Maintenance Planning System｜13-Week Scheduling in Thailand

A maintenance planning system turns approved work into execution. This guide shows Thai factories how to connect a 13-week rolling plan, ready backlog, weekly freeze, daily dispatch and actual feedback across labor, skills, parts, downtime and permits—and test the design in an RFP and PoC.

The gap a maintenance planning system closes

A factory receives calendar- and meter-based preventive work, operator requests, inspection findings, IoT alerts, improvement jobs, and tasks with legal or safety due dates. A work order, however, is not automatically ready to execute.

Consider a pump-bearing replacement. Knowing the asset, symptom and priority is insufficient unless:

scope and completion criteria are clear;
an employee or contractor with the required skills is available;
parts, consumables and special tools are usable;
production has approved equipment downtime;
LOTO, hot-work, work-at-height, confined-space and other controls are identified; and
upstream/downstream, quality, test-run and hand-back responsibilities are agreed.

The maintenance planning system exposes work that is “created but not prepared,” converts it into a ready backlog, and allocates it to limited resources and downtime. IBM’s maintenance-planning documentation likewise frames planning around the availability of tools, labor and facilities so that required work can be packaged and performed at an appropriate time.

Separate planning, scheduling and dispatching

The same application may support all three, but the work is different.

Planning prepares what and how: scope, method, hazards, skills, standard hours, parts, tools, drawings, permits and acceptance criteria.

Scheduling decides when prepared work will occur: it aligns the production plan, downtime, shifts, skill capacity, part arrivals and dependencies.

Dispatching turns the frozen schedule into today’s execution: it responds to absence, breakdown, a longer production run or a rejected part and adjusts sequence and assignment.

Without the distinction, planners spend every morning firefighting instead of preparing future work. Schedulers force incomplete orders onto dates, while technicians search for parts and permits. Workflow states and permissions should preserve these responsibilities.

See future load with a 13-week rolling plan

Thirteen weeks is a convenient example because it covers a quarter while remaining easy to refresh weekly. It is not an ISO requirement or an industry-mandated horizon. A site with major shutdowns and imported long-lead parts may need 26 weeks or more; a volatile process may detail eight weeks. Select a horizon that covers the plant’s preparation lead time.

Each week, move the window forward and add the newly visible demand. Include more than forecast PMs:

calendar- and projected-meter preventive maintenance;
condition findings that may approach an intervention point;
corrective backlog and improvement work;
legal, safety, quality and calibration deadlines;
opportunity work during a shutdown, product change, cleaning or modification; and
work requiring long-lead parts, contractors or special tools.

Do not freeze all 13 weeks at the same resolution

Fixing a precise date and person far into the future creates needless replanning. Raise detail as work approaches:

Week 1 — execution commitment: normally fix work, owner, downtime, permits and parts.
Weeks 2–4 — preparation commitment: close scope, reservations, contractor orders and downtime requests.
Weeks 5–8 — load balancing: move work while reviewing skill hours, downtime and supply risk.
Weeks 9–13 — demand outlook: see probable work and rough capacity; advance only long-lead actions.

Keep an earliest start, required finish and preferred week rather than one date alone. These boundaries tell the scheduler where work can move without violating risk or compliance.

Compare demand and capacity in the same unit

“Fifty jobs” cannot be balanced against “five people.” Break demand into standard hours by skill: mechanical, electrical, instrumentation, PLC, welding or required certification. Build capacity by subtracting meetings, training, leave, routine rounds and an explicit break-in reserve from shift hours.

An illustrative internal calculation might start with four electricians × 40 hours = 160 nominal hours. After 16 hours for meetings/training, 24 for routine response and 32 for emergency reserve, only 88 hours are assignable to planned work. The values 160, 16, 24, 32 and 88 are method examples—not benchmarks. Actual calculation must reflect shifts, overtime rules, overlapping skills and contracts.

Total labor may balance while electrical demand is overloaded and mechanical capacity is free. Stack weekly hours by skill and review them with part arrivals, downtime and outsourcing options.

