A shutdown, turnaround or outage is a planned period in which a plant is taken partly or fully out of operation for maintenance, inspections, statutory examinations and modifications that cannot be done while it runs. Shutdown and turnaround planning starts months ahead: collecting the work and freezing the scope, ordering parts with long lead times, booking contractors and scaffolding, and sequencing the work so the plant starts up again safely and on time. Every extra day costs production, and whether the stop stays within its planned days is largely decided during that preparation.
What a turnaround is, and the main types of shutdown
What sets a turnaround apart from routine maintenance is the volume of work packed into a short, fixed window, with the plant out of service and often with many outside specialists on site.
| Type of stop | Characteristics | Example |
|---|---|---|
| Major turnaround | Whole plant, weeks, fixed cycle | Boiler inspection and statutory re-examination |
| Short stop | One line or system, days | Cleaning a heat exchanger |
| Inspection-driven stop | An examination needs the plant down | Internal inspection of a pressure vessel |
| Unplanned stop | A failure forces the plant down | Leaking boiler tube |
Statutory inspections often set the rhythm of shutdowns; see statutory inspections. Even an unplanned stop can be put to use if you keep a list of ready-to-execute work that can only be done with the plant down.
Every day counts: the plant produces nothing while contractors, scaffolding, cranes and hired equipment are paid by the day. In a waste-to-energy plant, an extra day means waste that is not processed and power and heat that are not delivered; see waste-to-energy plant maintenance.
The phases of a turnaround over time
A turnaround is planned backwards from the first day of the stop, T-0; T-6 means six months before it and T+2 two months after start-up. The figure below shows the phases of a major three-week turnaround. The timings are indicative: for a short stop of a few days, months become weeks, and for very large turnarounds you start earlier.
- Initiation and strategy (T-12 to T-10). Set the objectives: what the stop must deliver, the number of days, the budget and the safety goals. Agree the date with production and sales, appoint a turnaround manager and core team, and bring out the lessons from the previous stop.
- Build and challenge the scope (T-11 to T-6). Collect all the work for the stop, challenge it and close with the scope freeze at T-6.
- Materials and contractors (T-9 to the start of the stop). Order long-lead items from T-9, as soon as it is clear you need them. Contract specialists, scaffolders and crane hire, and arrange storage and site cabins.
- Work preparation and scheduling (T-6 to the start of the stop). Turn each job into a work pack (see work preparation in maintenance) and build the turnaround schedule from those packs: critical path, permit and isolation plan, scaffolding plan and shift roster.
- Execution (the stop itself, three weeks). Shut down and make safe, inspect, repair, reassemble and test.
- Start-up and handover (end of the stop). Finish the punch list, run the pre-start-up check, start up in a controlled way and hand over to production.
- Review and close-out (up to T+2). Review the stop, update the history, close the work orders and record the lessons, up to two months after start-up.
Scope: which work belongs in the turnaround?
Work for the stop comes from many directions:
- inspection and measurement history, such as boiler tube wall thicknesses heading towards the minimum (see boiler tube thickness measurement);
- statutory inspections and re-examinations;
- preventive tasks that can only be done with the plant down;
- corrective work in the backlog with the status "waiting for shutdown";
- risk analyses such as an FMECA, and findings from failure investigations;
- projects, modifications and improvements from the previous stop.
Scope challenge: three questions per job
Test every job put forward with three questions. Does it have to be done during the stop, or can it be done while the plant runs? Does it have to be done in this stop, or can it wait until the next one given its condition and risk? And can it be prepared in time, with materials, drawings and someone to do it? Work that does not make the cut goes back to the regular maintenance schedule or to a later stop.
Work that is only needed if an inspection finds something goes in as conditional work: prepared, with materials ready where necessary, but carried out only once the finding is confirmed.
Scope freeze and late additions
After the scope freeze, work can only be added through a formal change procedure: a request with reason and risk, the extra hours and materials, and the effect on the critical path. The turnaround manager decides, together with production. A late job without materials or a work pack is often where an overrun starts.
Also reserve hours and budget for discovered work (emergent work): damage that only becomes visible once the plant is open. Base the size on your previous stops. If, for example, twenty hours of discovered work came on top of every hundred planned hours, build that allowance into the plan.
