Long activities and critical path are related but different ideas in turnaround (TA) scheduling. A long activity is any task whose duration is large compared with the rest of the shutdown – column internals, catalyst change, heater refractory, compressor overhaul. The critical path is the longest chain of logically linked activities from feed-out to feed-in; its activities have zero (or the lowest) total float, so every hour lost on them moves the startup date. A long activity is only critical if it sits on that chain. Planners find the critical path with a CPM forward and backward pass (Total Float = Late Finish − Early Finish), schedule turnarounds in hours on 24/7 calendars, and protect the critical and near-critical paths with resources, pre-work, PERT/Monte Carlo risk analysis and, when needed, crashing or fast-tracking.
Every turnaround has a published startup date, and a refinery or petrochemical unit can lose a large amount of margin for each day it stays offline beyond that date. Whether that date holds depends on a small number of activities – the critical path – and on the long activities that can quietly become critical when inspection uncovers new work. This guide explains both concepts with a fully worked CPM example, a float diagram, the scheduling standards planners use (AACE, DCMA, GAO, PMI) and practical tools to estimate durations and compression cost.
It is written from 26+ years of hands-on shutdown and turnaround delivery in oil & gas, petrochemical, fertilizer, cement and power plants, where the same question is asked at every morning meeting: “Are we still on the critical path, and what is the float on the next path behind it?”
Float in turnaround scheduling: total, free, zero and negative
Float (also called slack) is the amount of time an activity can slip without delaying something else. In a turnaround, float is measured in hours because the whole execution window is usually only 10 to 45 days and the schedule runs on a continuous 24/7 calendar with two 12-hour shifts. Four float values matter to planners:
| Float type | Formula | What it tells you | Turnaround meaning |
|---|---|---|---|
| Total float (TF) | TF = LF − EF (= LS − ES) | How long an activity can slip before it delays the project finish | Hours of slack before the startup (feed-in) date moves |
| Free float (FF) | FF = ESsuccessor − EF | How long an activity can slip before it delays its next activity | Slack a crew can use without disturbing the next crew or permit |
| Zero float | TF = 0 | Activity is on the critical path | Any hour lost is an hour added to the shutdown |
| Negative float | TF < 0 | Schedule cannot meet an imposed date (e.g. a contractual startup) | Recovery plan required now: crash, fast-track or re-sequence |
| Near-critical float | 0 < TF ≤ threshold | Paths that can become critical with small delays | Track daily; typical threshold 24–48 h or 10–15% of duration |
In Primavera P6, Microsoft Project and similar CPM tools, the critical path can be defined either as activities with total float ≤ 0 or as the longest path. In turnarounds with imposed milestones, calendars and constraints, the two definitions can differ – the longest-path option is usually the more reliable one (see P6 settings).
What is a long activity in a turnaround?
A long activity is a task whose duration is large compared with the other activities in the turnaround schedule. “Long activity” is a planning term, not a formal CPM term like float. Planners use it to flag the work packages that consume most of the execution window and most of the specialist manpower. In practice there are three kinds:
- Long-duration work packages – physically large jobs such as column tray replacement, reactor catalyst change, fired heater retubing or a turbine major overhaul.
- Level-of-effort activities – activities that span the whole turnaround (fire watch, HSE supervision, document control, scaffolding maintenance). They are long but never drive the critical path and should not be linked into it.
- Sequential process chains – a string of short tasks done by the same crew or on the same equipment (open → clean → inspect → repair → NDT → PWHT → box-up) that together become a long, high-risk chain.
For control, long work packages are broken into activities of about one shift (8–12 hours) at execution level. A 120-hour “column repair” bar hides progress; ten 12-hour activities with clear inspection hold points expose slippage on the first day.
Long activities and critical path: key differences
| Attribute | Long activity | Critical path activity |
|---|---|---|
| Definition | Large duration relative to other activities | On the longest logic path; total float = 0 (or lowest) |
| Formal CPM term | No – descriptive planning term | Yes – defined by CPM forward/backward pass |
| Delay impact | Delays the finish only if it consumes all its float | Every hour of delay moves the startup date |
| Can it change? | Duration can grow with discovery work | Path can shift to another chain during execution |
| Example | Heat-exchanger bundle cleaning with 60 h float | Column internals repair with 0 h float |
| Control focus | Duration estimate, pre-work, crew productivity | Daily float tracking, resource priority, recovery plans |
| Typical tool | Work-pack estimate, PERT three-point estimate | P6 longest path, near-critical filters, Monte Carlo |
How to calculate the critical path (CPM forward and backward pass)
The Critical Path Method (CPM) calculates early and late dates for every activity from its duration and logic links. The steps below are the same whether you use paper, Excel or Primavera P6:
- List activities and durations in hours, with one responsible discipline per activity.