Treat downtime as an equipment calendar

Receive equipment-specific availability, product changes, cleaning, holidays and planned shutdowns from production planning. “Maintenance on Saturday” is too vague when upstream overtime or a start-up trial consumes the window. Store shutdown start, usable maintenance time, test/quality verification and hand-back deadline separately.

Concurrent work may compete for common power, utilities, a crane, access route or LOTO authority. A planning system must overlay people, asset, location and shared-resource calendars.

Maintenance Planning System｜13-Week Scheduling in Thailand - figure 1

Make the ready backlog the gate to weekly scheduling

A ready backlog is not a list of high-priority work. It is work that meets defined preparation conditions and can receive an execution date. If Created, Planning, Awaiting Parts, Awaiting Downtime and Ready are not distinct, backlog hours say nothing about how much can actually be scheduled.

Definition of Ready

Set a plant-specific gate and verify it in the system:

Asset ID, scope, priority and required finish are approved.
Job plan and completion/acceptance criteria exist.
Standard hours are estimated by skill.
Required parts are in stock and reserved, or have a dependable arrival date.
Tools, fixtures, instruments and shared equipment are available.
LOTO, hot work, height, confined space and chemical-permit needs are identified.
Drawings, methods, SDS and quality criteria are current.
Production, quality and maintenance agree on downtime and hand-back ownership.
Contractor order, access, qualifications, insurance and induction are confirmed.
Predecessor and successor dependencies are known.

Minor jobs need not pass an identical heavy gate. Template Ready criteria by work type, risk and criticality. Emergency work requires a separate exception path that permits safe minimum information at start and forces missing data to be completed before technical close-out.

Measure the health of ready backlog

In addition to total backlog weeks, review by skill:

ready versus unprepared hours;
hours waiting for parts, downtime, engineering or approval;
remaining time to required finish;
jobs returned to planning for rework; and
revisions to standard hours after scope change.

Six weeks of total backlog with only 0.5 ready week will not support a stable next-week schedule. Twenty ready weeks may indicate stale priorities, obsolete work or lack of shutdown opportunity. Do not claim one universally correct level; set site thresholds from lead time, outage frequency, break-in rate and skill mix.

Create a production–maintenance commitment through weekly freeze

In the weekly scheduling meeting, allocate next week’s ready work against skill capacity, downtime, parts, permits and shared resources. Then freeze the baseline and treat later changes as exceptions.

Freeze does not mean “never change.” Safety, quality, major failure and production changes still require action. Record who changed the schedule, why, which job left and which job entered. Without reason codes, poor schedule compliance cannot be attributed to preparation, production change, absence, parts or breakdown.

Standard weekly-freeze process

Confirm available hours by skill and contractor capacity.
Confirm downtime, test-run and quality-verification windows with production.
Select candidates from required finish, criticality, risk and opportunities.
Recheck Ready evidence, reservations, permits and dependencies.
Assign work packages to day, shift and owner.
Leave explicit reserve capacity for emergency work.
Obtain production, maintenance, quality and safety approval.
Record the frozen version and timestamp; audit subsequent changes.

A plant may close its schedule at 15:00 Thursday, for example. That time is illustrative, not a standard. Derive it from stores kitting, contractor confirmation and permit pre-review lead times.

Define schedule compliance carefully

“Jobs completed as scheduled ÷ jobs scheduled” gives a two-hour job the same weight as a 20-hour job. Report count and labor-hour views, and define complete, partially complete, deferred, canceled and scope-changed.

Do not improve the number by scheduling only easy work or splitting incomplete orders. The metric should expose planning quality; using it as a single personnel score invites distortion.

Deliver the frozen plan through daily dispatch

Daily dispatch is not a meeting to rebuild the weekly schedule. It is a short control that verifies today’s safety, equipment condition, people, parts and downtime, then adjusts sequence and assignment.

Shift-start checks

carryover, asset hand-back and residual risks from the previous shift;
Ready status and current procedure for today’s work;
assigned people, skills, qualifications and contractor access;
kitted parts, tools and instruments;
permits and owners for LOTO, hot work and height;
actual production handover, test-run and quality window;
new safety, quality and equipment alarms; and
displaced job and impact when an emergency enters.

