Industrial Twin Lab
Build machine intelligence in the twin before trusting it in the machine.
A public research publication and deterministic concept demonstrator for digital twins, industrial AI, simulation, and engineering evidence.
Twin system / research boundary
Field update /
Data
Synchronized evidence
Context
Asset + operating limits
Decision and process orchestration
Validated recommendation
Actuation
Human-authorized execution
Phase 1 stops here
Never let AI perform its first experiment on the physical machine.
- Twin before intervention
- Human in command
- Evidence before deployment
Field update /
The twin is no longer only a model.
The Digital Twin Consortium framework connects data, context, orchestration, and actuation through a traceable digital thread.
Data
Synchronized evidence
Context
Asset + operating limits
Decision and process orchestration
Validated recommendation
Actuation
Human-authorized execution
Industrial Twin Lab studies the first three layers as concepts; live data acquisition and agent execution are future work. Actuation remains outside the Phase 1 boundary.
A discipline for machine intelligence
Twin Before Intervention
Physics Before Pure Correlation
Evidence Before Deployment
Demonstration asset
P-101 — Boiler Feed Water Pump
One fictional but engineering-realistic centrifugal pump connects the manifesto, Twin Capsule, experiments, fault laboratory, and AI Scientist. It is a synthetic teaching fixture, not a physical installation or source of plant evidence.
P-101 Twin Capsule
TWIN-P101-0.2.0
Asset identity
- Asset ID
- P-101
- Asset
- Boiler Feed Water Pump P-101
- Type
- Centrifugal Pump
- Driver
- Electric Motor
- Description
- Fictional centrifugal boiler feed water pump used only as a consistent Industrial Twin Lab demonstration asset.
- Hierarchy
- Enterprise → Fleet → Plant → System → Machine → Component
- service
- Boiler feed water
- configuration
- Single-stage centrifugal pump with electric-motor drive
- designFlow
- 240 m³/h
- designHead
- 112 m
- fluidDensity
- 1000 kg/m³
- hydraulicAssumptions
- Incompressible water at the stated reference density; equal inlet/outlet elevations and velocity heads. Pressures use the same absolute reference. Nominal values are rounded, not a validated pump curve.
Features
- pressure-ratioPressure RatioDischarge pressure relative to suction pressure.process; sources: suction-pressure, discharge-pressure
- flow-per-speedFlow / SpeedFlow normalized by rotational speed.process; sources: flow, speed
- power-per-flowPower / FlowElectrical power normalized by flow.process; sources: motor-power, flow
- bearing-de-delta-ambientDE temperature delta ambientDrive-end bearing temperature above ambient.physics; sources: bearing-de-temperature, ambient-temperature
- bearing-nde-delta-ambientNDE temperature delta ambientNon-drive-end bearing temperature above ambient.physics; sources: bearing-nde-temperature, ambient-temperature
- vibration-rmsVibration RMSRMS vibration for each bearing-housing channel. The fixture supplies scalar RMS levels, not waveform samples.vibration; sources: axial-vibration, radial-vibration
- vibration-kurtosisVibration KurtosisPer-channel waveform distribution-tail indicator; cannot be inferred from the two scalar RMS levels.vibration; sources: axial-vibration, radial-vibration
- twin-residualTwin ResidualMeasured value minus digital-twin prediction; unavailable until an independently validated prediction model exists.physics; sources: motor-power, flow, speed
- rolling-mean-30mRolling Mean 30mThirty-minute rolling temperature mean; requires a timestamped history absent from this single snapshot.temporal; sources: bearing-de-temperature
Signals
| Signal ID | Signal | Unit | Quantity | Location | Nominal fixture value |
|---|---|---|---|---|---|
| suction-pressure | suction pressure | bar | Pressure | Pump suction | 2.6 bar |
| discharge-pressure | discharge pressure | bar | Pressure | Pump discharge | 13.6 bar |
| flow | flow | m³/h | Volumetric flow | Discharge line | 240 m³/h |
| motor-current | motor current | A | Electrical current | Motor control center | 168 A |
| motor-power | motor power | kW | Electrical power | Motor control center | 101 kW |
| speed | speed | rpm | Rotational speed | Motor shaft | 2950 rpm |
| bearing-de-temperature | bearing DE temperature | °C | Temperature | Drive-end bearing | 68 °C |
| bearing-nde-temperature | bearing NDE temperature | °C | Temperature | Non-drive-end bearing | 64 °C |
| axial-vibration | axial vibration | mm/s RMS | Vibration velocity | Bearing housing | 1.9 mm/s RMS |
| radial-vibration | radial vibration | mm/s RMS | Vibration velocity | Bearing housing | 2.4 mm/s RMS |
| ambient-temperature | ambient temperature | °C | Temperature | Pump enclosure | 28 °C |
Operating envelope
- flowMinimum
- 180 m³/h
- flowMaximum
- 280 m³/h
- suctionPressureMinimum
- 2.2 bar
- dischargePressureMaximum
- 15 bar
- ambientTemperatureMinimum
- 5 °C
- ambientTemperatureMaximum
- 45 °C
Failure modes
- bearing-degradationbearing degradationProgressive degradation of bearing condition.Affected signals: bearing-de-temperature, bearing-nde-temperature, axial-vibration, radial-vibration
- impeller-degradationimpeller degradationLoss of hydraulic performance from impeller condition.Affected signals: flow, motor-power, discharge-pressure
- cavitationcavitationVapour-cavity formation associated with inadequate suction conditions.Affected signals: suction-pressure, axial-vibration, radial-vibration
- suction-restrictionsuction restrictionRestriction upstream of the pump inlet.Affected signals: suction-pressure, flow, motor-power
- seal-leakageseal leakageLoss of process fluid at the pump seal.Affected signals: flow, discharge-pressure
- motor-degradationmotor degradationReduced motor efficiency or electrical condition.Affected signals: motor-current, motor-power, speed
Model and uncertainty boundary
- Physics model implementation
- Not implemented
- Model validation
- Not validated in Phase 1
- Availability statement
- No physics model is implemented or validated in Phase 1.
