Condition monitoringManufacturing & robotics

Acoustic failure prediction

A single standoff microphone and a continuous-wavelet-transform diagnostic stack on a substation power transformer — time-frequency signal processing, a six-family feature bake-off, and a health-index / RUL prognostics pipeline with honestly measured uncertainty, reported with the misses left in.

DjiniousLabModelling & simulation

A grouped bar chart on a dark background comparing AUC by feature family (FFT, STFT, CWT, DWT and two others) across four fault classes — CWT-scale wins outright only on the partial-discharge class

CWT calibration error, worst tone (100 Hz–10 kHz): 0.14%CWT calibration error, worst tone (100 Hz–10 kHz)
partial-discharge burst localisation error: 0.0 mspartial-discharge burst localisation error
cooling-fan BPFO recovery error: 0.17 Hzcooling-fan BPFO recovery error
RUL 80% predictive-interval coverage (target ≥ 80%): 81.3%RUL 80% predictive-interval coverage (target ≥ 80%)

Hearing a transformer fail before it does.

A single standoff microphone, three meters from an oil-immersed power transformer's tank wall, feeding a continuous-wavelet-transform diagnostic stack. Where a stationary Fourier transform blurs a microsecond partial-discharge burst across an entire averaging window, the CWT localises it to the millisecond — and a health index with honestly measured uncertainty and a remaining-useful-life estimate turn that signal into a maintenance decision, with the classes where wavelets don't help reported as plainly as the one where they do.

One thing here is synthetic: the microphone.

Every recording in this program is generated, not captured — a physics model of five acoustic mechanisms (core magnetostriction, winding Lorentz force, and three fault modes) convolved through a validated three-mode tank-wall transfer path, at a declared 8.4 dB broadband SNR against switchyard ambient noise. Everything downstream of that recording — the transform, the features, the classifiers, the health index, the prognosis — is real code operating on those samples; replace the recording step with a real sensor and nothing downstream changes. The health index's own 90% split-conformal interval, the gate built specifically to test whether its stated uncertainty is honest, covered 82.5% of held-out cases (n=63) — reported as a FAIL in the verification notebook and the program dossier, not rounded up. The remaining-useful-life trajectories are simulated monthly health-index snapshots from the same physics model, not longitudinal field data, so the 81.3% RUL coverage figure above is a property of the estimator applied to a simulated trajectory, not a claim about real transformer wear. No claim on this page substitutes for field validation.

Four continuous-wavelet-transform scalograms on a dark background, one exemplar per class — normal, looseness, fan-bearing and partial-discharge — log10 CWT energy across 96 log-spaced scales from 60 Hz to 20 kHz — One scalogram per class from the same 96-scale continuous wavelet transform. The partial-discharge exemplar's wideband, non-stationary energy is exactly what a stationary Fourier transform cannot localise — the CWT recovers its burst to 0.0 ms, against a sub-millisecond target.

Where wavelets win — and where they don't.

AUC values from notebook 06's held-out bake-off: one Mahalanobis-distance classifier, scored per class against six feature families on the same 234-record dataset (117 records fit each family's per-class Gaussian, the other 117 held out and scored — never both for the same record). The FFT column shows the strongest Fourier-derived family, spectral-band — consistent with how the R-07 gate itself is scored, against the strongest Fourier family rather than the weakest. CWT-scale wins the partial-discharge class outright among all six families measured — but spectral-band trails it by only 1.3 AUC points, short of the ≥10-point margin the program's own gate requires, which is measured as a FAIL, not rounded up. On the three stationary classes, CWT does not win at all.