Factory & Manufacturing CMMS

On a plant floor, every minute of unplanned downtime has a direct cost: a late order, a stopped line, paid labour producing nothing. Yet many factories in the Kingdom still run maintenance on paper or spreadsheets that never tie a failure to a machine, never measure how long the repair took, and never reveal which assets consume the most downtime. A factory CMMS exists to close exactly that gap: turning maintenance from a reaction to breakdowns into a measured, improvable process.

Skyline's computerised maintenance management system (CMMS) is built in Saudi Arabia by a Saudi engineering firm. It is Arabic-native, not a global product superficially translated. More importantly, it ships genuine industrial reliability capability that runs inside the product, not just on a brochure page: per-work-order downtime tracking, automatic calculation of availability, MTBF and MTTR from work-order data, a transparent and documented predictive model, ISO 14224 failure codes, and spare-parts management with automatic reorder.

This page is written for the plant maintenance manager and the reliability engineer. We explain in detail how the system schedules preventive maintenance on a meter basis, how it separates downtime that actually stopped the line from downtime that did not, how the OEE factors are fed from this data, and how the predictive model flags which assets are at risk before failure — all with worked examples on the real fields in the product. For the full platform picture, see the complete CMMS guide.

The foundation in any factory is preventing failure before it happens. Skyline supports three preventive-maintenance engines working together:

1) Time-based preventive maintenance

Schedules that recur daily, weekly, monthly, quarterly or yearly, automatically generating a work order when due. When a factory is provisioned, the system arrives with a starter pack of ten industrial asset categories — Production Lines, CNC Machines, Conveyors, Compressed Air, Hydraulics, HVAC & Ventilation, Electrical Distribution, Material Handling, Tooling & Jigs, and Safety Systems — together with a 23-task preventive-maintenance template tied to those categories.

2) Meter-based (counter-based) preventive maintenance

Not every machine is serviced by time; many are serviced by usage. The system tracks meters per asset: runtime hours, cycle count, distance (km), energy (kWh), volume (litres/m³) or mass (kg). A trigger threshold (pm_trigger_at) is configured, and each time the meter crosses an interval boundary the system spawns a preventive work order automatically. A real example in the code: a chiller serviced every 500 runtime hours. The sweep runs on a periodic cron, handles multiple missed intervals at once (capped at ten) and uses row-level locking for idempotency so no duplicate work orders are created.

3) Operator rounds and inspection routes

The system turns ad-hoc operator rounds into structured early-warning data. Sequenced inspection routes are defined across several assets (route type: lubrication, inspection, meter reading, safety check, cleaning, security or general), and each stop on the route carries a checklist, meters to record, a photo-required flag and an is-critical flag. A fleeting observation on the line becomes a traceable, analysable record.

The difference between software that displays KPIs and software that actually computes them shows up in the formula. These are the four formulas exactly as implemented inside the product, with no ambiguity:

• MTTR — mean time to repair (hours): total repair hours ÷ number of repairs, where each repair time = completed-at minus started-at. Computed from completed corrective work orders only — preventive work orders are deliberately excluded so they do not distort the number.
• MTBF — mean time between failures (days): total operating period in days ÷ number of corrective failures in the period. The longer the period and the fewer the failures, the higher the reliability.
• Downtime % / Availability: downtime % = (sum of downtime hours ÷ total period hours) × 100. Availability = 100 minus downtime %.
• PM compliance %: (number of PM schedules due in the period that received a completed work order ÷ total due) × 100. This is the health indicator of the preventive programme itself.

All of these KPIs are generated automatically from work-order data with no manual calculation. Assets can be ranked in descending order by MTBF to identify the least reliable machines, so resources go to the assets that genuinely need them.

The phrase "predictive maintenance" is often used to sell an opaque AI model. Skyline deliberately does the opposite: a transparent, documented weighted model that computes a 0-to-100 health score for each asset from four signals we already collect inside the system. To be honest: this is not a deep-learning model, but a transparent weighted model that delivers the operational equivalent — a single risk number per asset plus a next-failure-window estimate.

The weights of the four signals:

• MTBF trend (weight 30%): the last 90-day MTBF versus the 12-month baseline. If the time between failures has shortened recently, something is degrading.
• Meter-reading slope (weight 25%): a linear regression on recent meter readings; a runtime-hour meter crossing 90% of its service interval bumps risk sharply.
• IoT telemetry breach frequency (weight 25%): IoT alerts opened in the last 30 days, capped at ten per asset so one chatty sensor cannot dominate.
• Age factor (weight 20%): asset age versus its expected useful life.

Health score = 100 minus (100 × weighted risk), banded as low risk (≥ 80), medium (60–79), high (40–59) and critical (below 40), with a next-failure date estimated by linear extrapolation. When an IoT sensor (vibration, temperature, pressure, current, rpm, humidity, power, runtime hours) breaches a configured threshold for an asset, the system opens a work order or alert automatically over an HMAC-secured webhook.

