Cases/CROSS/C18
Cross-Domain SignalsGrid substations / transformer plants / condition-based maintenance

Transformer Core Hum Online Monitoring

01

Pain Points

  • !The hum is a fundamental plus a string of harmonics; loosening/bias alters the harmonic structure and modulation, and manual listening can't catch the subtle changes.
  • !Substations are unattended and need 7×24 online monitoring with automatic early warning on deviation.
  • !Ambient and seasonal temperature shifts interfere, so judgment must be relative to each unit's own baseline rather than an absolute threshold.
02

The Tinia Approach

The modulation spectrum characterizes the hum's harmonic structure + TNR prominence + per-unit baseline + Z-Score deviation early warning, judged against each unit's own baseline to resist environmental interference.

Frequency weightingModulation-spectrum analysisTNR prominencePer-unit baselineZ-Score early warning
03

Nodes Used

Frequency Weightingweighting_filter
Suppress irrelevant bands
Modulation-Spectrum Analysismodulation_spectrum
Characterize hum harmonics / modulation structure
TNRtnr
Narrowband hum prominence
Baseline Statisticsbaseline_stats
Model each unit's normal state
Z-Score Anomaly Detectionzscore_anomaly
Baseline-deviation early warning

Core nodes; add or remove steps as needed for your data and standards — the node graph is always editable.

04

Expected Output

  • Trend over time of the hum's harmonic structure and modulation depth.
  • Deviation early warning relative to each unit's own baseline, resistant to environmental / seasonal interference.
  • A 7×24 online monitoring workflow streamable via SDK.
Reference Standards / Engineering PracticeIndustry practice (transformer condition-based maintenance / acoustic monitoring)

Want to run this workflow on your own data?

Request a Trial
CROSS

Other cases in this domain