- Uncertainty status
- Unquantified
- Uncertainty statement
- Uncertainty is unquantified because P-101 has no plant measurements or validated industrial evidence.
Limitations
- P-101 values and relationships are synthetic teaching fixtures, not plant measurements or validated industrial evidence.
- Read-only conceptual fixture; no operational technology connection exists.
- No result authorizes automatic control or setpoint changes.
Safety constraints
- Read-only conceptual fixture; no operational technology connection exists.
- No result authorizes automatic control or setpoint changes.
- Human engineering review is required before any physical-machine decision.
Provenance and disclosure
- Asset version
- ASSET-P101-0.2.0
- Twin version
- TWIN-P101-0.2.0
- Source
- Industrial Twin Lab fictional engineering fixture
- Origin
- Fictional / synthetic fixture
P-101 values and relationships are synthetic teaching fixtures, not plant measurements or validated industrial evidence.
P-101 Evidence Package
EXP-P101-BD-COMBINED-XGBOOST-WALKFORWARD
Conceptual demonstration — synthetic fixture results.
Hypothesis — not tested
Bearing-related features may reveal degradation before failure under held-out operating regimes.
- All seven metrics are authored teaching values, not measurements, fitted models, or benchmark results.
- Feature and algorithm adjustments prescribe the displayed scores; they cannot establish a winning model or feature set.
- Validation selections label illustrative scenarios; no dataset is split and no statistical validation is executed.
- The fixed seed 101 identifies this fixture family; no random sampling occurs. Tick 0 is a single snapshot, not elapsed plant time.
- Sensor counts and inference milliseconds are illustrative resource assumptions, not an executed feature pipeline or hardware timing.
Replay identity
- Behavior version
- BEHAVIOR-P101-0.2.0
- Experiment version
- EXPERIMENT-P101-0.2.0
- World version
- WORLD-P101-0.2.0
- Metric version
- METRICS-P101-0.1.0
- Snapshot tick
- 0 / fixture snapshot
Experiment record
- Model
- MODEL-XGB-0.1.0
- Model status
- experimental
- Dataset
- DATASET-P101-SYN-0.2.0
- Feature set
- combined
- Validation
- walk-forward
- Twin version
- TWIN-P101-0.2.0
- Asset version
- ASSET-P101-0.2.0
- Dataset version
- DATASET-P101-SYN-0.2.0
- Simulator version
- SIM-P101-0.2.0
- Feature pipeline
- FEATURES-P101-0.2.0
- Provenance model
- MODEL-XGB-0.1.0
- Code version
- ITL-PHASE-1-0.2.0
- Configured asset
- P-101
- Problem
- bearing-degradation
- Configured feature set
- combined
- Algorithm
- xgboost
- Configured validation
- walk-forward
- Random seed
- 101
- Timestamp label
- Synthetic fixture
Qualified evidence
| Evidence | Summary | Strength | Origin |
|---|---|---|---|
| EXP-P101-BD-COMBINED-XGBOOST-WALKFORWARD-SYNTHETIC-EVIDENCE | Deterministic synthetic fixture for conceptual comparison. | limited | Synthetic fixture |
Metric results
| Metric | Result |
|---|---|
| Detection Rate | 86 % |
| False Alarms | 0.9 alerts/month |
| Lead Time | 5 days |
| Inference Cost | 18 ms |
| Sensor Count | 11 sensors |
| Robustness | 80/100 |
| Explainability | 78/100 |
Operating regimes
- Nominal flow: 220–260 m³/h
- Rated speed: 2,900–3,000 rpm
- Ambient temperature: 5–45 °C
Limitations
- Synthetic fixture results do not establish plant performance.
- No control decision, alarm threshold, or maintenance action is authorized.
- Simulation fidelity and transferability require independent engineering validation.
Uncertainty, explainability, and provenance
- Uncertainty
- Illustrative uncertainty only; no confidence value is derived from an operating machine.
- Explainability
- XGBoost compared under Walk Forward.
- Dataset source
- Industrial Twin Lab synthetic experiment fixture
- Dataset disclosure
- This dataset is a deterministic synthetic fixture for conceptual comparison only; it is not plant data.
- Experiment source
- Industrial Twin Lab deterministic experiment fixture lookup
- Experiment disclosure
- Conceptual demonstration — synthetic fixture results. No model is trained or executed and no real plant data is used.
- Author agent
- Industrial Twin Lab synthetic fixture agent
Four separate authorities
- Train a model
- Not performed. A future training run requires its own authorization.
- Recommend an action
- No operational recommendation is issued by this fixture.
- Approve an inference model
- Requires separate engineering review for a named model version and use.
- Change a physical setpoint
- Outside Phase 1. An inference approval never grants this authority.
Evidence-backed machine knowledge.
The primary output is defensible machine knowledge: a reproducible body of evidence that an engineer can inspect, challenge, and govern—not a model score or autonomous act.
Continue to the manifesto