Executive summary
- A scheduled shutdown (turnaround, TAR) is the most expensive concentrated maintenance event in a process plant: 45 to 60 days, 25–150 M€ budget and 200.000–800.000 person-hours, with downtime costs on the order of 1–3 M€/day (illustrative example) at an average Spanish refinery.
- Time study is not applied here as it is on a discrete line: the standard is built for complete trades (scaffolder, pipefitter, seal mechanic, NDT inspector) under Seveso III, API 510/570 and lockout/tagout (LOTO) procedure constraints.
- The VA / NVA / IMP categories of work sampling remain valid, but turnaround VA is different from operating VA: here it is "time that brings the unit closer to a clean start-up", not "time that adds value to the product".
- Three statistical levers that do work: Tippett (1935) to size the sampling of non-productive time during peak workload, MTM-UAS for repetitive scaffolding tasks, and the history of the last 3 shutdowns as a provisional standard while the current one is being measured.
- This article proposes a 7-step template to deploy CronometrasApp as a field tool during the turnaround, compatible with export to Excel/PDF for the shutdown close-out reports.
1. What a turnaround is and why the classic stopwatch is not enough
A turnaround (in Spanish, parada programada) is a total or partial interruption of a process unit — a refinery, a cracking plant, a polyolefins unit — during which tasks are carried out that cannot be performed while the plant is running: regulatory inspections, equipment replacement, mechanical cleaning, process modifications. The industry's operational definition is summed up by the Wikipedia entry on Turnaround (refining):
"A turnaround (TAR) is a scheduled event wherein an entire process unit of an industrial plant, such as a refinery, petrochemical plant, power plant, or paper mill, is taken offstream for an extended period for work to be carried out." — Wikipedia, Turnaround (refining)
Unlike minor maintenance (which is done online or with micro-stops of a few hours), the turnaround concentrates between 60% and 80% of the annual maintenance cost of a unit (illustrative example) into a short window. In Spain, public data from Repsol Tarragona 2024 — 52 days, 150 M€ budget, 600.000 person-hours (Repsol Tarragona, press release 18-Mar-2024) — and Repsol A Coruña 2024 — more than 60 days, ≈25 M€ across eight fuel plants, 200 items of equipment worked on (Repsol A Coruña, Parada de combustibles 2024) — are representative examples of the order of magnitude.
Why the classic stopwatch is not enough:
| Characteristic | Manufacturing line | Refinery turnaround |
|---|---|---|
| Task duration | Seconds to minutes | Hours to days per task |
| Repetition | High (closed cycle) | Low (each task is unique) |
| Variability between operators | Low | High (each crew has its own trade) |
| Regulatory constraints | General OHS | Seveso III + API 510/570 + LOTO + ATEX |
| Cost of the minute | Product margin | Unit downtime (1–3 M€/day) |
| Learning curve within the event | Stable | Starts at zero every shift (new crews) |
The consequence is that measuring a turnaround well is not about measuring more, it is about measuring something else: trades, crews and critical windows, not machine cycles.
2. The regulatory framework that governs the time study
Before setting up the stopwatch, three frameworks must be understood, since they impose constraints on who can do what, with which certification, and with which physical lockout:
- Directive 2012/18/EU (Seveso III), transposed in Spain by Royal Decree 840/2015 (which updates the original RD 1254/1999). Refineries and petrochemical plants handling flammable or toxic substances above certain thresholds are classified as lower-tier or upper-tier establishments. In plant practice, this is managed through the work permit (PT) and its internal approvals. (Eurofins, Modification of RD 1254/1999; Tesicnor, Seveso III and example of application).
- API 510 (Pressure Vessel Inspection Code) and API 570 (Piping Inspection Code). These are the American Petroleum Institute codes governing the inspection, repair, alteration and re-rating of pressure vessels and piping in service. API 510 and 570 are typically coordinated with scheduled shutdowns because internal inspection requires the equipment to be depressurized, isolated and cleaned (Inspectioneering, API 510; NDT Tanknicians, API 510 vs 653 vs 570; Norman QC, API 510 vs 570).
- LOTO (Lockout / Tagout) and plant lockout and tagging procedures. Without confirmed physical isolation, equipment opening is not authorized. This adds an administrative waiting time between permit request and execution that does not exist in a discrete factory.