- Link the logic – mostly finish-to-start (FS) relationships; every activity needs at least one predecessor and one successor except the start and finish milestones.
- Forward pass: ES = largest EF of all predecessors; EF = ES + duration. The last EF is the project duration.
- Backward pass: LF = smallest LS of all successors; LS = LF − duration, starting from the project finish.
- Calculate float: TF = LF − EF; FF = earliest ES of successors − EF.
- Identify the critical path: the continuous chain of activities with TF = 0 (or the longest path when constraints exist).
Worked example: CPM for a distillation unit turnaround
The example below is a simplified distillation-unit turnaround with ten summary activities. Durations are in hours on a 24/7 calendar. It shows clearly why the longest activity is not automatically the only thing that matters, and why some long activities carry float.
| ID | Activity | Pred. | Dur. (h) | ES | EF | LS | LF | TF | FF | Status |
|---|---|---|---|---|---|---|---|---|---|---|
| A | Shutdown, cool-down, depressure, steam-out and purge | – | 36 | 0 | 36 | 0 | 36 | 0 | 0 | Critical |
| B | Positive isolation (blinding) and LOTO | A | 24 | 36 | 60 | 36 | 60 | 0 | 0 | Critical |
| C | Open column manways, gas test, entry permit | B | 12 | 60 | 72 | 60 | 72 | 0 | 0 | Critical |
| D | Column internals inspection and tray repair | C | 120 | 72 | 192 | 72 | 192 | 0 | 0 | Critical – longest activity |
| E | Exchanger bundle pull, hydrojet, re-install | B | 96 | 60 | 156 | 120 | 216 | 60 | 60 | Long, 60 h float |
| F | Piping replacement: weld, NDT, PWHT | B | 84 | 60 | 144 | 108 | 192 | 48 | 0 | Long, 48 h float |
| G | Column box-up and flange bolting (ASME PCC-1) | D | 24 | 192 | 216 | 192 | 216 | 0 | 0 | Critical |
| H | Hydrotest of replaced piping | F | 24 | 144 | 168 | 192 | 216 | 48 | 48 | Near-critical path |
| I | De-blinding and unit leak/tightness test | G, E, H | 24 | 216 | 240 | 216 | 240 | 0 | 0 | Critical |
| J | PSSR, start-up and feed-in | I | 36 | 240 | 276 | 240 | 276 | 0 | 0 | Critical |
How discovery work shifts the critical path
The critical path is not fixed – it moves whenever a non-critical activity grows by more than its total float. Turnarounds are especially exposed because much of the work is only fully defined once equipment is opened and inspected.
In the worked example, suppose eddy-current testing on exchanger E finds tube wall loss and the bundle must be partially retubed, adding 72 h. E now finishes at 156 + 72 = 228 h, which is later than G (216 h). Activity I can only start at 228 h, so the startup moves from 276 h to 288 h (+12 h). Only 12 hours were lost, not 72, because the first 60 hours of growth were absorbed by E’s float. From that moment, E is the new critical path and column work D gains 12 h of float.
Lesson for turnaround planners
- Report float consumed, not only percent complete, for every long activity at the daily meeting.
- Hold a discovery-work budget (contingency hours and approved emergent-work allowance) for equipment with uncertain inspection results.
- Re-run the schedule at least once per shift during peak execution so the new critical path is visible before the next shift starts.
Long activities that often drive turnaround critical paths
Which activity is critical depends on the unit, but the same long activities appear on the critical path of most refinery, petrochemical, fertilizer and power turnarounds. The durations below are indicative ranges for medium-to-large units on 24/7 execution; your own historical data always takes priority.
