Forgis/Mechanical-Components

Mechanical Components Vibration Dataset

Comprehensive multi-source mechanical vibration dataset for training cross-component fault diagnosis and prognostics models. Designed for the Mechanical-JEPA project.

Total: ~12,000+ samples | 9.5 GB | 16 sources | 5 component types

Quick Start

from datasets import load_dataset

bearings = load_dataset("Forgis/Mechanical-Components", "bearings", split="train")
gearboxes = load_dataset("Forgis/Mechanical-Components", "gearboxes", split="train")
sources = load_dataset("Forgis/Mechanical-Components", "source_metadata", split="train")

Two-Level Schema

source_metadata (16 entries): One row per source dataset with constant properties. bearings/gearboxes configs: Per-sample data linked via source_id foreign key.

Dataset Sources

Bearings (~10,000 samples from 10 sources)

Source	Samples	Component	Sensors	Unique Value
CWRU	40	Ball bearing	Vibration	Standard benchmark
MFPT	20	Ball bearing	Vibration	Variable load
FEMTO	3,569	Ball bearing	Vibration, temperature	Run-to-failure (RUL)
Mendeley	280	Ball bearing	Vibration	Speed transitions (action-conditioning)
XJTU-SY	1,370	Ball bearing	Vibration	Run-to-failure (RUL)
IMS/NASA	1,256	Ball bearing	Vibration	Run-to-failure (RUL)
Paderborn	384	Ball bearing	Vibration, current	Real + artificial faults
MAFAULDA	800	Shaft+bearing	Vibration, acoustic, tachometer	Imbalance, misalignment (shaft faults!)
Ottawa	180	Ball bearing	Vibration, acoustic	Cage faults, 3 health stages
SCA Pulp Mill	2,663	Industrial bearing	Vibration	Real industrial data
VBL-VA001	800	Shaft+bearing	Vibration (triaxial)	Misalignment, unbalance
SEU	140	Drivetrain bearing	8-ch (motor+gearbox)	Cross-component rig

Gearboxes (~1,225 samples from 4 sources)

Source	Samples	Component	Sensors	Unique Value
OEDI	20	Spur gear	Vibration (4-ch)	Healthy vs gear crack
PHM 2009	109	Spur gear	Vibration, tachometer	Challenge data
MCC5-THU	956	Spur gear	Vibration	Speed/load transitions
SEU	140	Planetary+parallel	8-ch (motor+gearbox)	Cross-component rig

Component Types Covered

Component	Sources	Fault Types
Bearings	CWRU, MFPT, FEMTO, Mendeley, XJTU-SY, IMS, Paderborn, Ottawa, SCA	inner_race, outer_race, ball, cage, compound, degrading
Gears	OEDI, PHM2009, MCC5-THU, SEU	gear_crack, gear_wear, missing_tooth, tooth_break
Shafts	MAFAULDA, VBL-VA001	imbalance, misalignment_horizontal, misalignment_vertical
Drivetrains	SEU	Combined motor+gearbox+bearing from single rig
Industrial	SCA Pulp Mill	Naturally occurring faults in real machinery

Sensor Modalities

Modality	Sources	Channels
Vibration (accelerometer)	All 16	1-8 channels per sample
Motor current	Paderborn, Mendeley (partial)	2-3 phase current
Acoustic (microphone)	MAFAULDA, Ottawa	1 channel
Tachometer	MAFAULDA, PHM2009, OEDI	1 channel
Temperature	FEMTO	Scalar in slow_signals
Torque	SEU, MCC5-THU	1 channel

Key Features

• 412 transition samples for action-conditioning (Mendeley speed ramps + MCC5-THU speed/load transitions)
• Episode/RUL fields for prognostics (FEMTO, XJTU-SY, IMS, SCA)
• Real industrial data from SCA pulp mill (not lab)
• Shaft faults (imbalance, misalignment) from MAFAULDA and VBL-VA001
• Acoustic data from MAFAULDA and Ottawa (microphone alongside accelerometer)
• Cross-component drivetrain data from SEU (motor+gearbox+bearing on single rig)

Per-Sample Schema

{
    "source_id": "cwru",            # FK to source_metadata
    "sample_id": "cwru_105",
    "signal": [[0.1, 0.2, ...]],    # (n_channels, signal_length)
    "n_channels": 2,
    "channel_names": ["DE_accel", "FE_accel"],
    "channel_modalities": ["vibration", "vibration"],
    "health_state": "faulty",       # healthy | faulty | degrading
    "fault_type": "inner_race",
    "fault_severity": None,
    "rpm": 1750,
    "load": 2.0,
    "load_unit": "hp",
    "episode_id": None,             # For run-to-failure
    "episode_position": None,       # 0.0 to 1.0
    "rul_percent": None,            # Remaining useful life
    "is_transition": False,         # Speed/load change
    "transition_type": None,        # ramp_speed | ramp_load
}

v2 Training-Ready Config (Planned)

A standardized config for direct model training:

• Fixed sampling rate: 12,800 Hz
• Fixed window: 16,384 samples (1.28 seconds)
• Vibration-only (single modality)
• Per-sample instance normalization
• Source-disjoint train/val/test splits

v2 Training-Ready Config (LIVE)

Standardized for direct model training. All sources resampled to common format.

# Load training-ready data (all splits in one, filter by 'split' column)
v2 = load_dataset("Forgis/Mechanical-Components", "v2_train", split="train")
v2_train = v2.filter(lambda x: x["split"] == "train")  # 20,143 samples
v2_val = v2.filter(lambda x: x["split"] == "val")      # 1,332 samples
v2_test = v2.filter(lambda x: x["split"] == "test")    # 6,363 samples

Parameter	Value
Sampling rate	12,800 Hz
Window length	16,384 samples (1.28 seconds)
Channels	1 (primary vibration)
Normalization	Per-sample instance norm
Splits	Source-disjoint (train/val/test)

Train (12 sources): CWRU, MFPT, FEMTO, XJTU-SY, IMS, OEDI, PHM2009, MCC5-THU, SEU, MAFAULDA, VBL, SCA-train Val (2 sources): Paderborn, Ottawa Test (2 sources): Mendeley, SCA-test

Citations

Please cite the original datasets. See source_metadata config for full citations per source.