Implication for the time study: every field activity carries an associated prior preparation (PT + isolation + atmosphere measurement + line breaking) that typically lasts 30–90 minutes (illustrative example) and is performed by personnel other than those doing the mechanical work. If the study does not capture this preparation, it systematically underestimates the standard time by between 15% and 30% (illustrative example).
3. The turnaround window: 5 phases with different metrics
The operating model used in the industry literature distinguishes five phases with different metrics and different VA/NVA/IMP categories. Here we adapt the classic phase scheme with the nomenclature used by the turnaround team:
| Phase | Typical duration | Objective | Key metric |
|---|---|---|---|
| F1 — Preparation (off-line, before shutdown) | 9–18 months | Engineering, permits, scope, contracts | % of tasks with closed work order |
| F2 — Shutdown and isolation (shutdown) | 3–7 days | Depressurize, empty, inert, LOTO | Time from last on-spec product to complete LOTO |
| F3 — Execution (the bulk of the work) | 30–50 days | Maintenance, inspection, changes | % direct vs indirect labor, tasks completed/day |
| F4 — Commissioning and testing | 5–10 days | Pneumatic/hydrostatic testing, step-by-step energization | Time between LOTO removal and first on-spec product |
| F5 — Start-up and stabilization | 3–5 days | Bring the unit to steady state | Time to sustained "on-spec" quality |
What is NOT VA in F3 (execution): many engineers arriving at their first turnaround confuse "time in the plant" with "productive time". Here VA is the time that brings the unit closer to a clean start-up; anything else is NVA or IMP. In a shutdown the size of Repsol Tarragona 2024 (600.000 person-hours estimated), the bulk of the work is concentrated in F3 (typical industry pattern); being able to distinguish in the work sampling between productive work (bundle pulling, channel head opening, pressure testing), waiting (scaffold not ready, PT unsigned, part missing) and logistical non-productive time (trip to the store, crane queue, refuelling) is the lever that moves the duration of F3.
4. Critical turnaround tasks: the catalog of 7 trades where the stopwatch makes the difference
From the real case history of refinery and petrochemical shutdowns, these are the seven task families where a robust time study generates the greatest return, ordered by daily opportunity cost:
- Extraction and reinsertion of heat exchanger tube bundles (heat exchanger bundle pulling). This is the most critical task in F3 in FCC units, hydrocrackers and vacuum units. The canonical sequence — channel head removal, cutting of bypass piping if applicable, bundle extraction with a bundle extractor (aerial, self-propelled or truck-mounted), chemical or mechanical cleaning, hydrostatic test, reinstallation — is suited to MTM study for the extractor preparation micro-motions and to stopwatch time study for the long pulling cycles.
- Opening and closing of distillation columns (trays manways, internals). Opening of the 20–80 manways of an atmospheric or vacuum distillation column (illustrative example), cleaning of internals (trays, packing, distributors), NDT inspection and closing. A task very intensive in rolling scaffolding and small-parts store.
- Furnace tube replacement (furnace tube replacement) in process heaters (crude, vacuum, FCC heating furnaces). It involves cutting, removal, preparation of new tubes with welding qualified to ASME Section IX, hydrostatic testing.
- Catalyst changeout in reactors (catalyst changeout) in continuous units such as hydrodesulfurization (HDS) or hydrocracker. It involves unloading, inerting, reactor opening, discharge of spent catalyst, loading of fresh catalyst, closing.
- API 510 / 570 inspection: non-destructive testing (NDT) — magnetic particle, ultrasonics, eddy current, radiography — on vessels and piping. The time standard here is built by trade and test type, not by overall item.
- Scaffolding and dismantling (scaffolding). This is the task with the greatest variability and the one that most overlaps with IMP-LOG. The difference between a scaffolding crew that knows the unit and one that enters for the first time can be 3:1 in productivity (illustrative example).
- Thermal insulation and painting after the testing phase. Usually subcontracted on a turnkey basis, but the time study is critical to accept the final certification of the scope.
For each one, the return on the time study is measured as hours saved per task × average man-hour cost × number of repetitions in the shutdown.
5. 7-step template for applying CronometrasApp in the turnaround
The field tool in a shutdown cannot be the same as in a discrete factory. This template adapts CronometrasApp to the Seveso + LOTO + multi-crew context. It fits the actual export flows to Excel and PDF that the app supports.
