Under no load condition, magnetizing current is required to set up the main flux. When a load is connected to the transformer secondary, an EMF is induced in the secondary winding which opposes the main flux. This secondary flux reduces the primary-induced EMF which in turn causes more primary current to compensate the flux produced by the secondary winding. To produce more flux, more magnetising current is needed. So why is the magnetising current always constant?
The magnetising current is, by definition, that current required to set up the flux in the core.
Whether the secondary is supplying a load current or not, the core flux is generating a voltage in both primary and secondary windings that is (to first order, neglecting primary loss resistance) constant. So the flux is constant. So the magnetising current is constant.
When the secondary supplies a current, that will also cause the primary to draw additional current. These currents are however load currents, with the primary and secondary currents cancelling each other out, to leave only the magnetising current causing flux in the core.
Why is the magnetizing current in a transformer always constant irrespective of the load?
Ignoring core loss current (a step too far), there are two different currents taken by the primary winding: -
And clearly, providing there is still the same applied voltage at the primary terminals, the magnetization current remains constant (if series losses LP and RP are small).
Image from my basic website.
When a load is connected to the transformer secondary, an EMF is induced in the secondary winding which opposes the main flux.
The EMF is there all the time irrespective of a load being present.
By definition, the magnetizing current is the current required to set up the flux with an open secondary.
Yes, when the secondary is delivering current to a load, the primary must draw more current, but this fraction of the current is taken to be different from the magnetizing current.
Note the phrases "by definition" and "taken to be" -- you can view this any way you want, but in the end you can make the math easier if you take the transformer to be a linear device* and take the magnetization current to be a separate thing from the load current, and the primary current required to supply it.
* It isn't a linear device, but it's close enough that you can take it as one for the first cut, or even the final cut if you allow for your linear model not being 100% accurate.
Yes, exactly -- all that you're missing is a formula to express how the currents and fluxes vary with load. Your doubt concerns whether, and how, the magnetization nearly cancels out. Maybe they go up by ratio? Heck, maybe it even goes down instead?*
The simple truth is that: for every mA drawn at the secondary, the same current (divided by ratio) is added to the primary. A superposition.
*In fact it does, simply because more current flow, means more voltage drop, so some of the supplied primary voltage, and received secondary voltage, disappear as losses. Thus, core flux also decreases (taking, for the moment, the assumption that core flux remains proportional to applied voltage (EMF) over frequency). Resistance is usually small (~5% at full load) so the flux only drops by about as much at full load.
Magnetizing current is only dependent on the primary voltage.
This is critical to understanding the acoustic hum when a huge transformer is interrupted and then put back online. The magnetizing flux is stored in the Transformer steel core from the phase of cutoff voltage. If the cut-in voltage phase in not the same then an offset flux is stored which can saturate cores momentarily and create a loud hum with tons of forces on the windings of xxx MVA transformers, until it dissipates and balances out.
The normal test for large power transformer core saturation is to test it in production at 10% over or the max voltage rating and measure the dominant 5th and any other harmonics relative to the fundamental with some acceptance criteria.
It might be any harmonic that is dominant and that depends on the BH slope near saturation and thus the slew rate of the voltage induced current.
A different effect is primary inductance drops 10% which is the max rated current value.
Each industry tends to choose a test method according to customer requirements. Wind farms and solar panels often raise commutation spikes that can increase voltage and thus magnetizing transients and thus harmonics with a different impedance.
Sometimes these grid tied inverters see neighbouring farms with high harmonic currents get absorbed by better GTI's and this current of different harmonics flowing in both directions can blow the fuse. I saw data to suggest this was happening on Duke Energy grids.