so that if the window tightens, you cut from the bottom, not from whatever happens to be unfinished.
How far ahead should you plan a turnaround?
Big turnarounds need a runway of roughly a year to eighteen months for scope development, with detailed planning starting no later than about 90 days out and scope frozen around 60 days out. Small or single-unit outages compress this, but the sequence is the same: develop, plan, freeze, execute. Starting the detailed work at 30 days or less is the reliable way to run long and over budget.
What are the steps to run a turnaround?
A turnaround runs on a repeatable sequence. Here is the framework, from the first worklist to the post-event review:
- Charter and initiate. Name a turnaround manager, set the window dates, fix the budget envelope, and define what "success" means in numbers, restart date, cost, safety, and quality of restart. Everything downstream references this charter.
- Develop and challenge scope. Build the master worklist, attach a reason code and a criticality rank to every item, and challenge each one. Kill or defer anything that does not need the asset down.
- Freeze the scope. Lock the worklist at the milestone. After the freeze, new work goes through a formal change process with a cost and schedule impact, no quiet add-ons. This single discipline saves more turnarounds than any other.
- Plan and package the work. Turn every job into a work package: procedure, permits, parts, tools, labor hours, and safety plan. Stage the parts. This is maintenance planning and scheduling at high intensity, and it is where wrench time is won or lost.
- Order long-lead materials early. Anything with a lead time longer than the planning horizon gets its own delivery milestone on the schedule. A missing forging or specialty valve stops a critical path cold, the same logic behind good spare parts optimization.
- Build the integrated schedule. Sequence packages by the critical path, load crews so trades are not stacked on top of each other, and identify the handful of jobs that actually gate the restart date. Everything else has float; the critical path does not.
- Execute with daily control. Run the outage against the schedule with daily progress reviews, a live punch list, and one source of truth for status. When jobs live in spreadsheets and radios, visibility collapses within a shift.
- Start up and review. Commission in a controlled sequence, close permits, and hold a cold-eyes review while memories are fresh. What ran long, what surprised you, what to add to next cycle's scope, capture it before the crew scatters.
| Milestone | Typical timing | What must be true |
|---|---|---|
| Scope development | T-12 to T-18 months | Worklist seeded from backlog, inspections, and reliability history |
| Detailed planning starts | T-90 days | Every item has a reason code and a criticality rank |
| Long-lead materials ordered | T-75 days | Delivery milestones on the schedule for every long-lead item |
| Scope freeze | T-60 days | Worklist locked; changes go through formal review |
| Integrated schedule complete | T-30 days | Critical path known; crews loaded; packages staged |
| Execution window | T-0 | Daily control, live punch list, single source of status |
Turnaround planning: the reference numbers
Anchors for scoping and defending a turnaround plan. Treat the ranges as directional industry practice, not a single study:
- 12–18 months of scope development for a major turnaround, with detailed planning starting by roughly T-90 days and scope frozen near T-60 days the widely used process-industry planning horizon.
- 30–40%+ savings opportunity exists for plants that shift work from reactive to planned, per the U.S. Department of Energy Federal Energy Management Program (PNNL O&M Best Practices). Turnarounds are how you clear the reliability debt that reactive work creates.
- Scope discovered in the field is the most expensive kind, because it arrives without staged parts or a schedule slot, the argument for freezing scope early and inspecting hard before the window.
How do you keep an outage under control while it runs?
You run it against one live source of truth, review progress every day, and manage the punch list as ruthlessly as the schedule. The plan only survives contact with the field if everyone is looking at the same status. When a discovered job appears, it goes through change control with a critical-path impact before it gets a crew, not as a favor squeezed in between other work.
This is where a connected work-management layer earns its keep. When work packages, parts reservations, permits, and completion status all live in one system, a supervisor can see the critical path move in real time instead of reconstructing it from radio calls. Harmony's work and inventory intelligence is built to spot exactly those gaps across systems (see the platform), so a missing part or a slipping critical-path job surfaces as an alert instead of a surprise at shift change. No rip-and-replace of the CMMS you already run, see how a connected plant handles it in the CLS case study.
Where does turnaround planning fit in the reliability picture?
A turnaround is the concentrated expression of a plant's whole maintenance program. The CMMS holds the backlog and job history that seed the worklist; predictive and condition-based inspection data decide what actually needs opening; and disciplined spare parts inventory management makes sure the parts are on the shelf when the window opens. Plants with strong day-to-day reliability run shorter, calmer turnarounds because less has been deferred and fewer surprises are waiting inside the equipment. Even lean-team plants can run a tight outage, see maintenance for small plants for how.