THULab/praxis_mcc5_thu_motor

MCC5-THU Motor Replay Dataset (TsFile)

Apache TsFile version of narteybrown/praxis-mcc5-thu-motor.

Overview

Motor test-bench recordings from the MCC5-THU bearing dataset: vibration and three-phase current channels plus key-phase, torque and speed signals, captured at 1 s resolution across many fault-type / operating-mode test runs.

• Tables: 2 TsFile table(s), 25,920 rows in total.
  • mcc5_thu_motor_train: 20,700 rows, 1970-01-01 00:00:00 → 1970-01-01 00:01:29 (UTC)
  • mcc5_thu_motor_val: 5,220 rows, 1970-01-01 00:00:00 → 1970-01-01 00:01:29 (UTC)

Schema (TsFile structure)

mcc5_thu_motor_train.tsfile

• Time (INT64, milliseconds) — the source time_s column.
• TAG: unit_id — device dimension; query one device with WHERE unit_id=....
• FIELD (15): cycle (int64), source_archive (string), source_file (string), speed_key_phase (double), torque_nm (double), motor_vibration_x_0_1g (double), motor_vibration_y_0_1g (double), motor_vibration_z_0_1g (double), motor_current_a_0_1a (double), motor_current_b_0_1a (double), motor_current_c_0_1a (double), fault_type (string), operating_mode (string), torque_setpoint_nm (double), rpm_setpoint (double)

mcc5_thu_motor_val.tsfile

• Time (INT64, milliseconds) — the source time_s column.
• TAG: unit_id — device dimension; query one device with WHERE unit_id=....
• FIELD (15): cycle (int64), source_archive (string), source_file (string), speed_key_phase (double), torque_nm (double), motor_vibration_x_0_1g (double), motor_vibration_y_0_1g (double), motor_vibration_z_0_1g (double), motor_current_a_0_1a (double), motor_current_b_0_1a (double), motor_current_c_0_1a (double), fault_type (string), operating_mode (string), torque_setpoint_nm (double), rpm_setpoint (double)

Conversion notes

• The source ships a train and a val split; each becomes its own TsFile.
• time_s is a per-run stopwatch (0-89 s at 1 s resolution) that restarts for every test run, so unit_id is declared as the TAG (device) dimension: each of the 230 test runs is one TsFile device with its own 90-point timeline, and WHERE unit_id='...' returns a single run. Without the TAG 230 rows would share every second and the time axis would drift by ~23% of its span.
• Time = round(time_s * 1000) ms. cycle, fault_type, operating_mode, the setpoints and all vibration/current channels are FIELDs. No column is dropped.

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("mcc5_thu_motor_train.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: narteybrown/praxis-mcc5-thu-motor
• Source card title / publisher: MCC5-THU Motor Replay Dataset
• License: mit (as declared by the original dataset).