1. Complete Functional Overview

MES Enterprise turns Dolibarr into a full Manufacturing Execution System (MES), giving industrial companies, factories, workshops and production units real-time control of their shop floor without any external MES software. It targets discrete and process manufacturing: metalworking, plastics, food processing, textile, chemical, pharmaceutical and mechanical industries.

1.1 Functional scope

·       Production scheduling with an industrial Gantt, workshop / machine / operator load and conflict detection.

·       Work order execution and real-time tracking across a seven-state lifecycle.

·       Shop floor terminal optimised for tablets, touch screens and industrial terminals.

·       Shop floor structure: workshops, production lines, workstations, machines and operators.

·       Production declaration with automatic roll-up of quantities, times and costs onto the work order.

·       Material consumption tracking with supplier lots and waste.

·       Lot and serial number management with automatic generation, QR code and barcode.

·       Full upstream and downstream traceability, including the product genealogy view.

·       In-process quality control driven by control plans (characteristic, target, tolerances, frequency, sampling).

·       Non-conformity management with root cause, corrective action, owner, due date and financial impact.

·       Scrap management by defect type, with cost valuation and Pareto analysis.

·       Machine downtime management across the seven standard downtime families.

·       Integrated maintenance: preventive, corrective and predictive.

·       Automatic OEE computation by machine, line, workshop and factory.

·       Industrial KPI cockpit across six families.

·       Real-time industrial dashboards and a twelve-report decisional reporting suite.

·       Local industrial AI: failure prediction, anomaly detection, scrap root-cause Pareto, scheduling optimisation and production forecast.

1.2 Business objects

The module ships 22 business objects, each with a list view, a card view, per-column search, a column selector and status management.

Domain	Objects
Shop floor	Workshop, Production line, Workstation, Shift
Machines	Machine / equipment, Downtime, Maintenance
Workforce	Operator
Production	Work order, Operation (routing), Schedule slot, Declaration, Material consumption
Traceability	Lot / serial, Traceability event
Quality	Control plan, Control, Non-conformity, Scrap, Defect type
Performance	OEE record, KPI record

 

2. Architecture

2.1 Engine architecture

·       Production Control Engine - work order lifecycle, routing operations and roll-up of declared quantities, times and costs.

·       Shop Floor Monitoring - real-time machine states, running work orders and a merged production event feed.

·       Machine Data Collection - run/stop hours, failure counts, IP address and protocol fields ready for machine feeds.

·       OEE Engine - availability x performance x quality, computed and aggregated with weighting.

·       Quality Control Engine - control plans, measurements, conformity verdicts and automatic non-conformity creation.

·       Traceability System - upstream/downstream genealogy from supplier lot to customer shipment.

·       Downtime Management - seven downtime families feeding availability, MTBF and MTTR.

·       Workforce Management - operators, skills, certifications, shifts, productivity and efficiency.

·       Industrial Dashboard - gauges, KPI cards, kanban, workshop map, timeline and alerts.

·       AI Manufacturing Analytics - deterministic local rules engine over the module's own data.

2.2 Native Dolibarr architecture

·       Every object extends CommonObject and uses the standard create / fetch / fetchAll / update / delete lifecycle.

·       Module descriptor modMesEnterprise (numero 517300) declares rights, menus, boxes, CSS and JS through module_parts.

·       table_element carries no llx_ prefix; the live prefix is resolved through $db->prefix() / MAIN_DB_PREFIX.

·       A schema self-heal runs after _load_tables(): Dolibarr only ever runs CREATE TABLE, so missing columns are added with ALTER TABLE on upgrade.

·       delete_menus() runs at the top of init() so menus never duplicate on re-activation.

·       Every menu URL is module-local, so it cannot collide with another module's identical menu row and abort activation.

·       All 456 language keys are namespaced under a Mes prefix: Dolibarr merges every active module's language file into a single translation table, so unprefixed keys would render another module's values.