Reliability is incomplete without a standard classification of failures. The system ships an ISO 14224 failure-code catalogue of 39 default codes available to every tenant: 14 problem codes (what the technician observed: leak, noise, vibration, overheat, low flow, fails to start, unplanned stop, alarm, damage, wear, corrosion, control fault…), 14 cause codes (why it happened: mechanical failure, electrical failure, sensor drift, normal wear, material fatigue, corrosion, insufficient lubrication, contamination, human error, design defect, installation error, environmental, power-supply issue…), and 11 remedy codes (what was done: repaired in place, replaced, calibrated/adjusted, cleaned, lubricated, tightened, reset…). Each code carries an English label and an ISO 14224 category.

An important honesty note: the Arabic labels for these codes are Skyline's own explanatory translation, not shipped inside the product; the catalogue ships English labels and categories only.

At work-order closure, the problem / cause / remedy triplet linked to this catalogue is captured via the Skyline mobile app. Web closure, by contrast, captures structured fields for root-cause category, root-cause description, corrective action, preventive recommendation and a failure code. The digital sign-off at closure is an optional post-completion step available after the work order is completed — it is not enforced.

For proactive analysis, the system provides an FMEA table per asset or asset class: failure mode, effect on system, cause, severity (1–10), occurrence (1–10), detection (1–10), and a Risk Priority Number RPN = severity × occurrence × detection, with a recommended action, owner and due date. Failure modes are ranked by risk before they ever become actual downtime.

Much downtime is not in the repair itself but in waiting for the spare part. The system manages spare-parts and MRO inventory with configurable thresholds per item: minimum quantity, reorder quantity, maximum quantity, preferred vendor, lead-time days and an auto-reorder flag. When stock drops below the minimum, the system alerts or initiates a reorder.

When a part is linked to a work order, an auto-reserve service reserves the required parts against that work order or flags a shortage, with clear states: not applicable, pending, reserved, shortage, consumed. So the maintenance engineer knows before starting the repair whether the part is actually on hand, and downtime is shortened. For high-value assets, the system supports IAS 16 depreciation with five methods (straight-line, declining balance, double-declining, sum-of-years, units of production), so repair-versus-replace decisions are made against the asset's actual book value.

Reporting a fault must be as easy as taking a photo, or it will not be logged. The system supports multiple work-order intake channels, distinguished by a creation-source field: manual entry, auto-generated from PM, a mobile-app request, REST API, import, email, WhatsApp or IoT. An operator can scan a QR code on the machine to open an instant request linked to the correct asset, send the report over WhatsApp, or have an email turned into a work order via a rule engine. The system runs in the cloud or fully on-premise inside the factory's own servers for data sovereignty, and the CMMS interface is available in roughly 30 languages including Arabic.

This is the actual starter pack a new factory is provisioned with. These are not generic marketing phrases but realistic preventive-maintenance tasks with frequencies and technical references that a maintenance engineer can adopt as-is or adapt to their machines. Note: task frequencies and category names are fully configurable per site.

Note the depth here: CNC geometric-accuracy checks with a ball-bar and laser interferometer, hydraulic-oil particle-count analysis, motor insulation testing with a megger — these are tasks an actual maintenance engineer writes, not a marketing page. Each is generated as a work order when due, so no task is forgotten and no record is lost.

Skyline's CMMS is built in Saudi Arabia by a Saudi engineering firm, and it is Arabic-native rather than a translated product — an authentic Arabic interface for line operators, local support that understands the Saudi plant environment, and a fully on-premise option. Explore the industrial-city pages CMMS Riyadh, CMMS Dammam and CMMS Jeddah, our asset management service, and Skyline Care maintenance and support contracts.

The central idea in factory reliability is Overall Equipment Effectiveness (OEE). The industry-standard definition is OEE = Availability × Performance × Quality (an industry definition, not a Skyline output). Many tools display the number without ever tying it to maintenance data. We will be honest here about what Skyline actually measures and what it does not.

What does the system actually compute? Every work order carries downtime_start and downtime_end timestamps and a production_impact flag (did this failure actually stop the line or not). From the sum of downtime hours the system computes the downtime percentage = (sum of downtime hours ÷ total period hours) × 100, and from it Availability. This is the Availability factor of the OEE equation, computed from real maintenance data rather than estimates. For clarity and honesty: the system computes Availability and downtime%; the Performance and Quality factors are fed from production data, and the system does not claim a fully auto-computed three-factor OEE dashboard.

The example below shows how the work-order log becomes the Availability factor. Assume a filling line operates within a 720-hour production window (30 days) in a month, with the following downtime events recorded, each tagged with an ISO 14224 cause code:

Step-by-step calculation (as the system performs it):

• Downtime % = (14.0 ÷ 720) × 100 = 1.94%
• Availability factor = 100% − 1.94% = 98.06%
• MTTR (mean time to repair) = total repair time ÷ number of repairs = 14.0 ÷ 5 = 2.8 hours (from completed corrective work orders only)
• MTBF (mean time between failures) = total period in days ÷ number of corrective failures = 30 ÷ 5 = 6 days

This completes the backbone of reliability measurement: the Availability factor for the OEE equation from downtime data, and MTTR and MTBF from the corrective work-order log. Sorting downtime by ISO 14224 cause code immediately shows that mechanical failures and contamination consume the largest share of downtime hours — which is exactly where the preventive-maintenance budget should be directed.