Step 1 — Create one project per shut-down unit
In CronometrasApp, open a project for each shut-down unit. The folder structure should reflect the WBS (Work Breakdown Structure) of the turnaround engineering, not the plant structure. For example, for a crude/vacuum unit: Unit 100 - Crude → 100-A Heat Exchangers → 100-A-01 E-101 Tube Bundle → Separate studies for extraction, cleaning, reinstallation and testing.
Step 2 — Define elements by trade, not by equipment
Each element should represent one trade performing a specific operation, not an abstract operation. Examples of well-defined elements:
Scaffolder — Assembling 2×2×3 m module at E-101 channel headPipefitter — Cutting and preparing T-301 manway level 4Mechanic — Removing E-101 front coverNDT inspector — Magnetic particle testing of S-201 flange weld
A poorly defined element (which must be avoided): Maintenance E-101 — it groups trades and hides real productivity.
Step 3 — Size the observation phase with Tippett
For work sampling in the field during F3 (where the classic work sampling technique does apply, described in depth in Work Sampling in Wind Farms), size it with Tippett's formula (1935):
For a p = 0.5 (maximum variance, conservative) and e = 0.03 (precision ±3%):
For e = 0.05 (±5%): N ≈ 400. For e = 0.10 (±10%, useful for management dashboards): N ≈ 100.
# Verificación numérica
p = 0.5
for e in (0.03, 0.05, 0.10):
N = round(4 * p * (1 - p) / e**2)
print(f"e={e}: N={N} observaciones")
# e=0.03: N=1111
# e=0.05: N=400
# e=0.10: N=100
The actual sampling schedule is spread over the 30–50 days of F3: if 30 observations/day are taken, 1.111 observations take ≈37 days — perfectly within the window.
Step 4 — Time with two capture approaches
Combine two approaches in the same project:
- Continuous mode: for tasks lasting between 2 minutes and 4 hours (manway opening, cover removal, pneumatic test). Lap reset or cumulative, as appropriate. This is the native mode of CronometrasApp.
- Micro-tasks with predetermined time systems (MTM) (extractor preparation, instrument valve opening, fitting connection): MTM-1 or MTM-UAS are industry methods that the team applies with their own tables if they know them, coding them in the element description field; they are not an app feature.
The choice of mode is decided before the task starts and is maintained throughout the observation to avoid sample contamination.
Step 5 — Categorize with a 6-state taxonomy adapted to the turnaround
The classic work sampling VA/NVA/IMP taxonomy works, but the categories must be redefined for the shutdown context:
| Code | Category | Definition adapted to the turnaround | Example |
|---|---|---|---|
| VA | Work that brings the unit closer to start-up | Direct mechanical task listed in the PT that advances the scope | Manway welding, pressure test |
| VA-WAIT | VA with active waiting on a downstream condition | Work started but blocked by an unsigned permit, crane occupied, part unavailable | NDT inspector waiting for sandblasting to finish |
| NVA | Movement or check that does NOT advance the scope | Trip to the store, searching for tools, refuelling equipment | Scaffolder crossing the unit to pick up a module |
| IMP-PT | Non-productive due to permits / LOTO / coordination | Unsigned permit, waiting for LOTO, delayed atmosphere measurement | Crew stopped 45 min due to uncoordinated line lockout |
| IMP-BREAKDOWN | Non-productive due to equipment breakdown | Broken crane, extractor with damaged hose, generator out of service | Waiting for a spare hydraulic pump for the extractor |
| IMP-WEATHER | Non-productive due to rain, wind, extreme cold | Suspension of work at height for wind > 50 km/h | Scaffold idle 2 h due to a storm |
The IMP-PT and IMP-WEATHER categories are the most underestimated in novice studies: in a typical shutdown on the Spanish Mediterranean coast, IMP-PT can consume between 8% and 15% of direct labor time (illustrative example), and it is the one most easily reduced through better permit coordination and grouping of tasks by area.
Step 6 — Calculate the standard with ILO/TAL allowances and record the learning ratio
The standard time of each element is calculated with the canonical formula:
Where TN is the normal time (observed average × performance rating factor) and S is the allowance. The allowances to be applied are those recommended by the ILO (International Labour Organization) — the classic classification for fatigue and personal allowances — plus the TAL allowances specific to the collective agreement. The sum of ILO + TAL covers:
- Fatigue allowances according to the type of physical effort (posture, weight handled, monotony).