2.3 Data model

Table	Purpose
llx_mes_workshop	Workshops: capacity, calendar, shift model, OEE target, hourly cost
llx_mes_line	Production lines: capacity, takt time, state, yield
llx_mes_station	Workstations: type, machine, cycle and setup time, state
llx_mes_machine	Machines: brand, model, serial, state, run/stop hours, MTBF, MTTR, OEE, IP, protocol
llx_mes_operator	Operators: badge, skills, certifications, shift, productivity, efficiency
llx_mes_shift	Shifts: hours, breaks, planned hours, headcount
llx_mes_order	Work orders: quantities, dates, priority, status, times, costs, lot
llx_mes_operation	Routing operations attached to a work order
llx_mes_schedule	Schedule slots with load and conflict detection
llx_mes_declaration	Production declarations: quantities, times, rate, labour and machine cost
llx_mes_consumption	Material consumption with supplier lot and waste
llx_mes_lot	Lots and serial numbers with QR code, barcode and quality status
llx_mes_trace	Traceability events, upstream and downstream
llx_mes_control_plan	Control plans: characteristic, target, tolerances, frequency, sampling
llx_mes_control	Quality controls with measured value and conformity
llx_mes_nonconformity	Non-conformities with root cause and corrective action
llx_mes_scrap	Scrap by defect type with cost valuation
llx_mes_defect	Defect type catalogue
llx_mes_downtime	Machine downtime by family, with duration and cost
llx_mes_maintenance	Maintenance interventions: preventive, corrective, predictive
llx_mes_oee	OEE records per machine and per day
llx_mes_kpi	Industrial KPI records with target, deviation and trend

 

 

3. Manufacturing Workflows

3.1 Work order lifecycle

A work order moves through seven states, stored language-independently in mes_order.of_status:

State	Meaning
planned	Created and scheduled, not yet released to the shop floor
launched	Released to the shop floor, ready to start
running	In production; declarations are being recorded
suspended	Paused, typically after a breakdown or a material shortage
done	Production finished
closed	Closed after final checks; consumed by reporting as completed
cancelled	Cancelled; excluded from performance figures

 

3.2 Shop floor execution flow

·       The operator selects the work order, then scans or selects the badge, the machine and the lot.

·       Start sets the status to running and stamps the real start date (only once).

·       Declare quantity creates a declaration, derives the good quantity and the real rate, then rolls the totals back onto the work order.

·       Declare scrap creates a scrap record linked to the work order, defect, machine, operator and lot.

·       Declare breakdown opens a downtime record, switches the machine to the breakdown state and suspends the work order.

·       Close sets the status to done and stamps the real end date.

3.3 Scheduling flow

·       Work orders are placed on the Gantt over a 7 to 60 day horizon.

·       Load is computed per workshop, machine and operator and compared to capacity; above 100 percent the resource is flagged as overloaded.

·       Conflicting slots are surfaced with their reason, for example a machine already assigned over the same window.

 

4. OEE Engine

4.1 Formula

OEE (Overall Equipment Effectiveness), known as TRS in French, is computed as:

OEE = Availability x Performance x Quality

Factor	Formula	Source data
Availability	run time / planned time	mes_oee.time_run, time_planned
Performance	real rate / theoretical rate	mes_oee.real_rate, theoretical_rate
Quality	good quantity / total quantity	mes_oee.qty_good, qty_total

 

4.2 Engine guarantees

·       Each factor is clamped to 100 percent: a machine cannot be more than 100 percent available, and running above nominal speed must not inflate OEE beyond its physical ceiling.

·       When the real rate was not measured, it is derived from the produced quantity divided by the run time.

·       Aggregation is weighted, never a naive average: availability is weighted by planned time, the rates by run time, and quality by quantity. A 30-minute record therefore cannot outweigh an 8-hour shift.

·       The same engine serves the dashboard, the OEE page, the KPI cockpit, the reporting suite, the AI analysis and the box widget, so every surface shows the same figure.

4.3 Reliability indicators

·       MTBF (Mean Time Between Failures) = cumulated run time / number of breakdowns.

·       MTTR (Mean Time To Repair) = cumulated repair time / number of breakdowns.

·       Both are computed from the machine's downtime records filtered on the breakdown family.

 

5. Quality Management

5.1 Control plans

·       A control plan defines the characteristic, the control type, the unit, the target value, the minimum and maximum tolerances, the frequency, the sample size and the method.

·       Plans are attached to a product and optionally to a workshop.

5.2 Controls and non-conformities

·       A control records the measured value, the sample size, the conforming and rejected counts, a photo, a comment and the validating user.

