Before01
Company profile
- Hand-written rounds
- Fragmented records
- Limited MES visibility
AUTONOMOUS
Digital Twin
A digital twin that connects the physical and the virtual
SYSTEM ARCHITECTURE
System architecture
Visualization, data/model, transport and physical — four layers connected as one data flow.
04
Digital twin visualization
Visualization Layer
3D build · twin capabilities
3D Visualization
Entity Mapping
Rendering Pipeline
(HDRP)
Data Connector
(C# Scripts)
Texture Atlas
ML-Agents
Digital Twin
Real-Time Data
3D Physics Simulation
Cross-Platform
Quality Prediction
Smart Alarm
Industry Safety
03
Data / model layer
Data & Model Layer
Collect · store · link · analyze
Data collection
OPC-UA collection
NODE → SubmodelElement semantic tags applied automatically
IoT collection
Live state values bound in
Data storage
AASX package
JSON serialization
Version and history
Submodel version tags / LifeCycle state tracking
Data integration
AAS Registry
Supports enterprise-wide resource integration
API Gateway
REST API interface
Analytics / AI models
Digital twin model engine
Unity / syncs physical state with the virtual model
AI / ML submodels
Prediction and anomaly models registered
02
Transport layer
Transport Layer
Standardization · validation · encryption
OPC-UA based integration interface
Data collectionProtocol standardizationData validationServer-client structureEncryption / authentication
Modbus TCPOPC-UAPROFINETCC-LINKRS-232TCP IP SocketREST API
01
Physical layer
Physical Layer
Control · robots · identification · sensors
Equipment controllers
PLC
Production equipment
CNC
Machine tools
Robots
Cobot
Industrial robot
AGV / AMR
Identification
Barcode
RFID
Machine vision
IoT sensors
Power meter
Temperature/humidity sensor
Vibration sensor
PROCESS
Implementation process
Four stages, from data collection and infrastructure through model development and integration testing to rollout.
Data collection and infrastructure
- Installing on-site sensors and IoT
- Building the data pipeline
- Starting baseline data collection
Digital twin model development
- 3D physics modeling
- Real-time data integration
- Implementing the simulation engine
- Developing AI/ML prediction models
System integration and validation
- Integrating legacy systems
- Running integration tests
- User acceptance testing (UAT)
Rollout and enhancement
- Deploying to the production environment
- Training administrators
- Enhancement based on feedback
EFFECT
Benefits
What a digital twin leaves behind — in output, defect rate, working environment and cost.
- 01
Higher output
Running 24/365 without interruption and handling tasks far faster than a person does, cutting processing time.
Uptimeoutputtact time - 02
Fewer defects
Sensor data raises an alert the moment something drifts, and past data plus the virtual environment explain the cause so the same mistake does not repeat.
Defect rateanomaly detectionrecurrence - 03
A better working environment
Hazards in the areas people work in are simulated to raise safety, and emergency scenarios are rehearsed to sharpen the safety manual.
Hazard zonesemergency responseincidents - 04
Lower costs
AI reads each machine's energy consumption pattern to strip out waste, and optimal work paths and process order cut labor and logistics cost.
Energy uselabor costlogistics cost
CASE STUDY
Case study
A factory rebuilt virtually, with power data linked in and carried through to predictive maintenance.
Customer industry
Printed circuit board (PCB) manufacturing
After02
What was built
- Plant rebuilt in 3D
- Meter data per machine
- Predictive maintenance
Results
A digital twin that replaced manual rounds with live visibility