- Personal needs allowances (5–7% of TN).
- Special conditions allowances for the turnaround: confined space work, ATEX, work at height, heat (>35 °C on furnace platforms), cold (winter shutdowns in Tarragona or Cartagena with gusts of >50 km/h).
In CronometrasApp, the allowances field of the study accepts multiple percentage entries, so the calculation can be done element by element and exported in the report.
Learning factor within the event (learning curve): in a shutdown, each crew starts from zero every shift because crews rotate, incoming shifts arrive with subcontracted personnel different from the previous shift, and trades change teams every 6–8 days. The provisional standard during the first 5–7 days of each trade must incorporate a Wright learning factor (typical curve L = 0.85 for skilled trades) until stability is observed:
where b = log₂(L) ≈ -0.234 for L = 0.85 (85% curve). This correction is what explains why the first days of a trade are systematically slower and why planning engineers who apply the "historical" standard from day 1 see deviations that look like measurement errors when they are in fact the learning curve.
Step 7 — Export the close-out report for two distinct audiences
The final report delivered after the turnaround has two audiences:
- Internal audience (engineering + plant management): Excel with all the element studies, VA/NVA/IMP ratios by trade and by unit, comparison with the previous shutdown (T-4), identification of the 10 tasks with the greatest deviation between planned and actual standard.
- Compliance audience (where applicable): PDF with the time studies of the tasks subject to API 510/570 together with the personnel qualification certificates of those who performed the inspection (API 510/570 require a qualified inspector, not time studies) and the record of the allowances applied.
CronometrasApp covers this dual flow: the export to Excel delivers the data per study for the internal audience to work with (pivot tables, filters), and the export to PDF consolidates the studies into an auditable document.
6. Simulated practical case: FCC unit shutdown
Simulated practical case. Example built for educational purposes from public industry case history. It does not correspond to a real Cronometras implementation or to a specific client: the figures are plausible but do not correspond to an identified real plant.
- Unit: FCC (fluid catalytic cracking) of medium size, shut down together with the coker unit.
- Total duration F2+F3+F4: 48 days.
- Budget executed: ≈95 M€.
- Total person-hours: ≈420.000.
- Distribution of VA / VA-WAIT / NVA / IMP-PT / IMP-BREAKDOWN / IMP-WEATHER over the total person-hours of the shutdown (F2+F3+F4) (measured with 800 instantaneous observations, p = 0.5, e ≈ 0.035):
| Category | % observed | Person-hours estimated | Observations |
|---|---|---|---|
| VA | 38% | 159.600 h | Direct mechanical tasks |
| VA-WAIT | 14% | 58.800 h | Active waits due to downstream blockages |
| NVA | 19% | 79.800 h | Trips, store, logistical waits |
| IMP-PT | 12% | 50.400 h | Uncoordinated permits, LOTO not prepared |
| IMP-BREAKDOWN | 8% | 33.600 h | Crane, extractor, generator breakdowns |
| IMP-WEATHER | 4% | 16.800 h | Suspension due to rain, wind, cold |
| Other IMP | 5% | 21.000 h | Lunch, training, additional PPE |
Interpretation (without inventing): a 30% reduction in IMP-PT (permit coordination, grouping of tasks by area, 24h advanced planning) equals 15.000 h saved in F3 — on the order of 1.7 days of shutdown (15.000 person-hours ÷ ≈8.750 person-hours/day, with the ≈420.000 person-hours over 48 days of the assumption). If the downtime cost is 1.5 M€/day, the estimated saving is around 2.5 M€ against an implementation cost for the measurement and coordination system of approximately 0.3–0.5 M€.
The payback is realized in a single shutdown.
7. Typical mistakes when timing a turnaround for the first time
From documented experience and shutdown close-out reports, these are the seven mistakes that keep recurring:
- Confusing time in the plant with productive time. In F3, being physically in the unit does not mean advancing the scope. The time study must record what the labor is doing at each instant, not where it is.
- Applying the planned standard from day 1 without correcting for the learning curve. The first 5–7 days of a trade are systematically slower; applying the historical standard generates deviations that look like measurement errors.
- Forgetting the invisible administrative time: issuing the PT, physical isolation, atmosphere measurement, handover signature. These 30–90 minutes per task (illustrative example) are the easiest to underestimate and the most critical for planning.