Technicians should receive mobile work orders and record not only Start/Hold/Complete but also hold reason, added scope, consumed parts, readings and findings. Where connectivity is unreliable, the RFP should require offline access for today’s work and duplicate-safe synchronization.

Agree break-in rules before the shift

If every request called “urgent” breaks in, the frozen baseline has no meaning. Define categories for safety, environment, quality, total outage, degraded capacity and delivery impact, and name the urgency approver.

When work enters, update the required finish, reservation and downtime for the displaced job. Dragging it to tomorrow without propagating impact conceals future overload. Returning daily decisions to the 13-week plan completes the loop.

Maintenance Planning System｜13-Week Scheduling in Thailand - figure 2

Feed plan-versus-actual variance into next week

Completion needs more than a check mark. Capture variance to improve the next estimate and preparation—not to assign blame.

Variance to capture

planned versus actual hours, by skill and headcount;
scheduled start/finish versus actual start, restoration and close;
reserved versus consumed parts, returns and additional issues;
planned versus actual outage, wrench time and waiting;
expected versus added or unnecessary scope;
expected failure mode/cause versus finding;
waits for permit, drawing, tool, scaffold, cleaning or quality; and
test result and rework after maintenance.

Do not close the discussion with “the technician was slow.” The standard may be stale, access poor, kit incomplete, permit owner absent, production handover late or diagnosis wrong. Separate causes that future planning can change.

Weekly plan-versus-actual review

Before rolling the window, review the largest variances, break-ins, deferrals, rework, part waits and outage overruns. For each, decide whether to update the job plan, standard hours, bill of materials, Ready criteria, downtime rule or capacity factor.

As an illustrative calculation, if the same recurring task was planned for four hours but took six, 5.5 and six hours, do not simply change the standard to the 5.8-hour average. First determine whether roughly two hours came from scaffold waiting or required scope. The values 4, 6, 5.5 and 5.8 are method examples, not benchmarks.

Connect KPIs across the horizons

Use leading measures as well as outcomes:

skill-capacity shortfall against 13-week demand;
unresolved part, contractor and downtime commitments in Weeks 2–4;
ready hours versus candidate hours for next week;
post-freeze changes and reasons;
schedule compliance by count and labor hours;
planned/actual labor variance and waiting reasons;
break-in rate and due-date effect on displaced work; and
PM completion within tolerance and recurring failures.

Define formula, population, start/end timestamps and exclusions. Comparing sites with different definitions creates an input contest rather than improvement.

Put preventive and predictive work into the 13-week plan

Preventive maintenance initiates work from calendar, operating hours or cycle rules before failure. Predictive maintenance estimates deterioration or intervention timing from condition data such as vibration, temperature, current, oil or pressure. The trigger differs, but both must pass through planning and scheduling.

Forecast preventive work

Do not wait until the PM due date to show demand. Expand projected meter as well as calendar PMs into the 13-week window. Define fixed versus completion-based frequency, early/late tolerance, seasons and non-operating days, job-plan sequences and behavior when prior work remains open. IBM’s PM materials describe time- and meter-based frequency, job plans and work-order scheduling; turn the plant’s rules into RFP scenarios rather than copying a product screen.

Give predictive candidates confidence bands

Do not treat a prediction 13 weeks away like committed work. Use bands such as Watch, Prepare and Commit. Watch verifies data; Prepare reserves a part and possible downtime; Commit advances a technically confirmed work order to Ready. These labels are a design example and make no promise about model accuracy.

For sensor and threshold design, see our guide to a condition-based maintenance system for Thai factories. The focus here is turning an alert or diagnosis into executable work with a part, skill and downtime window.

Predictive maintenance is not automatic shutdown. Verify data quality, operating mode, criticality, production and part lead time; retain accepted, rejected and deferred reasons. Feed findings back to model evaluation so predictive action joins the decision cycle.

ISO 55001:2024 and maintenance planning

ISO 55001:2024 specifies requirements for an asset-management system. It is not a feature list for maintenance-planning software, and buying software does not by itself demonstrate conformity or certification.