| Activity | Indicative duration | Why it is long | Critical path risk | Proven mitigation |
|---|---|---|---|---|
| Process and static equipment | ||||
| Column / tower internals (trays, packing, distributors) | 5–12 days | Confined-space entry, many trays, inspection and repair hold points | High | Pre-ordered tray parts, extra manways, 24/7 crews, inspection pre-planning |
| Reactor catalyst change (incl. inert entry) | 7–14 days | Cooling, unloading, inert-entry rules, sock/dense loading | High | Specialist contractor, pre-shutdown oxidation/cool-down plan, ready catalyst logistics |
| Fired heater tubes and refractory | 10–20 days | Tube cutting/welding, PWHT, refractory cure and dry-out curve | High | Prefabricated coils, refractory pre-cast shapes, dry-out included in schedule |
| FCC reactor/regenerator cyclones and refractory | 20–35 days | Large refractory areas, hot work, crane lifts | Very high | Pre-fabricated cyclones, detailed lift plans, refractory specialists |
| Heat exchanger bundle pull, clean, retube | 3–10 days per string | Crane/bundle-extractor availability, hydrojetting, tube testing | Medium (rises with discovery) | Spare bundles, pre-booked extractors, ECT results planned before TA |
| Rotating equipment and piping | ||||
| Compressor / steam turbine major overhaul | 10–21 days | Disassembly, OEM inspection, alignment, run-in | High | OEM supervisor booked early, spare rotor and parts on site |
| Piping replacement and tie-ins (weld, NDT, PWHT, hydrotest) | 4–10 days | Welder qualification, radiography windows, PWHT cycles | Medium | Pre-fabricated spools, hydrotest at workshop where allowed, NDT shift planning |
| Start and finish chains | ||||
| Shutdown, decontamination and isolation | 2–4 days | Cool-down rates, chemical cleaning, blinding sequence | Always on critical path | Blind list verified pre-TA, pre-assembled spades, chemical cleaning contractor ready |
| Box-up, de-blinding, leak test and start-up | 2–5 days | Bolting QA, tightness testing, PSSR, staged heat-up | Always on critical path | Bolting per ASME PCC-1, punch-list closure plan, PSSR team mobilised early |
Schedule quality standards and Primavera P6 critical path settings
A critical path is only trustworthy if the schedule behind it is logically sound. Turnaround planners increasingly check schedule quality against published guidance: AACE International RP 49R-06 (identifying the critical path), AACE RP 37R-06 (schedule levels of detail), the US GAO Schedule Assessment Guide (GAO-16-89G) and the DCMA 14-point schedule assessment. The DCMA metrics were written for long projects in days, so the right-hand column shows a common turnaround adaptation.
| Check | DCMA guideline | Turnaround adaptation | Why it matters |
|---|---|---|---|
| Missing logic | ≤ 5% of activities | 0% – every activity has a predecessor and a successor | Open ends create false float and hide the true critical path |
| Leads (negative lags) | 0% | 0% | Leads hide overlap and distort float |
| Lags | ≤ 5% | Minimal; use explicit activities for cure, cool-down, PWHT soak | Lags are invisible to crews and hard to statuse |
| Finish-to-start relationships | ≥ 90% | ≥ 90% | FS logic gives the clearest, auditable critical path |
| Hard constraints | ≤ 5% | Only contractual or permit milestones | Constraints override logic and can create negative float |
| High float | ≤ 5% with TF > 44 working days | Review any TF > 25% of TA duration | Large float usually means missing logic |
| Negative float | 0% | 0% in the baseline; any later value triggers recovery | Shows the startup date cannot be met |
| High duration | ≤ 5% with duration > 44 working days | Execution activities about 1 shift (8–12 h); long packages broken down | Short activities expose slippage early |
| Critical path test | Pass | Add delay to a critical activity: finish must move by the same amount | Proves the critical path is continuous |
| CPLI | ≥ 0.95 | ≥ 0.95 | Shows whether remaining float can still meet the finish |
| BEI | ≥ 0.95 | ≥ 0.95 at each shift / daily update | Shows whether baseline work is being finished on time |
Primavera P6 settings that change your critical path
- “Define critical activities as”: choose Longest Path for turnarounds with multiple calendars and constraints; “Total float less than or equal to 0” can flag false critical activities or miss real ones.
- Calendars: use a 24/7 hourly calendar for execution and separate calendars for crane, NDT or day-shift-only crews; mixed calendars change float values.
- Multiple float paths: run multiple float path analysis (free float or total float option) to list path 1 (critical) and paths 2–5 (near-critical).
- Retained logic vs progress override: use retained logic for statusing so out-of-sequence progress does not shorten the critical path artificially.
- Constraints: limit to permit, regulatory or contractual milestones; check for negative float after every update.
Duration risk on long activities: PERT and Monte Carlo
Long activities carry the largest duration uncertainty, especially those that depend on inspection results. A PERT three-point estimate converts optimistic (O), most likely (M) and pessimistic (P) durations into an expected duration and a spread:
Expected duration Te = (O + 4M + P) / 6 | Standard deviation σ = (P − O) / 6 | P80 ≈ Te + 0.84σ
For the whole turnaround, a Monte Carlo schedule risk analysis (AACE RP 57R-09) is more reliable than adding PERT values along one path, because it captures merge bias – when several near-critical paths join at box-up or de-blinding, the chance of finishing on time is lower than for any single path. Many owners set the published startup date at P50–P80 confidence and hold the difference as schedule contingency.
