Shutdown maintenance project management: running the 'shutdown days' as a controlled project

I. Why shutdown maintenance is a 'project', not a 'job'

The annual shutdown or campaign is a one-off, clearly-scoped, resource-intensive temporary project: dozens to hundreds of people, hundreds of jobs, a window of only days to weeks. It has every feature of a project — scope, time, cost, quality, safety, stakeholders. Managing it with a 'daily maintenance' mindset (just do it when it comes, add what's missing) inevitably loses control: jobs clash, materials arrive late, isolation is missed, rework. So shutdown maintenance must be run as project management, with WBS (work breakdown), schedule (critical path), safety (energy-isolation master diagram), acceptance (Punch list). This echoes M01's 'planning beats ad-hoc'.

II. Shutdown plan: window, scope, resources

The planning stage answers three things: window (which day to stop, how many days, off-peak/peak avoidance), scope (what to repair, what not to, avoiding 'touch everything while at it' causing scope creep), resources (people, materials, tools, contractors, lifting). Scope is derived from 'maintenance list + defect ledger', not memory. Resources must calculate the critical path — the slowest job chain sets total duration; bottleneck jobs (e.g. large-part lifting, dryer hydraulic test) get priority scheduling. The plan also keeps a buffer: real maintenance always has surprises, zero buffer = sure overrun.

III. Work package: job / material / tool / risk

Each job is packaged into a work package: work content, standard (technical parameters, torque, clearance), required materials (incl. spare-part number, echoing M06), tools, lifting and scaffolding, risk and isolation requirements, acceptance criteria, responsible person. The work package turns 'the veteran's experience in his head' into 'a reproducible work order'. Materials must arrive early and be checked — find one missing seal on site and the whole critical path stalls. The risk field must state which energy-isolation classes this job involves (echoing M05's six energy classes of LOTO).

IV. Installation and documentation system (As-Built)

Maintenance changes (roll change, repiping, crown adjustment, bearing change) must leave an installation record and As-Built (as-built drawing): actual dimensions, actual clearances, actual torque, serial numbers. Many machines get 'messier with each repair' because changes are not recorded, and years later no one knows 'what this machine truly looks like now'. The documentation system also includes: maintenance log, defect-closure record, spare-parts consumption, anomalies and deviations. These link directly to M01's RCFA and M06's ledger — one proper shutdown should let the next shutdown 'stand on this one's shoulders', not guess from zero.

Digital practice: use mobile terminals to enter data in real time at the job point (photo + form), avoiding 'back-fill records later' causing distortion. Updating the as-built drawing and BOM (M06) in sync is the landing point of the equipment full-lifecycle data loop.

V. Schedule and safety control

The execution stage has many parallel jobs; control two things: schedule — daily stand-up on the critical path, red/yellow/green warning on bottleneck jobs; safety — energy-isolation master diagram (which equipment, which energy classes, who locked), cross-operation (no entry below lifting, hot-work isolated from oil cleaning), contractor management. Safety control outranks schedule: any 'skip isolation to catch up' trades lives for time. Shutdown-period accidents are a high share of the paper industry precisely because 'stopped means let guard down'.

VI. Maintenance quality acceptance (Punch List)

Each job's completion needs acceptance criteria: whether dimensions/clearances/torque meet spec, whether trial run is smooth, whether any leak. After all jobs, form a Punch List: unclosed items, responsible person, deadline. No start-up until the Punch List is cleared — starting with open items turns small hazards into shutdown accidents. Acceptance also checks against the work package's technical standard, not 'looks about right'. Much rework comes from 'soft acceptance', surfacing only after start-up at the cost of another shutdown.

VII. Start-up verification checklist

Start-up is not 'just power on' but a verification checklist: isolation-release confirmation (all LOTO removed, no one left inside), single-point jog → interlock → with stock, re-measure key parameters (echoing M09 start-up sequencing, M07 hydraulic acceptance four numbers, M08 ventilation three numbers), compare with the pre-maintenance baseline (M04/M10 snapshots). The 72 hours before start-up are the 'new baseline' building period; record all anomalies into the file, distinguishing 'always like this' from 'introduced by maintenance'. This checklist follows the same line as M09 §XI and M07 §XI start-up sequencing — from local to global, stabilize before linking.

VIII. Interface with M01 RCFA / PM

Shutdown maintenance is the 'concentrated practice field' of M01's maintenance methodology: the defect ledger accumulated by daily BM/PM/CBM is the input to maintenance scope; hidden defects found during maintenance feed back to RCFA for root cause; the maintenance Punch List is the basis for PM plan optimization. Conversely, without daily defect-data support, scope is 'by guess', either missing repairs or over-repairing. So maintenance is not an isolated event but the 'centralized cashing-in' of daily O&M data.

