Industrial Solutions & Insights

Field notes from a real FDE engagement: the customer never asked for AI and had never heard of FDEs — they simply bought mes-line, our line-control system. Because mes-line ships with MCP support, AI came through the door with the product: when they added a KEYENCE device on the fly, AI pinned down the register values and had it collecting data the same day — faster than waiting for a point list. The brain is summoned from the cloud to work on-site, while production data never leaves the factory.

mes-line 2.0 is here: 106 MCP tools across 11 domains let AI handle data acquisition, monitoring, changeovers, dispatching and HMI development in plain language — while 34 high-risk actions stay behind human approval. AI does the work; humans hold the trigger.

Tsingley opens Remotely — free remote operations for manufacturers. Register and complete company verification, then manage up to 30 devices under one company account, with RDP tuned for small industrial screens and a Web terminal built for AI Agents. China + Singapore nodes, expanding.

We once branded mes-line as an "AI Agent". Looking back, that was a beautifully executed directional mistake. mes-line should expose its capabilities through MCP so AI can call it — not pretend to be a chatbot. This post is both a correction and a crash course in the 5 AI concepts we learned along the way.

Traditional maintenance knows two states: running and broken. But most equipment operates in the gray zone — still turning, efficiency declining. PHM digitizes health into H(t), a continuous scalar that drives adaptive capacity control from full-load to derated to forced shutdown.

CNN classifies known faults but isn't explainable. VAE detects unknown anomalies but can't classify them. The rule engine is explainable but has limited coverage. Tri-fusion isn't about "more is better" — remove any one path, and you create blind spots.

Inspection data, LOT numbers, CPK reports — each line stores its own in separate SQLite databases. Online does one thing: aggregates it all onto the factory server for unified query, analysis, and API output.

Traditional MES manages top-down — when the network drops, the line stops. mes-line inverts this: each line runs independently, recipe-driven, offline-capable, add lines without touching the architecture.

Production control, quality inspection, equipment maintenance, data aggregation — most factories end up with four separate silos. Tsingley puts them all on one architecture: Cloud-Network-Edge-Device.

Equipment inspection is the most important and most easily faked TPM activity. A real case: 52 inspection items moved from 4 paper pages to phone-scan walk-through with a confirmation chain and anomaly closure — inspection stops being a formality.

Tightening is the most common and most failure-prone process — too little torque means loosening, too much means fracture. Tightening traceability: real-time torque/angle/curve judgment, instant reject of defects, CPK process-capability analysis, and fully traceable data.

A digitalization solution for shoe-molding workshops: temperature monitoring for vulcanizers/freezers/ovens, pressure monitoring for heel-back/wall presses, output comparison across toe-lasting/lasting/needle-detection — health score + OEE score make traditional shoe machines achieve automatic production.

A robot-welding digitalization solution for rail-transit equipment manufacturing: the edge controller serves vision/back-gouging/welding/slag-removal subsystems; five software modules deliver cloud-network-edge-device integration from MES to terminal equipment. Phase 1 vision inspection covers appearance; Phase 2 online welding-quality monitoring (CNN+VAE+rule engine) covers process signals — dual mistake-proofing, zero missed defects.

The digital upgrade of equipment nameplates: one QR code, scan to see the full profile — documents, images, videos, maintenance records and technical support in one place. Built for equipment makers and integrators, ready to ship with the machine.

CPK is the process capability index; 6σ is the quality methodology. The ONLINE platform builds CPK/OEE/6σ analytics on real line data — full-dimension traceability across item/spec/LOT — upgrading quality management from inspection to process control.

A WMS+WCS stereo warehouse solution for flexible sheet-metal lines: automatic scheduling of laser cutting / CNC punching / robots / bending stations, recipe-based automatic in/out, stacker cranes and robot coordination — with auto-computed outbound coordinates and work-in-process coding as core technologies.

Most manufacturing LLM projects fail not because the model is weak but because the data is incomplete. The ONLINE platform prepares complete, clean, traceable data first — via MES-LINE auto-collection, auto-labeling and full-dimension management — so the LLM can actually drive decisions.

Quality monitoring for ultrasonic welding: unlike continuous condition monitoring of turbines, this captures the in-process signal of each single weld (power/amplitude/frequency/displacement) and judges OK/NG with CNN plus a rule engine.

Traditional MES manages plans and work orders but cannot reach equipment data — the digital factory becomes a castle in the air. TSingley's MES II uses a three-loop control model to perfect the execution layer (Device → Efficiency → Work Order), complementing — not replacing — customers' existing MES/APS.

The real-world headache: Mitsubishi, Siemens, Modbus and Fanuc CNC protocols mixed on one line. The TSingley DAQ engine turns 'protocol hell' into plug-and-play with 5-protocol simulation, register planning, live debugging and a clean data flow into MES-LINE.

Online quality monitoring for EV battery manufacturing: three-channel real-time acquisition of power/amplitude/frequency with three-way fusion judgment — closing the gap left by sampling inspection.

Almost all neural-network applications in manufacturing are machine vision. Diagnosis of vibration, temperature and current signals — the core of PHM equipment health — is nearly empty territory.

Equipment health management for wind turbines: 32 channels of vibration/temperature/RPM, CNN diagnosis + VAE anomaly detection + RUL remaining-life prediction, all inferring at the edge — turning reactive repair into predictive maintenance.

A digitalization solution for automotive electronics (ECU) plants: zero-intrusion acquisition without touching PLCs or lines, paperless inspection and changeover control, and an ONLINE + MES-LINE two-layer architecture from devices to group-level quality control.

A CNC line digitalization solution for machining shops: solving mixed-protocol device access, hard data collection, and manual work completion reporting — data is captured as production happens.

Mapping classic control theory's PID three loops (current / velocity / position) onto the factory: Device Loop → Efficiency Loop → Work-Order Loop. The TSingley MES II Industrial IoT three-loop control model is the mathematical foundation of plant-wide execution.