PERT duration and crash-cost calculators
Use these two quick tools during planning reviews. All calculations run in your browser; nothing is sent to a server.
PERT three-point duration estimator
Crash cost slope (schedule compression)
10 steps to control the critical path in a turnaround
- Freeze scope on time. Lock the worklist around T−6 months and route late additions through a formal approval gate with schedule impact shown.
- Build logic from job packs. Each work package should have defined steps, hold points, permits and resources before it enters the schedule.
- Schedule in hours on a 24/7 calendar, with execution activities of about one shift and separate calendars for constrained resources.
- Run a schedule quality check (missing logic, constraints, lags, critical path test) before baselining.
- Resource-load and level welders, riggers, cranes, NDT and PWHT crews; a resource-constrained path is often the real critical path.
- Move work out of the window through pre-shutdown activities: scaffolding, insulation removal, prefabricated spools, on-stream inspection and staged materials.
- Quantify duration risk on long and discovery-prone activities with PERT ranges and a Monte Carlo run.
- Track float every shift. Report critical path 1 and near-critical paths 2–5, float trend and hours ahead or behind.
- Control emergent work with an approved budget and a fast decision process; never let it enter the schedule without logic.
- Recover early using crashing or fast-tracking on the cheapest critical activities first, and record lessons learned for the next TA.
Schedule compression: crashing vs fast-tracking
| Technique | What it does | Typical TA examples | Main risk | Use when |
|---|---|---|---|---|
| Crashing | Adds resources to a critical activity to shorten its duration | Second crew on column trays, extra bundle extractor, additional welders and NDT shift | Higher cost, congestion, lower productivity in confined spaces | Cost slope is lower than the value of earlier start-up |
| Fast-tracking | Runs sequential activities partly in parallel | Start box-up of lower trays while upper trays are being repaired; parallel NDT and coating | Rework and safety conflicts (SIMOPS, dropped objects, permits) | Activities are separable and permits allow simultaneous work |
| Re-sequencing | Changes logic to use float on other paths | Move hydrotest crew from a float path to a critical path test | Creates new near-critical paths | Resource is the real constraint |
| Scope deferral | Removes non-essential work to the next window | Defer non-integrity improvement jobs | Unplanned failure before next TA | Risk-based inspection (API 580/581) confirms deferral is acceptable |
Critical path KPIs for daily turnaround control
| KPI | Formula / definition | Target or trigger |
|---|---|---|
| Hours ahead/behind critical path | Forecast finish − baseline finish (h) | 0 h; recovery plan if behind |
| Critical path float trend | TF of path 1 at each update | Stable at 0; any negative value = action |
| Near-critical paths | Number of paths with TF ≤ threshold | Track paths 2–5 daily |
| CPLI | (Critical path length + critical path TF) / critical path length | ≥ 0.95 |
| BEI | Activities completed / activities baselined to finish by status date | ≥ 0.95 |
| SPI (earned value) | EV / PV (man-hours) | ≥ 0.95; review if trending down |
| Emergent work | Added man-hours / frozen-scope man-hours | Within approved allowance (often 10–20%) |
| Schedule compliance per shift | Planned activities completed in the shift / planned | ≥ 85–90% |
Common critical path mistakes in turnarounds
- Treating the longest activity as the only critical item and ignoring near-critical paths.
- Linking level-of-effort activities (supervision, fire watch) into the logic, which creates a false critical path.
- Using hard constraints instead of logic, producing negative float that hides the real problem.
- Statusing progress without updating remaining duration, so float looks healthier than it is.
- Leaving cure, cool-down, PWHT soak and refractory dry-out as hidden lags instead of visible activities.
- Allowing emergent work to start without adding it to the schedule with logic.
Key takeaways: long activities and critical path
- The critical path is the longest logic chain from shutdown to start-up; its activities have zero (or the lowest) total float.
- Long activities consume the most time and manpower, but they are critical only when they sit on that chain or lose their float.
- Calculate float with a forward and backward pass: TF = LF − EF and FF = ESsuccessor − EF.
- Discovery work on long, inspection-dependent activities is the most common reason the critical path shifts mid-turnaround.
- Protect the start-up date with a quality-checked hourly schedule, near-critical path tracking, PERT/Monte Carlo risk and early recovery actions.
Frequently asked questions
What is a long activity in turnaround scheduling?