Long-lead items, contractors and logistics
Long-lead items are parts with lead times measured in months: heat exchanger bundles, boiler tubes and bends in special alloys, large valves, grate bars and refractory materials. Do not wait for the scope freeze; order as soon as you know you need them. The route from request to receipt is covered in purchasing spare parts. Plan incoming inspection and storage as well, and kit the materials per work pack before the stop.
Contractors should be booked early, because specialist trades are scarce when many plants stop at the same time. Agree scope, quality requirements, reporting and site induction beforehand, and name contact persons on both sides. EU-OSHA's basic rules for maintenance work point out that communication with contractor workers can be difficult in terms of time, language and organisation. More on managing contractors and following up their reports is in managing maintenance contractors.
Scaffolding is often underestimated. Many jobs cannot start until a scaffold is up and inspected, so scaffolding easily ends up on the critical path. Plan per work pack when each scaffold goes up, is inspected and can come down.
Logistics covers cranes, laydown and work areas, access control and routes, but also where people spend their time. In the 2005 explosion at the Texas City refinery, cabins for contractors working on turnarounds of neighbouring units stood right next to a unit that was being started up. According to the investigation report of the US Chemical Safety Board (CSB), all fifteen fatalities occurred in or around those cabins. Place cabins, workshops and assembly points outside the hazard zones of units that stay in operation or are being started up.
Planning a turnaround: sequence, critical path and hold points
The sequence of a turnaround is largely fixed:
- Shut down and make safe: take the plant out of service, drain, flush or gas-free, isolate and blind.
- Scaffolding and access: erect and inspect scaffolds, open manholes, ventilate.
- Inspect first: as early as possible, so discovered work surfaces while there is still time to arrange materials and people.
- Repair and replace: the planned work plus the discovered work that has been approved.
- Close up and reassemble: gaskets and bolting to specification, blinds out, insulation back on.
- Test and release: pressure and leak tests, trips and instruments proven, hold points signed off.
- Start up and hand over: only once everything has been ticked off.
Steps 1 and 5 mirror each other. The UK HSE's technical guidance on maintenance procedures lists failing to drain or isolate plant before dismantling among the possible sources of major accidents, along with blinds (spades) left in or removed, gaskets left out and bolts torqued incorrectly. A blind list, on which every blind fitted and removed again is signed off, is the simplest safeguard.
The critical path is the longest chain of activities that wait on each other; that chain determines how long the stop lasts. In a boiler it might be scaffolding in the furnace, inspection, tube replacement, pressure test and refractory repair. Whatever lies on the critical path gets the best preparation, first pick of people and a daily progress check. Keep an eye on the near-critical paths too.
Hold points are moments where work stops until an inspection or approval has taken place, for example before a vessel is closed or after welds that need examination. Schedule them as activities with a time and a responsible person, so the inspector is there at that moment.
Permits and isolations are open in far greater numbers during a stop than in normal operation. Prepare isolation packages per system using the isolation points from the P&ID, and set up a single point where permits are requested and issued. How permits to work, LOTO and last-minute risk assessments fit together is covered in safe maintenance.
Shifts. If you work long shifts, watch for fatigue. At Texas City, the CSB concluded that operators were likely fatigued after working 12-hour shifts for 29 or more consecutive days. Build in rest days and keep an eye on overtime, especially in the final week; what Dutch working time law allows in 60-hour weeks is covered in recording overtime.
During the stop: daily rhythm, progress and discovered work
A turnaround runs on a fixed daily rhythm: a short start-of-shift briefing per crew and a daily coordination meeting with turnaround management, production, safety and the main contractors. Safety first, then progress, bottlenecks, decisions on discovered work and the look-ahead for the next 24 to 48 hours. That is also where you coordinate work that interacts, such as hot work next to gas-freeing or lifting over other crews; many people on a turnaround do not know the plant.
Tracking progress with an S-curve
Measure progress in earned hours: the planned hours of the work that has been completed, with partly finished jobs counted pro rata. Hours spent are not progress. Plot cumulative earned hours as a percentage of the total against the plan; the result is an S-curve, because the first days (shutting down, scaffolding) and the last (testing, start-up) contain fewer work hours than the middle.
The figure below shows a 21-day stop. On day 12, the plan says 65 per cent of the hours should have been earned; the actual figure is 58 per cent.
