Industrial Automation glossary

Automated camera inspection of PCBs or finished parts

Used mainly in electronics — checks solder joints, missing components, polarity before parts reach the test station

A programmatic channel that lets two systems talk to each other

A device that converts an electrical signal into physical motion — cylinder, motor, valve

Light/screen system showing line status — everyone sees normal vs. abnormal instantly

The rectangle AI draws around a detected object or defect

Conveyor belt — the backbone of any automated production line

Time to produce one part from start to finish

Aligning a vision system or sensor to a reference — critical for measurement accuracy

How confident the AI is in its classification (0–1) — used to set decision thresholds

A 2D barcode for unit-level traceability — encodes more than a 1D barcode

Time a machine is stopped — split into planned (PM) and unplanned (breakdown)

Enterprise resource planning — systems like SAP, Oracle, Odoo

Processing data on the device at the site instead of sending it to the cloud

AI lets a defect pass through (worse than false positive — bad parts ship to customer)

AI flags a good part as defective — wastes yield, tune the threshold

The touch screen operators use to control a machine

Hardening / annealing of metal — temperature and time must be tracked precisely

Input/output pins on a PLC or controller — wired to sensors and actuators

International standard for integrating enterprise systems with control systems

The data format most modern systems use to exchange information — human-readable

Visual signal system for triggering production / replenishment — from Toyota, used worldwide

Production philosophy that eliminates waste — pioneered by Toyota Production System

Manufacturing execution system — the middle layer between PLCs below and ERP above

A protocol for PLCs, sensors, and controllers to talk — old but supported everywhere

A lightweight messaging protocol — widely used in IoT and edge devices

Read text from images automatically — used for expiry dates, serial numbers, etc.

Overall Equipment Effectiveness = Availability × Performance × Quality

World-class target is ≥ 85% — typical factories sit at 60–70%

Modern protocol for machine-to-machine communication — secure and cross-vendor

An industrial computer that runs machines — runs 24/7 without crashing

Popular brands in Thailand: Mitsubishi (FX, Q), Siemens (S7), Omron (CP, NJ), Allen-Bradley

A robot/tool that picks parts from A and places them at B — used in assembly

Strict defect rate unit — e.g. 50 PPM = 50 defects per million parts

Design that prevents human error — e.g. a USB that can only be inserted one way

Return on investment — often expressed as months to payback

Responding within a guaranteed time window — not necessarily the fastest, but on time

Supervisory control and dashboard system — sits above the PLC layer

High-precision motor — fine control of position and speed

A factory where every layer (sensor → PLC → MES → ERP → AI) is connected and logged

A method to push defects below 3.4 PPM — DMAIC is the core framework

Target time per part required to meet customer demand

The ability to trace every part back to raw materials and production conditions

Total Productive Maintenance — proactive equipment care to cut unplanned downtime, everyone participates

Dedicated hardware that runs vision algorithms + ingests camera signals — not a regular PC