A long activity is a task whose duration is large compared with the other activities in the turnaround, such as column internals repair, reactor catalyst change, fired heater retubing or a compressor overhaul. It is a descriptive planning term, not a formal CPM term, and it is only critical if it lies on the longest logic path to start-up.
Is every long activity on the critical path?
No. A long activity can have plenty of total float if it runs in parallel with a longer chain. In a typical example, a 96-hour heat exchanger job can carry 60 hours of float while a 120-hour column job is critical. It becomes critical only if discovery work or delays consume all of its float.
What does zero float mean in project scheduling?
Zero float means an activity has no scheduling flexibility: its early and late dates are the same, so any delay directly delays the project finish. Activities with zero total float form the critical path in a schedule without imposed constraints.
What is the difference between total float and free float?
Total float is how long an activity can slip without delaying the project finish (late finish minus early finish). Free float is how long it can slip without delaying the early start of its immediate successor (successor early start minus the activity early finish). Free float is always less than or equal to total float.
What is negative float in a turnaround schedule?
Negative float appears when the logic-driven finish is later than an imposed date, such as a contractual or commercial start-up date. It shows how many hours must be recovered, and it should trigger a recovery plan using crashing, fast-tracking, re-sequencing or scope deferral.
How do you calculate the critical path?
List activities with durations and logic, then do a forward pass (early start equals the largest early finish of predecessors; early finish equals early start plus duration) and a backward pass (late finish equals the smallest late start of successors; late start equals late finish minus duration). Total float equals late finish minus early finish, and the continuous chain with zero or lowest float is the critical path.
Should I use longest path or total float to define the critical path in Primavera P6?
For turnarounds with multiple calendars, constraints and hourly scheduling, the Longest Path option is usually more reliable. The total float less than or equal to zero option can flag activities as critical only because of calendars or constraints, or miss real drivers.
Which activities usually drive the critical path in a refinery turnaround?
Common drivers are shutdown and decontamination, positive isolation, column and reactor internals, catalyst change, fired heater tubes and refractory, FCC reactor and regenerator work, major compressor or turbine overhauls, box-up and de-blinding, and start-up with the pre-startup safety review.
How can the critical path be shortened during a turnaround?
Crash the critical activity with the lowest cost slope by adding crews or equipment, fast-track activities that can safely overlap, re-sequence resources from paths with float, move work into the pre-shutdown phase, or defer non-integrity scope after a risk-based review. Re-check the critical path after each change.
How detailed should a turnaround schedule be?
Execution-level turnaround schedules are normally built in hours on a 24/7 calendar, with activities of about one shift (8 to 12 hours), clear hold points and one responsible crew per activity, so slippage on long work packages is visible within a shift.
Test your knowledge
1. An activity has ES = 60 h, EF = 156 h, LS = 120 h and LF = 216 h. What is its total float?
2. Which statement about long activities is correct?
3. What does negative float indicate?
4. PERT expected duration for O = 96 h, M = 120 h, P = 180 h is:
5. Adding a second crew to a critical activity is called:
Read the complete Aienginear plant turnaround guide for T-minus milestones, scope freeze, job packs, isolation and PSSR, then use the calculators for hydrotest pressure and flange bolt torque.
Related Aienginear guides and tools
References and standards
- AACE International RP 49R-06 – Identifying the Critical Path; RP 37R-06 – Schedule Levels of Detail; RP 38R-06 – Documenting the Schedule Basis; RP 57R-09 – Integrated Cost and Schedule Risk Analysis Using Monte Carlo Simulation of a CPM Model.
- U.S. Government Accountability Office, GAO-16-89G – Schedule Assessment Guide: Best Practices for Project Schedules.
- Defense Contract Management Agency (DCMA) – 14-Point Schedule Assessment.
- Project Management Institute – Practice Standard for Scheduling; PMBOK® Guide.
- ISO 21502 – Project, programme and portfolio management: Guidance on project management.
- OSHA 29 CFR 1910.119 – Process Safety Management (pre-startup safety review, management of change, mechanical integrity).
- API RP 580 / API RP 581 – Risk-Based Inspection; API 510 and API 570 – Pressure Vessel and Piping Inspection Codes.
- ASME PCC-1 – Guidelines for Pressure Boundary Bolted Flange Joint Assembly; ASME B31.3 – Process Piping.
- Oracle Primavera P6 EPPM / Professional documentation – scheduling options for critical activities, longest path and multiple float path analysis.
Written by Muhammad Raza
Founder of Aienginear. Engineering and digital solutions professional with 26+ years managing projects, plant shutdowns and turnarounds in oil & gas, petrochemical and power generation.