·       Conformity is evaluated against the plan's tolerance band.

·       A failing control automatically opens a non-conformity carrying severity, root cause, affected quantity, corrective action, owner, due date and financial impact.

5.3 Scrap

·       Scrap is declared against a defect type from the catalogue, with cause, machine, operator, lot and work order.

·       Total cost is valued as quantity x unit cost; a recoverable flag separates reusable scrap from outright loss.

·       Pareto analysis with cumulative percentage isolates the vital few defects that carry most of the volume.

 

6. Traceability System

·       Search accepts a lot number, a serial number, a barcode or a QR code, using an exact match first and then a partial match.

·       Upstream: raw materials consumed, supplier lots and parent lots.

·       Work order: status, quantities and real start / end dates.

·       Operators and machines that touched the lot, with the timestamp of every declaration.

·       Quality controls performed, with the measured value and the conformity verdict.

·       Downstream: traceability events and the customer shipment reference.

·       Traceability events carry a direction (upstream / downstream), an event type, a date, a linked reference and a quantity.

 

7. Machine Monitoring

Machine state	Meaning
running	In production
idle	Available, not currently producing
setup	Being set up or changed over
maintenance	Under maintenance
breakdown	Failed, production stopped
offline	Out of service

 

Downtime families:

Family	Planned	Typical cause
breakdown	No	Tool failure, spindle stop
maintenance	Yes	Scheduled preventive maintenance
setup	Yes	Start-of-run setup
changeover	Yes	Series or mould changeover
nomaterial	No	Material shortage
nooperator	No	Waiting for a qualified operator
quality	No	Stopped on a quality drift

 

 

8. AI Manufacturing Analytics

The analytics engine is local by default: every insight is derived from the module's own data with deterministic industrial rules, so no production data ever leaves the instance. The setup page can point MESENTERPRISE_AI_PROVIDER at OpenAI or a self-hosted Ollama for free-text generation; the analytics themselves remain local either way.

8.1 Failure prediction

·       A risk score from 0 to 100 combines the three signals a maintenance engineer actually watches: a short MTBF (fails often), a long MTTR (hard to repair) and a high recent breakdown count.

·       The MTBF threshold is calibrated against the fleet median, so the rule adapts to the plant rather than to a hard-coded constant.

·       A machine currently in the breakdown state and a stop-time ratio above 15 percent both raise the score.

8.2 Anomaly detection

·       The last 7 days are compared with the previous 21 on OEE, scrap rate and downtime minutes.

·       An OEE drop of 5 points or more is attributed to whichever of the three OEE factors moved the most, which is what turns a number into an actionable finding.

·       A scrap rate 40 percent above its baseline, and downtime 50 percent above the weekly average, are reported.

8.3 Other analyses

·       Scrap root causes: Pareto with cumulative percentage and cost.

·       Scheduling optimisation: late work orders, conflicting slots and overloaded resources.

·       Production forecast: a 7-day projection from the trailing 28-day daily mean adjusted by the recent trend.

·       Natural-language questions are routed by keyword to the matching analysis, for example: Why did line 2 OEE drop this week?

 

9. API & Integration

9.1 Dolibarr module integration

Dolibarr module	Integration
Products	Manufactured products and raw materials are native Dolibarr products
Third parties	Work orders carry the customer; demo customers are native Societe records
Stock / Warehouses	Material consumption references native warehouses
Users	Managers, technicians and validators are Dolibarr users; operators link to a user account
Projects / Agenda	Work orders can be reconciled with projects and planning
GMAO / CMMS	Machines and downtime align with the maintenance module's asset model
QHSE	Non-conformities and corrective actions feed the quality system
BI / Dashboard Builder	All MES tables are exposed for cross-module reporting
GED / OCR / e-Signature	Control reports and maintenance reports can be filed and signed

 

9.2 Industry 4.0 readiness

·       Every machine carries an IP address and a protocol field (OPC-UA, MQTT, Modbus TCP), so a collector can be attached without a schema change.

·       The declaration, downtime and OEE tables are the natural write targets for a machine feed: the shop floor terminal and a future collector share the same roll-up logic.

·       The real-time page is a wall board that already refreshes on a timer; an MQTT or OPC-UA push can drive it later.