IX. Interfaces with Part B / the grades

Maintenance content differs by grade: graphic paper (M10) weighs calender rolls / dryer condensate; board (M11) weighs large dryer cylinder / multi-layer headbox lip / coating system; tissue (M12) weighs Yankee flaw detection and creping blade (safety-related, must-check). But the management framework is common: work package, isolation master diagram, Punch List, start-up verification. Part B's three lectures are the 'technical content library' of maintenance — the maintenance standards for hydraulic loading (M07), ventilation balance (M08), drive speed chain (M09) all live there. Combine the management framework (this lecture) with the technical content (Part B + Part C) to get a complete maintenance plan.

X. Typical maintenance scenarios

Annual shutdown: full-machine teardown level, roll/bearing change, dryer pressure test, major retrofit, window days to weeks, densest resources. Campaign: targets one system (e.g. dryer section, pulp line), shorter window. Breakdown extended to plan: after a sudden failure repair, use the shutdown window to handle related hidden risks together, avoiding 'fixed east, leaked west'. Different scenarios, different planning granularity, but the four-phase framework is unchanged.

XI. Digitalization and knowledge retention

Maintenance digitalization is not just 'paperless' but 'knowledge retention': mobile terminals enter data in real time, as-built drawings auto-archived, defects linked to RCFA, spare-parts consumption written back to BOM (M06). The value of retention is 'smarter next maintenance' — don't step in the same pit twice, do the same job faster and steadier. This is also where M01 self-improvement comes from: turn each maintenance into organizational memory, not personal experience. A mature maintenance system sees downtime loss decrease with each cycle, because each time it reinforces the data base.

XII. Contractor safety and briefing management

During maintenance, contractors often make up over half the workforce, from different units with varied work habits and risk awareness — the highest-frequency accident source in the shutdown period. Entry must pass three gates: qualification review — special-work certificates (welder / lifting / height work), insurance, medical exam valid; safety briefing — explain the energy-isolation master diagram person by person, the machine's specific risks (e.g. Yankee high-pressure steam, dryer high temperature), emergency evacuation route; on-site supervision — dedicated person watching cross-operation and isolation effectiveness, no formality. Two iron rules: no briefing, no work; no supervision, no operation. Many mills treat contractors as 'outsiders' and lower standards, yet bear the blame themselves when accidents happen — contractors must be brought under the same LOTO and work-permit system, the premise of zero accidents in maintenance.

XIII. Maintenance-period fire protection and hot-work management

During maintenance, cutting, welding and other hot work are frequent, often near lube stations, hydraulic lines and paper-dust buildup — fire risk is higher than in running periods. Hot-work 'four pieces' are all mandatory: hot-work permit (approval + risk identification + combustible clearance confirmation), combustible-gas detection (qualified to start, re-test during work), supervision + fire extinguisher in place (dedicated person with extinguisher on site), oil/dust isolation (M05: water forbidden for oil fires, use smothering or dry powder; open flame strictly forbidden in dust zones). Maintenance-period fires are a high share, rooted in the 'stopped so let out a breath' complacency — the more stopped, the more fire must be watched like in running periods; a physical isolation belt must be kept between the hot-work point and the oil station.

XIV. Maintenance and M06 spare-parts linkage

The work package's material list (incl. spare-part numbers, echoing M06) must be checked for completeness against inventory before shutdown; key parts — large dryer bearings, main seals, hydraulic power-unit components, creping blades — should have a 'maintenance-dedicated safety stock' to avoid one missing seal stalling the whole critical path on site. More importantly write the missing-part emergency flow early: borrow from sister machines, activate certified substitutes, or move the item into the Punch List as deferred — but deferred items must never be safety-related or pressure-bearing parts. Spare-part completeness directly sets shutdown-window utilization, a hard indicator locked at the planning stage.

XV. Digitalization: BIM / AR and mobile terminals

Maintenance digitalization is not 'paperless' gimmickry but turning the veteran's experience into retrievable data assets. Three pieces: mobile terminal enters as-built dimensions/defects/torque and photos at the job point, avoiding 'back-fill' distortion; BIM model overlays equipment assets — open to see 3D location + history + linked drawings, so newcomers find the right part; AR remote lets experts annotate and guide complex teardown through the screen, easing scarce-trade shortages. Value is 'direct recall next maintenance' — same job faster and steadier, same pit not stepped in twice. This is also where M01 self-improvement lands: each maintenance reinforces organizational memory, not just a few blurry photos.

XVI. Maintenance review and continuous improvement

Maintenance end ≠ project end. The review meeting must quantitatively output five things: scope achievement rate (what repaired, what missed), schedule deviation root cause, safety incidents and near-misses, Punch List closure rate, spare-parts consumption vs budget. Each deviation feeds back to RCFA (M01) and the PM plan — e.g. a roll repeatedly rumbling, the review pinned it to a wrong assembly-clearance standard in the work package, fixed and rolled into the next work package. Review discipline decides 'whether smarter next time': mills that only repair without reviewing step in the same pit yearly, downtime loss rises instead of falling; mills that keep reviewing do the same maintenance shorter each time, loss decreasing per cycle. Writing review conclusions into the M06 ledger and M01 knowledge base is the closed loop.

