Rotating Component Test Data Analyser

Automated multi-channel sensor data analysis pipeline for rotating component test engineering.

Raw DAQ-style CSV data goes in. Filtered signals, metrics, FFT diagnostics, fault-signature detection, plots, and a structured PDF report come out.

Why This Project Matters

Manual test-data review is slow, inconsistent, and easy to repeat incorrectly. This project shows how a test engineer can move from raw sensor data to a repeatable pass/fail report with diagnostic evidence, including detection of a planted bearing-defect signature.

The same structure can be adapted to motor test benches, drivetrain endurance tests, rotating machinery validation, aerospace component tests, and other sensor-heavy workflows.

Features

• Multi-channel analysis for vibration, force, torque, and temperature
• 4th-order Butterworth low-pass filtering for noise reduction
• FFT-based vibration analysis with automatic dominant-frequency peak detection
• Configurable acceptance limits with pass/fail evaluation
• Automated PDF report generation with plots and engineering summary
• Optional real-time DAQ simulator with live FFT display
• Deterministic validation through scripts/validate.py and GitHub Actions

Headline Result

The synthetic test data simulates a rotating component at 3,000 RPM with 1x, 2x, and 3x shaft harmonics, sensor noise, and a planted bearing-defect signature near 187 Hz.

The pipeline correctly recovers:

Detection	Frequency	Meaning
1x shaft	50 Hz	Fundamental rotation rate
2x shaft	100 Hz	Shaft harmonic
3x shaft	150 Hz	Higher-order shaft harmonic
Bearing defect	187 Hz	Planted fault signature surfaced by FFT analysis

That is the core engineering value: distinguishing expected rotational harmonics from a diagnostic fault signature.

Pipeline

Raw sensor CSV
  -> Butterworth low-pass filtering
  -> RMS, peak, and mean metrics
  -> Single-sided real FFT
  -> Acceptance-limit evaluation
  -> Automated PDF report

Technical Details

Item	Value
Sampling rate	10,000 Hz
Filter	4th order Butterworth low-pass
Cutoff frequency	2,000 Hz
FFT method	Single-sided real FFT
Channels	Vibration, force, torque, temperature
Report output	PDF plus PNG plots

Results Snapshot

Deterministic synthetic-data run:

Check	Result
Samples analysed	50,000
Vibration RMS	1.6020 g
Force peak	271.9539 N
Torque RMS	22.0745 Nm
Temperature peak	59.5630 C
Dominant FFT peaks	50.0 Hz, 100.0 Hz, 187.0 Hz, 150.0 Hz
Diagnostic note	187 Hz bearing-defect zone activity detected

Visual Results

Time-Domain Signals

FFT Vibration Spectrum

Metrics vs Acceptance Limits

Repository Structure

.
+-- src/
|   +-- test_analyser.py      # Batch analysis and PDF report generation
|   +-- daq_realtime.py       # Real-time DAQ simulation and live FFT display
+-- docs/
|   +-- images/               # Generated plot previews for GitHub
|   +-- test_report_sample.pdf
+-- scripts/
|   +-- validate.py           # Deterministic validation checks
+-- .github/workflows/        # GitHub Actions CI
+-- requirements.txt
+-- LICENSE
+-- README.md

Installation

Linux / macOS

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Windows PowerShell

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt

Usage

Run the batch analyser with generated synthetic data:

python src/test_analyser.py

Run deterministic validation checks:

python scripts/validate.py

Validation confirms the channel limits, the 50 Hz shaft frequency, the 187 Hz bearing-defect signature, and the generated PDF/plot outputs.

Run with your own CSV file:

python src/test_analyser.py path/to/sensor_data.csv

Expected CSV format:

time_s,vibration_g,force_N,torque_Nm,temperature_C
0.0001,1.234,252.1,22.3,42.0

Run the real-time DAQ simulator:

python src/daq_realtime.py

Output

The batch analyser writes files to output/:

• test_report.pdf - structured report with metrics, FFT analysis, plots, and pass/fail verdict
• plot_time_domain.png - four-channel time-domain overview
• plot_fft.png - vibration spectrum with dominant frequency peaks
• plot_metrics.png - metrics compared against acceptance limits

A sample report is available at docs/test_report_sample.pdf.

Skills Demonstrated

• Signal processing - DAQ workflows, Butterworth filtering, FFT analysis, peak detection
• Test engineering - multi-channel measurement, acceptance limits, automated pass/fail, fault-signature detection
• Reporting automation - programmatic PDF report generation with ReportLab
• Verification engineering - deterministic seeds, CI-runnable validation, reproducible outputs
• Python toolchain - NumPy, Pandas, SciPy, Matplotlib, ReportLab

Relevant for rotating machinery testing, condition monitoring, NVH, motor and drivetrain testing, aerospace test engineering, and automotive endurance testing.

Author

Prajwal Bekal
M.Sc. Mechatronics and Cyber-Physical Systems, Deggendorf Institute of Technology
GitHub | LinkedIn

License

MIT - see LICENSE.