Datasets:

THULab
/

oil012_sample

license: cc-by-nc-4.0
task_categories:
  - tabular-classification
  - tabular-regression
  - time-series-forecasting
language:
  - en
tags:
  - tsfile
  - timeseries
  - synthetic
  - oil-and-gas
  - upstream
  - iot
  - rig-telemetry
  - predictive-maintenance
  - condition-monitoring
  - remaining-useful-life
  - sensor-fusion
  - anomaly-detection
  - xpertsystems
  - time-series
  - format:tsfile
pretty_name: OIL-012 — Synthetic Rig Sensor IoT Dataset (Sample) (TsFile)
size_categories:
  - 100K<n<1M
configs:
  - config_name: default
    data_files:
      - split: alarm_events
        path: alarm_events.tsfile
      - split: drilling_parameters
        path: drilling_parameters.tsfile
      - split: environmental_conditions
        path: environmental_conditions.tsfile
      - split: hydraulic_systems
        path: hydraulic_systems.tsfile
      - split: maintenance_records
        path: maintenance_records.tsfile
      - split: power_systems
        path: power_systems.tsfile
      - split: rigs_master
        path: rigs_master.tsfile
      - split: rul_labels
        path: rul_labels.tsfile
      - split: sensor_fusion_features
        path: sensor_fusion_features.tsfile
      - split: sensor_health
        path: sensor_health.tsfile
      - split: sensors_master
        path: sensors_master.tsfile
      - split: telemetry_streams
        path: telemetry_streams.tsfile
      - split: thermal_monitoring
        path: thermal_monitoring.tsfile
      - split: vibration_analysis
        path: vibration_analysis.tsfile

OIL-012 — Synthetic Rig Sensor IoT Dataset (Sample) (TsFile)

Apache TsFile version of xpertsystems/oil012-sample.

Overview

A free, schema-identical preview of XpertSystems.ai's enterprise drilling rig IoT telemetry dataset for predictive-maintenance ML, condition-monitoring ML, sensor-fusion modeling, and remaining-useful-life forecasting. The sample covers 30 rigs across 10 global basins and 8 rig types, with 134,622 rows including 94,745 sparse-format telemetry events linked across 15 tables.

Tables: 14 TsFile table(s), 134,622 rows in total.
- alarm_events: 278 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.002000 (UTC)
- drilling_parameters: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- environmental_conditions: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- hydraulic_systems: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- maintenance_records: 29 rows, 2026-01-01 00:10:00 → 2026-01-01 14:40:00 (UTC)
- power_systems: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- rigs_master: 30 rows, 1970-01-01 00:00:00 → 1970-01-01 00:00:00.029000 (UTC)
- rul_labels: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- sensor_fusion_features: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- sensor_health: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- sensors_master: 660 rows, 1970-01-01 00:00:00 → 1970-01-01 00:00:00.659000 (UTC)
- telemetry_streams: 94,745 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.656000 (UTC)
- thermal_monitoring: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)
- vibration_analysis: 4,320 rows, 2026-01-01 00:00:00 → 2026-01-01 23:50:00.029000 (UTC)

Schema (TsFile structure)

`alarm_events.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (8): alarm_id (string), rig_id (string), alarm_code (string), severity (string), subsystem (string), duration_sec (int64), anomaly_score (double), acknowledged_flag (int64)

`drilling_parameters.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (11): event_id (string), rig_id (string), operating_state (string), wob_klbs (double), torque_ftlb (double), rpm (double), rop_ft_hr (double), standpipe_pressure_psi (double), flow_rate_gpm (double), hook_load_klbs (double), stick_slip_index (double)

`environmental_conditions.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (6): rig_id (string), ambient_temp_f (double), humidity_pct (double), wind_speed_mph (double), environmental_stress_index (double), weather_impact_factor (double)

`hydraulic_systems.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (6): rig_id (string), pump_pressure_psi (double), hydraulic_pressure_psi (double), hydraulic_temp_f (double), flow_rate_gpm (double), cavitation_risk_score (double)