Seven percentage points behind, just past halfway: that calls for a decision today. If the gap is on the critical path, add people or an extra shift, or split the work; if it is on work with float, rescheduling is often enough. You can see how such a progress curve can be kept per turnaround under maintenance planning software.
Deciding on discovered work
Discovered work comes in every day. Decide per finding: do it now, defer it until after the stop or to the next stop with a recorded risk assessment, or apply a temporary measure. Book executed discovered work against the reserve and keep track of how much is left.
Start-up, close-out and review
Punch list and pre-start-up check
Keep a punch list from day one, marking for each item whether it must be done before start-up or can wait until after. Before start-up, walk through a fixed check with production: all permits closed, isolations and blinds removed according to the lists, tests completed, trips and instruments proven, the area cleaned up and everyone out of the zones being started up.
Start-up is a high-risk phase
Plan the start-up as an activity with its own time, not as whatever is left at the end. At Texas City, the day before the explosion, a supervisor told instrument technicians to stop checking critical alarms because the unit was starting up and there was not enough time, according to the same CSB report. At a petrochemical plant in Moerdijk, the Netherlands, two major explosions occurred in 2014 while reactors were being warmed up after a short planned maintenance stop in which the catalyst had been replaced. The Dutch Safety Board described start-up as a common safety-critical process and found that modifications to the plant and procedures had not always led to a new risk analysis; when the plant switched to another type of catalyst, its reaction with the warming-up liquid had not been retested either. Contractors were also still working close to the reactor.
The lesson for planning: every change made during the stop gets its own risk assessment before start-up, and the start-up instructions must match the plant as it now is.
Review and record
Review the stop within a few weeks: duration and hours against plan, cost, safety, discovered work and late additions. Update the history with findings, measurements and replaced parts, adjust maintenance plans and inspection intervals, close the work orders and return surplus materials. The lessons go into the initiation phase of the next turnaround.
Common turnaround mistakes
- Scope keeps growing after the freeze. Every late job without preparation raises the chance of an overrun.
- Too little inspection data beforehand. If you only see the condition once the plant is open, you are planning on hope.
- Long-lead items ordered too late. A bundle with a six-month lead time cannot be sorted out in three months.
- Scaffolding underestimated. Jobs stand still because the scaffold is not up or not inspected.
- No reserve for discovered work. Every finding then comes at the expense of planned work or the end date.
- Start-up as an afterthought. The checks that get dropped because time has run out are exactly the checks you needed.
Frequently asked questions
What is the difference between a shutdown, a turnaround and an outage?
The terms are often used interchangeably. Shutdown is the general term for taking a plant out of operation. Turnaround comes from refining, among other industries: shutting a unit down after a normal run for maintenance and repair, then putting it back into operation. Power stations usually speak of an outage. In Dutch the common word is revisiestop. Whatever the name, the approach of preparing, executing and closing out is the same.
How far in advance should turnaround planning start?
That depends on the size of the stop. For a major turnaround lasting several weeks, preparation often starts a year or more ahead, leaving time for scope, long-lead items and contracts. A short stop of a few days can be prepared in weeks. The deciding factor is the longest lead time in your scope: start early enough to order, receive and inspect that part before the stop begins.
Who leads a shutdown or turnaround?
Usually a turnaround manager with a dedicated team: planners and schedulers, someone for safety, purchasing and stores, and a production representative who releases the plant and takes it back. What matters is one person with final responsibility and a mandate over scope, schedule and budget, including decisions on late additions and discovered work. For major turnarounds, that team starts work in the initiation phase.
How much contingency should you plan for discovered work?
Use your own history. Look in the reviews of previous stops at how many hours of discovered work came up per planned hour, and build that allowance into hours, budget and, where needed, time on the critical path. If you lack that data, start with a cautious estimate and record during the stop what actually comes up. Scheduling inspections early in the stop helps you spot discovered work in time.
How often does a plant need a turnaround?
That depends on the plant and the requirements around it: statutory re-examinations, wear on parts that can only be reached with the plant down, the life of catalysts or refractory linings, and agreements on availability. In waste-to-energy plants, one major stop per line every twelve to eighteen months is common. The longer the run between two stops, the bigger and heavier the stop usually becomes.