The Transformer‘s “Memory” and Its “Reflex”: From Residual Flux Estimation to Inrush Current Suppression

Introduction

When a transformer is reenergised after being deenergised, it can sometimes draw a surge current several times its rated value for a few cycles. This is called inrush current. The surge itself lasts only seconds, but the consequences can persist for years: massive electromagnetic forces capable of damaging winding structures and accelerating insulation ageing; differential protection may misoperate, causing the transformer to be incorrectly tripped; in severe cases, the surge can even trigger “sympathetic inrush” in neighbouring transformers, setting off a chain reaction.

Inrush current is not random. Its source lies inside the core – in a phenomenon called residual flux.

What Is Residual Flux?

A transformer core is made of stacked silicon steel laminations, which exhibit hysteresis: when energised, the core is magnetised; when deenergised, the flux does not return to zero but retains a portion – this is residual flux. It is like the core’s “memory” of its magnetised state before deenergisation.

When the transformer is reenergised, the grid voltage generates new flux in the core. If the direction of this new flux aligns with the residual flux, the two superimpose and the core is driven into deep saturation. Under saturation, the core‘s magnetic reluctance collapses, and the transformer draws a massive current from the grid – this is inrush current.

The Limits of Conventional Methods

The most direct way to suppress inrush current is to know the magnitude and polarity of the residual flux before closing, and then select the optimal closing instant so that the new flux opposes the residual flux, cancelling it out. This is the principle of controlled switching – one of the most effective means of inrush suppression.

But residual flux cannot be measured directly. The conventional approach is either to close “blind” – ignoring residual flux and accepting the inrush risk – or to use DC demagnetisation: applying a DC current after deenergisation to force the core back to zero flux. Demagnetisation requires dedicated equipment, an outage, and considerable time – it cannot be performed before every reenergisation.

A New Approach: Capturing What the Core Leaks

In 2026, a completely new approach emerged – leakagefluxbased residual flux estimation.

The principle is straightforward. Although residual flux inside the core is invisible, it generates a weak leakage magnetic field in the space surrounding the transformer. By placing sensors outside the transformer and collecting leakage flux signals from the air, the residual flux state inside the core can be estimated.

The method works as follows: sensors are arranged at specific positions outside the transformer to collect leakage flux data. Using a preestablished mathematical model typically a relationship curve between leakage flux and residual flux derived from finiteelement simulations – the measured leakage flux values are substituted into the formula to calculate the residual flux density.

This method offers four key advantages:

No equipment modification. Sensors are placed outside the transformer, without touching internal structures.

No outage required. Measurements can be taken without deenergising the transformer.

Simultaneous magnitude and polarity. It provides not only the residual flux magnitude but also its direction.

Practical accuracy. It maintains low measurement error even under lowresidualflux conditions.

From Passive Endurance to Active Prediction

With accurate residual flux data, controlled switching is no longer a “blind guess”. The circuit breaker can be closed at the optimal instant, so that the new flux opposes the residual flux and the two cancel each other – suppressing inrush current at its source.

The deeper significance is that transformer energisation moves from “gambling” to “calculation”. Operators no longer need to brace for inrush surges and protection misoperations; they can know what the core “remembers” before closing, and make the optimal decision accordingly.

Residual flux is the core‘s “memory”. Inrush current is the “reflex” triggered by that memory. By capturing the weak leakage flux signals that the transformer “leaks” into the surrounding air and using them to estimate residual flux – like reading footprints to infer a pedestrian’s path – this technology transforms transformer energisation from guesswork into a calculated operation.