- Not differentiating productivity between trades. Scaffolders, pipefitters, mechanics and NDT inspectors have very different productivity ratios. A single standard for "maintenance" hides reality and blocks decision-making.
- Closing the study without classifying IMP-PT. If the team does not capture the permits/LOTO category, continuous improvement is impossible — management will see "slower than planned" without knowing where to attack.
- Not comparing with the previous shutdown (T-4). Without history, there is no benchmark. The inter-shutdown comparison is the metric management really wants to see.
- Forgetting cleaning and scaffold removal at the end. The dismantling phase, if not planned, consumes 3–5 additional days (illustrative example) that look like "unforeseen events" but are 100% known.
8. Close-out checklist for the turnaround time study
Before closing the shutdown report, this 10-point list must be at 100%:
- CronometrasApp project created per shut-down unit with the engineering WBS.
- Elements defined by trade, with Tippett sample size calculated for each F3 phase.
- ILO + TAL allowances applied element by element, with justification of special conditions (ATEX, confined space, height, temperature).
- Learning curve corrected for the first 5–7 days of each trade (Wright factor L = 0.85 unless justified).
- VA/NVA/IMP categorization complete, including IMP-PT and IMP-WEATHER.
- Comparison with the T-4 shutdown (4 years ago) in VA ratios and in the average cost of the top 10 tasks.
- Excel report exported for the internal audience with data by trade/unit.
- PDF report exported with the studies and the compliance documentation for the API 510/570 inspections.
- Qualification record of the personnel who performed the NDT inspections (UNE-EN ISO 9712 levels 1/2/3).
- Lessons learned documented and delivered to the planning team of the next shutdown (T+4 to T+6 years depending on the unit).
9. Conclusion
The time study of a turnaround is not a scaled-up version of the factory stopwatch: it is a different discipline combining work sampling techniques (Tippett), stopwatch time study (continuous and lap-reset modes), MTM for micro-motions, standards with allowances (ILO + TAL), and learning curve (Wright) on an engineering WBS within the Seveso III + API 510/570 + LOTO/ATEX framework. The economic lever is real: in an average shutdown of 95 M€ and 1.5 M€/day of downtime, a 30% improvement in IMP-PT pays for the measurement system in a single shutdown.
For a 360° plant audit combining turnaround with routine maintenance, time tracking and document traceability, the editorial umbrella ProdCont brings together the tools of the Micaot ecosystem: CronometrasApp for this type of study, WorkSamp for work sampling in the plant under normal operation, Induly for time tracking and traceability of subcontracted personnel during the turnaround, and Asetemyt for selecting the scaffolding, pipefitting and NDT inspection crews that execute the scope.
Sources
- Wikipedia, Turnaround (refining) — operational definition of turnaround/TAR.
- Diari de Tarragona, Repsol faces the most significant scheduled shutdown in terms of investment: 150 million — Tarragona 2024: 52 days, 150 M€, 675 items of equipment, 270 lines, 585 valves.
- Repsol A Coruña, Parada de combustibles 2024 — more than 60 days, ≈25 M€ (24,8 M€), 200 items of equipment, 8 plants.
- Cadena SER, Repsol invests 67 million in the scheduled shutdown of the Cartagena industrial complex — Cartagena 2025: 67 M€, shutdown extended until October.
- REIC, Fall Turnaround Planning for Petrochemical and Refinery Facilities — shutdown planning, 10–45 day windows.
- Intertek, Best Practices for Planning a Refinery Turnaround — turnaround planning cycle.
- Inspectioneering, API 510 — Pressure Vessel Inspection Code.
- NDT Tanknicians, API 510 vs 653 vs 570 — coordination of API 510/570 with scheduled shutdowns.
- Norman QC, API 510 vs API 570 — when both codes are needed.
- Eurofins Environment, Modification of RD 1254/1999 (Seveso III) — Seveso III framework in Spain.
- Tesicnor, Seveso III and example of application to an existing installation — transition Seveso II → III.
- Imubit, Turnaround Planning for Petrochemical Plant Operations — job packages and scope packaging.
Don't want to do it yourself? We do it for you from €800
If you prefer to outsource it, Cronometras delivers a complete time and motion study from a single video: elemental breakdown, OIT/TAL fatigue allowances, standard time with tolerances, and a report in PDF + Excel + JSON ready to present to management.