The changes described by ISO/TC 251 nevertheless relate to this planning cycle:

Decision-making and value: select jobs and downtime using criticality, risk, cost and opportunity.
Data and information: control the meaning and quality of labor, condition, part and configuration data.
Life cycle: plan modification, renewal and disposal as well as use and maintenance.
Predictive action: predict future parameter behavior to support intervention decisions.

ISO does not require a 13-week, weekly or daily cadence. What matters is that the organization applies, records and improves its own decision criteria. When contractors enter the plan, our equipment maintenance outsourcing guide for Thailand helps define ownership and return of history.

ISO 14224:2016 provides a standardized basis for collecting and exchanging reliability and maintenance data for equipment in the petroleum, petrochemical and natural-gas industries. Its sector scope matters: other factories should not claim direct applicability. Its principle of consistent equipment hierarchy, failure and maintenance definitions is still a useful primary reference when making planned and actual data comparable.

Use ISA-95 as a language for integration boundaries

ISA-95 can frame the boundary between business planning/logistics and manufacturing operations/control. When ERP provides parts/purchasing, MES/APS provides production and downtime candidates, OT provides meters/condition and CMMS provides availability/work plans, define the system of record for each item.

Allowing ERP, MES and CMMS to freely update the same downtime window produces conflict. Define owner, direction, timestamp, retry, deduplication and degraded operation for every interface. This does not mean ISA-95 prescribes a CMMS database schema.

Conditions that make the plan executable in Thailand

Multilingual work packages

Keep asset IDs, part numbers and permit codes language-neutral. Approve methods and hazards in Thai, English and Japanese where needed. Do not rely solely on automatic translation for safety instructions; control glossary, revision and approver. Daily changes must be visible in the assignee’s working language.

Parts kitting and import lead time

Inventory “1” may be reserved, under inspection, in another store or non-substitutable. Ready should distinguish physical confirmation, reservation, kitting and delivery to point of use. Make customs and holiday risk visible in Weeks 5–13. Store a confirmed date and risk note rather than one falsely precise arrival date.

Contractors and permits

Include availability, quotation/order, access documents, qualification, safety induction, insurance, tool entry and supervisor in one preparation gate. After work, capture actual hours, parts, readings, cause and recommendation in searchable work-order fields—not only an attached PDF.

Interpreting DIPROM’s current activity

DIPROM’s Division of Digital Industry runs industrial digitalization activities. On 20 August 2026 it posted an activity follow-up referring to AI Predictive Maintenance and AI Quality Control. This is an example of continued capability-building in Thailand; it does not establish any company’s subsidy eligibility or guarantee cost or failure reduction. Verify program conditions in current official information.

Demonstrate planning and scheduling in the RFP

Instead of another generic feature matrix, load one connected scenario and demonstrate the path from 13 weeks to the day:

Forecast 100 future PMs, 30 corrective jobs and five condition candidates across 13 weeks.
The counts 100, 30 and five are demo assumptions, not product-performance thresholds.
Compare weekly demand and available capacity by skill and show overload.
Update Weeks 2–4 after a part delay and production-downtime change.
Block work without Ready evidence from the weekly schedule.
Detect competition for one asset, location, LOTO owner or crane.
Freeze a version and audit before/after values and approval reason.
Redispatch for absence and emergency; return displaced work to the horizon.
Record actual hours, waits, parts, readings and added scope on mobile.
Create candidates to revise standard hours and job plans from variance.

Non-functional acceptance

Accept or reject concurrency, response time, volume, offline behavior, access, audit, backup, recovery, API limits, Thai display and local support hours. Test risks such as slow performance before weekly freeze, loss of today’s plan during an outage and inability to restore the baseline.

A 12-week implementation PoC

The following 12 weeks and 20 assets are illustrative assumptions, not a standard duration or benefit guarantee. The objective is to run at least one real planning-and-feedback cycle, not merely evaluate screens.

Weeks 1–3: baseline and ownership

Collect existing work for 20 assets and assign Planning, Scheduling and Dispatching responsibilities. Define skill capacity, downtime, part state, permits, required finish and Ready. Measure current schedule compliance and break-in reasons with fixed definitions.