`maintenance_records.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (9): maintenance_id (string), rig_id (string), component_id (string), maintenance_type (string), subsystem (string), failure_mode (string), labor_hours (double), repair_cost_usd (double), post_maintenance_health_index (double)

`power_systems.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (6): rig_id (string), voltage (double), amperage (double), motor_current (double), load_pct (double), current_harmonic_distortion (double)

`rigs_master.tsfile`

Time (INT64, milliseconds) — synthetic row index over the stable source order; this table has no clock column.
TAG: none — the table is stored as a single device.
FIELD (13): rig_id (string), rig_type (string), operator_name (string), basin_name (string), country_region (string), rig_age_years (double), automation_level (double), offshore_flag (int64), hthp_flag (int64), environmental_stress_index (double), base_health_index (double), daily_operating_hours (double), commissioned_year (int64)

`rul_labels.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (6): rig_id (string), component_id (string), remaining_useful_life_hr (double), failure_probability (double), failure_horizon_days (int64), subsystem_health_score (double)

`sensor_fusion_features.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (6): rig_id (string), fusion_vector_id (string), anomaly_score (double), health_index (double), sensor_fusion_consistency (double), maintenance_priority_score (double)

`sensor_health.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (5): rig_id (string), calibration_score (double), sensor_drift_pct (double), packet_loss_pct (double), edge_latency_ms (double)

`sensors_master.tsfile`

Time (INT64, milliseconds) — synthetic row index over the stable source order; this table has no clock column.
TAG: none — the table is stored as a single device.
FIELD (9): sensor_id (string), rig_id (string), sensor_type (string), subsystem (string), unit (string), sensor_age_years (double), sampling_profile (string), calibration_score (double), edge_device_id (string)

`telemetry_streams.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (8): rig_id (string), sensor_id (string), sensor_type (string), subsystem (string), unit (string), sensor_value (double), quality_flag (string), operating_state (string)

`thermal_monitoring.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (5): rig_id (string), component_temp_f (double), bearing_temp_f (double), thermal_gradient (double), thermal_stress_index (double)

`vibration_analysis.tsfile`

Time (INT64, milliseconds) — the source event_ts column.
TAG: none — the table is stored as a single device.
FIELD (7): rig_id (string), vibration_rms (double), axial_vibration (double), torsional_vibration (double), harmonic_index (double), stick_slip_index (double), shock_event_flag (int64)

Conversion notes

One TsFile per source CSV (14 tables) for this drilling-rig telemetry sample; each table is stored as a single device, so the rig/well/tank/sensor ids stay ordinary FIELDs.
Time = the table's own timestamp column as INT64 milliseconds (event timestamps, daily record/production dates, or test dates). Master/inventory tables that carry no clock (rigs_master, sensors_master, wells_master, tank_master, the label tables and scada_telemetry) use a synthetic row index as Time.
No column is dropped. Rows that share a timestamp are ordered by a 1 ms step in stable source order; daily tables keep their exact calendar day.
Not converted: equipment_failures.csv — the file ships with a header only and zero data rows.

Usage

Install the Apache TsFile Python SDK (pip install tsfile) and read a converted file:

from pathlib import Path
from tsfile import TsFileReader

path = Path("alarm_events.tsfile")
with TsFileReader(str(path)) as reader:
    schemas = reader.get_all_table_schemas()
    print("tables:", list(schemas))
    table_name = next(iter(schemas))
    table = schemas[table_name]
    columns = [column.get_column_name() for column in table.get_columns()]
    print("columns:", columns)
    field_names = [
        column.get_column_name()
        for column in table.get_columns()
        if column.get_column_name() not in {"Time", "time"}
    ]
    if field_names:
        with reader.query_table(table_name, field_names[:3], batch_size=1024) as result:
            batch = result.read_arrow_batch()
            if batch is not None:
                print(batch.to_pandas().head())

Source & license

Original dataset: xpertsystems/oil012-sample
Source card title / publisher: OIL-012 — Synthetic Rig Sensor IoT Dataset (Sample)
License: cc-by-nc-4.0 (as declared by the original dataset).