·       Lots carry QR codes and barcodes for scanner-based identification.

9.3 Exports

·       Every one of the twelve reports exports to CSV, Excel and Word, and is printable.

·       Exports are streamed from the same data structure as the on-screen table, so an export can never disagree with the screen.

·       CSV is emitted with a UTF-8 BOM so Excel detects the encoding correctly.

 

10. Security

10.1 Permission model

Eight functional right groups produce 23 permissions (read / write / delete per group; reporting is read-only; plus a module administration right). Read and write are granted by default so the module is usable immediately after activation, while delete and admin start closed.

Group	Covers
production	Work orders, operations, schedule slots, declarations, consumption
shopfloor	Workshops, lines, workstations, shifts
machine	Machines, downtime, maintenance
operator	Operators
quality	Control plans, controls, non-conformities, scrap, defect types
traceability	Lots, traceability events
performance	OEE records, KPI records
reporting	Industrial reporting (read only)

 

10.2 Security measures

·       Every page checks its right group before rendering; unauthorised access raises accessforbidden().

·       All objects are multi-entity aware (ismultientitymanaged) and filtered through getEntity(), so MultiCompany installations stay isolated.

·       All write actions are CSRF-protected with newToken(); the shop floor terminal posts a fresh token per action.

·       SQL values are escaped through the Dolibarr database layer, and identifiers are cast to int or float before interpolation.

·       Output is escaped with dol_escape_htmltag(); the reporting exports strip tags.

·       Creation, modification and author are stamped on every record (date_creation, tms, fk_user_creat, fk_user_modif), giving a native audit trail.

·       GDPR: operators hold professional data only (name, badge, job, skills, certifications); no personal or sensitive data is collected. Demo records are tagged and removable in one click.

 

11. Technical Architecture

Item	Value
Technical name	mesenterprise
Descriptor	modMesEnterprise
Module numero	517300
Version	1.0.0
Family	MES / Industrie 4.0
Table prefix	llx_mes_
Business objects	22
Database tables	22
Permissions	23 across 8 right groups
Menu entries	1 top + 32 left
Language keys	456 per language
Languages	French, English, Spanish, Italian, German
Dolibarr	16.0 and above
PHP	7.1 and above, PHP 8.x compatible
Database	MySQL / MariaDB, PostgreSQL compatible
Licence	GNU GPL v3
Publisher	DoliResources - www.doliresources.com

 

11.1 Performance

·       Every foreign key column is indexed; index names are kept well under the 64-character limit through short per-table tokens.

·       Dashboard and KPI aggregates are computed with SQL SUM / COUNT rather than by loading object collections.

·       List queries are paginated through plimit() and sorted through the database layer.

·       Computed rates and money values are explicitly rounded, so instances configured with a high decimal precision do not print eight decimals.

·       The CSS and JS are served through module_parts with a public cache limiter.

11.2 Front end

·       Industrial charter: industrial blue, technology cyan, performance green, alert orange, breakdown red and factory grey.

·       Responsive design: the KPI grid, chart rows and terminal buttons reflow below 900 px; a print stylesheet hides navigation and shadows.

·       Charts are rendered with the native Dolibarr DolGraph layer, so they follow the instance's chart engine.

·       OEE gauges are pure CSS conic gradients: no JavaScript charting dependency.

·       Touch targets on the shop floor terminal are at least 76 px high so they stay usable with gloves.

·       Icons are Font Awesome 5 Free names only, matching what Dolibarr 16 to 23 bundles.

11.3 Demonstration dataset

·       A versioned dataset guard (DATASET_VERSION) re-seeds the demo automatically when the module is upgraded, which a plain boolean guard would never do.

·       Every demo record is tagged with the import key MESDEMO; removal deletes only tagged rows and never touches real data.

·       The generator is deterministic: a re-seed reproduces the same dataset, so screenshots and documentation stay stable.

·       The dataset covers 3 workshops, 6 lines, 12 workstations, 14 machines, 12 operators, 4 shifts, 26 work orders, 108 operations, 71 declarations, 18 lots, 34 controls, 30 downtime records, 338 OEE records and more.

 

Copyright (C) 2026 DoliResources - www.doliresources.com - Distributed under the GNU GPL v3 licence.