XVII. Maintenance-period lifting and scaffolding safety

During maintenance, large-part lifting (dryer, roll, beam) and scaffolding erection are high-risk operations, managed separately. Lifting: rigging and slings must be inspected on entry (flaw detection / load plate), lifting points calculated, no entry below, stop lifting beyond wind-speed limit; command by standard signals / radio, no 'eyeball the lift'. Scaffolding: acceptance tag (green usable / red forbidden), inspect before each shift start, no overload stacking. Once these two go wrong it's mass casualties, so stricter than daily — many maintenance accidents lie precisely in the luck of 'throw up a frame, jury-rig a lift'.

XVIII. Maintenance and M04/M10 baseline comparison

Start-up verification (§VII) is essentially 'compare with baseline'. Baseline comes from: M04's OEE and history snapshots, M10's operating snapshots, each grade lecture's rated parameters. After maintenance, the first roll / first batch must be compared parameter by parameter — machine speed, line load, steam consumption, vibration, product indices back to or better than baseline; deviations go to Punch List or RCFA. Without a baseline you can't tell 'new problem introduced by maintenance' from 'always like this', and review is impossible. Suggest exporting a 'baseline package' before each major overhaul, archived with the work package.

XIX. Maintenance budget and cost control

Maintenance is not 'cheaper is better' nor 'pricier is steadier'; the key is 'spend on bottlenecks and risks'. Three budget blocks: materials (planned spares + safety stock), outsourcing (contractor labour + lifting/flaw-detection specials), hidden cost (downtime loss = unit-time output value × window duration). Control levers: lock scope to prevent creep (§I), optimize critical path to compress window (directly cuts hidden cost), complete spares to cut emergency-purchase premium (§XIV). A good maintenance budget should minimize the cost of 'every stopped hour' — consistent with management-accounting thinking, downtime burns money.

XX. Maintenance scheduling and critical-path practice

The critical path is not drawn on paper and done — it is a live diagram updated daily. Practice points: daily stand-up marks each job's status red/yellow/green, yellow card → intervene (add people / add equipment / reorder); bottleneck jobs (large-part lifting, dryer pressure test) get a dedicated owner + backup plan, switch instantly if stuck; parallel jobs scheduled by 'area / time-slot matrix' to avoid cross clashes. Much overtime comes from 'pretty critical-path chart, no one watching on site' — the scheduling meeting must have someone own the path, not pin it on the wall as decoration.

XXI. Change management and records in maintenance

During maintenance it's common to find 'more complex than planned, need to change the approach' — that's a change. Change management three things: record (what changed, why, who approved), re-assess risk (does the new approach introduce new isolation gaps), sync documents (As-Built updated in real time, not back-filled). Most dangerous is 'silently changed on site, documents not kept up' — after start-up no one knows the machine was modified, next maintenance steps in the pit. Change records are the dynamic patch of As-Built (§IV), must be archived into the work package.

XXII. Maintenance safety culture and near-miss reporting

Zero accidents rely not only on systems but on 'safety culture' — everyone dares to stop and report. Push near-miss reporting: hazards that didn't cause an incident but were close (one isolation step missed, lifting nearly skewed) are also reported and analyzed, killing accidents before they happen. Culturally 'report problems, don't punish people; hide problems, then punish' — otherwise the front line will cover up. Maintenance is temporary, crowded, fast-paced, most prone to 'good enough'; near-miss reporting is the daily mechanism that pulls this tendency back, in the same vein as M05's safety culture.

XXIII. Maintenance-period environmental and waste management

Maintenance-period environmental and waste management is often overlooked: used oil (hydraulic / lube) is hazardous waste, must be collected by category and handed to qualified units; wash water with fibre/chemicals cannot be discharged directly; old seals/bearings and other oil-bearing solids likewise. Many mills 'busy with shutdown, relax on environment', then get fined. Suggest listing 'waste list + collection point + transfer record' in the work package, checked as mandatory like LOTO. The environmental red line is in M05 general safety; during maintenance it must be localized — shutdown is not an environmental exemption period.

XXIV. Memory aids for this lecture

• Shutdown maintenance is a project not a job: has WBS, critical path, energy-isolation master diagram, Punch List.
• Three planning elements: window, scope (back-track defect ledger), resources (bottleneck jobs first).
• The work package turns experience into a work order: content / standard / material / tool / risk / acceptance / responsible person.
• As-Built must be kept: actual dimensions, clearances, torque, serial numbers, or the machine gets 'messier with each repair'.
• Safety outranks schedule: no LOTO, no work; stagger cross-operations.
• No start-up until Punch List cleared; no soft acceptance, or rework after start-up.
• Start-up verification by checklist: isolation release, jog → interlock → with stock, compare with pre-maintenance baseline.