Have the engineers and consulting engineering firms working for utility institutions such as Eskom and the major metropolitan utilities truly examined the fundamental electrical causes behind recurring transformer failures, cable burn-offs, and unexplained control system malfunctions?
Or are we still treating symptoms while ignoring the waveform in front of us?
In my previous article — Fundamentals of Power Quality — I explained how harmonic distortion, voltage imbalance, and waveform instability are no longer rare anomalies. They are engineered consequences of modern electrical networks dominated by non-linear loads, inverters, drives, and power electronics.
Yet we continue to hear phrases like:
“Unknown cause.”
“Unexplained insulation failure.”
“Cable fault for no obvious reason.”
Let’s ask the harder question.
🔥 Why Do Transformers Suddenly Fail — Sometimes Catastrophically?
Transformers do not simply “go up in flames” without physics being involved.
Harmonic currents increase eddy current losses and stray flux heating. Triplen harmonics overload neutrals. Voltage imbalance causes negative sequence currents that create additional rotor heating in rotating machines and stress upstream supply equipment. Distortion elevates RMS current beyond nameplate assumptions.
These are measurable phenomena.
If harmonic distortion and imbalance are present — and not properly quantified — then overheating is not mysterious. It is predictable.
Have full harmonic spectrum measurements been performed at the Point of Common Coupling?
Has Total Demand Distortion (TDD) been evaluated under maximum loading?
Have resonance conditions between capacitor banks and network impedance been studied?
Or are we relying on basic RMS measurements and assuming the waveform is clean?
🔌 And What About Cables “Burning Off”?
Conductors fail when thermal limits are exceeded.
Harmonic currents increase I²R losses. Skin effect worsens at higher frequencies. Neutral conductors — often undersized — carry cumulative triplen harmonic currents in three-phase systems.
When insulation degrades and joints overheat, it is not random.
It is electrical stress operating outside design assumptions.
Were these networks designed for inverter-dominated loads?
Were they assessed for distortion before large-scale renewable or drive integration?
🖥 Why Do Sensitive Control Systems Suddenly Malfunction?
Across manufacturing plants, major mines, and industrial facilities, sophisticated PLCs, VSDs, protection relays, and monitoring systems occasionally fail without an obvious mechanical trigger.
But power electronics are sensitive to:
- Voltage distortion
- Transient spikes
- Negative phase sequencing
- Neutral displacement
- Rapid harmonic fluctuations
When distortion levels fluctuate, control systems can misread signals, trip incorrectly, or reset unexpectedly.
We often blame “software glitches.”
Have we ruled out waveform instability?
Have we examined sequence components?
Have we performed power quality logging during fault events — or only after the failure?
The Uncomfortable Question
In a network increasingly populated by:
- Solar PV inverters
- Variable Speed Drives
- EV charging systems
- LED lighting arrays
- High-density data systems
…harmonic generation is not optional. It is inherent.
If up to 80% of power quality disturbances originate within facilities themselves, then the responsibility to measure, model, and mitigate lies closer to home than we might prefer to admit.
Are we designing networks for yesterday’s linear loads?
Or today’s electronic reality?
A Call for Engineering Honesty
This is not criticism. It is a professional challenge.
Before labeling failures as “unknown,” we must ask:
- Has a full harmonic study been conducted?
- Has negative sequence current been quantified?
- Has transformer derating been properly calculated?
- Has resonance analysis been performed?
- Have compliance limits been actively monitored?
Power quality problems are rarely invisible.
They are simply unmeasured.
Transformers do not ignite spontaneously.
Cables do not burn for no reason.
Control systems do not fail without electrical cause.
The waveform always tells the story.
The question is — are we listening?