Weeks 4–6: 13-week forecast and ready backlog

Expand PM, corrective and condition candidates into the horizon. Prepare parts, contractors and downtime for Weeks 2–4. Verify that unready reasons are searchable and the evidence for promotion to Ready is retained.

Weeks 7–9: weekly freeze and daily dispatch

Balance skill capacity and downtime, then freeze next week. Deliberately test absence, delayed part, extended production and emergency failure. Verify change reason, displaced work and due-date effect.

Weeks 10–12: actual variance and scale decision

Review labor, wait, part and outage variance. Update standard hours, job plans, capacity factors and Ready criteria. Evaluate data completeness, meeting time, field burden and auditability, then decide Scale, Revise or Stop.

Maintenance Planning System｜13-Week Scheduling in Thailand - figure 3

Build the business case from planning loss

Use plant variance rather than a vendor’s universal percentage. Suppose an 80-hour weekly schedule loses five hours to parts, four to production handover, three to method clarification and four to replanning: 16 hours total. Measure what Ready and freeze can actually avoid, then apply internal labor or downtime value.

The values 80, 5, 4, 3, 4 and 16 are illustrative assumptions—not industry averages or promised benefits. Prevent double-counting part waiting and downtime. Cost should include data cleanup, configuration, interfaces, devices, training, operation, local support, upgrades and exit migration as well as licenses.

Maintenance planning system FAQ

Is a 13-week rolling plan mandatory?

No. It is an example that spans a quarter. Cover the longest relevant part, contractor and downtime-preparation lead time while retaining a window the team can refresh weekly. A plant can combine a 26-week load view with a six-week detailed plan.

What is a ready backlog?

It is work that meets the site’s executable criteria for scope, skill hours, parts, tools, permits, downtime, drawings and contractors. Priority or part availability alone does not make work Ready.

Are emergencies prohibited after weekly freeze?

No. Respond to safety, quality and major outage, but record classification, approver, displaced work, reason and due-date impact. Freeze is a baseline that makes change learnable, not a ban on flexibility.

How complete must the maintenance-system asset register be?

For the planning scope, it must identify asset ID, hierarchy, criticality, owner, downtime condition and related PM/parts. Master-data and initiation design remain essential upstream; this article deliberately focuses on preparation through execution.

How does preventive versus predictive maintenance affect planning?

Preventive demand can usually be projected from calendar or usage. Predictive timing changes with condition and analysis, so confidence bands such as Watch, Prepare and Commit help prepare parts without prematurely committing work. Both pass through Ready and scheduling.

What should maintenance-record digitization prioritize?

For planning improvement, capture scheduled and actual starts/finishes, labor by skill, waiting reason, reserved/used parts, outage variance, added scope and failure finding. A PDF completion report alone is hard to feed into next week’s capacity and standards.

Should maintenance DX begin with AI?

Not necessarily. If ready backlog, freeze and actual variance do not work, more predictions create more unprepared backlog. Stabilize the path to execution, then connect AI/IoT for failure modes that are detectable, actionable and valuable.

Is there a benchmark for implementation cost or benefit?

No universal amount or reduction rate applies. Asset count, users, sites, interfaces, mobile, languages, data quality and local support change the case. Measure avoidable part wait, handover wait, replanning, break-ins and labor variance in a PoC.

Does ISO 55001:2024-ready software guarantee certification?

No. ISO 55001 covers policy, objectives, decision-making, roles, competence, data, operation, evaluation and improvement. Software can support consistent action and evidence, but the management system belongs to the organization.

Conclusion: close 13-week, weekly and daily horizons with actual variance

A maintenance planning system is not a calendar that places work orders on dates. Forecast demand against skills, parts and downtime in a 13-week rolling plan; create executable work through Definition of Ready; commit with a weekly freeze; respond through daily dispatch; and feed labor, waits, parts, outage and scope variance into next week. Preventive and predictive work become maintainable only when this operating cadence carries them into safe execution.

TOMAS TECH helps Thai factories define the 13-week rolling plan, ready backlog, weekly scheduling, daily dispatch and CMMS/EAM–MES–ERP–OT interfaces, from requirements through PoC even before a product is selected. Contact TOMAS TECH with your current weekly-planning method and main preparation constraints.