Condition monitoring and reliability

Synthetic predictive maintenance datasets for condition monitoring

Predictive Maintenance records provide time-series signals for degradation, vibration, bearing temperature, electrical current, and remaining-useful-life research workflows.

Generate a dataset Explore original records

Engine maturity

Experimental Engine — Under Scientific Review

This domain is an early-stage research prototype. Its models, units, temporal behaviour, and event semantics are being audited. Outputs should not be treated as empirically validated observations.

engine 3.2.0 · maturity experimental · validation audit_pending

Research scope

The records support modelling experiments and teaching. They do not represent a calibrated copy of a specific industrial fleet or machine installation.

Degradation trajectories, vibration zones, bearing temperature, current, and RUL-oriented fields

Synthetic condition-monitoring data for anomaly detection and reliability studies

Column schema

Every record in this domain emits the same 14 columns . Types and ranges below are read from the generator itself, so they cannot drift from the files it produces.

Column Type Range
Timestamp datetime —
Machine_ID string —
Machine_Type categorical —
Temperature_C float 20, 150
Vibration_mm_s float 0.1, 30
Current_A float 0.5, 50
Pressure_kPa float 0, 1200
RPM integer 0, 5000
Noise_dB float 40, 120
Oil_Temp_C float 20, 120
Operating_Hours integer 0, 100000
Degradation_Index float 0, 1
RUL_Hours integer 0, 100000
Machine_Status categorical —

Generative models

Values come from explicit equations, not from a trained generative model. Each component below is documented in the methodology.

Degradation

D(t) = 1 - exp(-(t/L)^beta) with Weibull shape beta between 1.8 and 3.0.

Vibration

ISO 10816

Vibration severity zones coupled to the health index.

Coupled signals

Temperature, current and pressure driven by the same underlying degradation state.

Remaining useful life

RUL derived directly from the degradation trajectory, giving exact supervised targets.

Research uses

The common thread is known ground truth: the generator records the true state alongside the observation of it, which a real deployment cannot supply.

Remaining-useful-life regression with exact RUL labels

Early-fault detection benchmarking

Survival analysis across controllable degradation rates

Class-imbalance experiments with tunable failure frequency

Compared with measured datasets

Measured datasets in this area remain the reference for site realism. A generator answers a different question: what a method does when the true state is known and the scenario can be set deliberately. These are complements, not substitutes.

Measured dataset What it provides What a generated record adds
NASA C-MAPSS Turbofan engine degradation simulation with RUL targets. Tunable degradation shape and failure prevalence across many seeds.
MetroPT Metro train compressor sensor readings with failure reports. Exact degradation index at every step, not only the failure timestamps.
AI4I 2020 Synthetic milling-machine maintenance benchmark. Six coupled sensor channels plus a documented Weibull degradation model.

Scope and limits

What is currently established

  • — The generator runs and emits its declared column contract.
  • — Output is deterministic for a given domain, parameter set and seed.
  • — Scientific audit of its models, units, temporal behaviour and event semantics is still outstanding.

What it does not establish

  • — Site realism, or transfer to any specific real deployment.
  • — Calibration against a named site, sensor network, fleet or population.

Read the full research record →

Browse Predictive Maintenance generated records

Each card links to the original citable record. IoTSyn keeps that URL, its dataset file, and its citation guidance unchanged.

Open the full library

Frequently asked questions

What degradation model is used?

A Weibull degradation curve D(t) = 1 - exp(-(t/L)^beta) with shape parameter beta between 1.8 and 3.0, with vibration severity zones following ISO 10816.

Does this represent a specific machine fleet?

No. The records support modelling experiments and teaching. They are not a calibrated copy of any industrial installation.

Is remaining useful life labelled?

Yes. RUL_Hours and Degradation_Index are emitted alongside the sensor channels, giving supervised targets with exact ground truth.

Can failure frequency be tuned?

Yes. Scenario parameters control degradation rate, which makes the class balance of failure events adjustable for imbalance experiments.

Explore other synthetic-data topics

Synthetic Smart City datasets for urban IoT research

Open topic page →

Synthetic Smart Home sensor data for IoT research

Open topic page →

Synthetic Medical IoT datasets for health-data research

Open topic page →

Synthetic IIoT network traffic datasets for security research

Open topic page →

Synthetic connected-vehicle telemetry datasets

Open topic page →