Learning Forum

[pepecarte22]

Hello everyone, I hope you are all doing well. I would like to ask for some assistance from the community.

I am modeling a transformer in Ansys Maxwell 2D and need to determine its efficiency. The load connected to the transformer is a single fixed resistor, so there are two straightforward ways to calculate the output power:

• The RMS voltage across the load squared, divided by the load resistance
• The average value of the product of the load voltage and current

I have verified that both methods yield exactly the same result.

The main issue arises when calculating the input real power. Since I do not directly know the power factor, I considered two simple approaches:

• The average of the product of the primary voltage and current
• The sum of the load power and all transformer losses (i.e., the average winding and core losses)

However, these two methods do not agree. Specifically, the first method gives 610.5 W, while the second yields 561.4 W, which is a significant discrepancy.

To clarify, the transformer losses were calculated as:

• StrandedLoss, representing RI^2 losses (where R is the DC resistance of the windings)
• CoreLoss, which includes eddy current, hysteresis, and excess losses

All windings are defined as stranded, eddy effects are disabled for all geometries, and core loss is enabled only for the magnetic core structure. Additionally, the simulation reports that SolidLoss is zero in this case.

When defining the expression for the average of the product of primary voltage and current in the transient report, I selected a steady-state interval containing an integer number of fundamental periods, ensuring that the waveforms start and end at the same positions/values of the waveform. The same approach was applied to the loss calculations.

A few additional details:

• There is a significant DC component in the primary current, along with some harmonic content, which is expected in my study due to magnetic saturation.
• The DC offset and harmonics should not affect the average power calculation, and I have verified this.
• The primary voltage is a pure 60 Hz sinusoid, with no harmonics or phase shift.
• The simulation uses a time step of 0.1 ms, providing more than 160 points per period, which I believe to be sufficient for accuracy.
• The mesh is well refined.

I also computed the power factor using FFT by extracting the fundamental component phase, then determining the displacement factor and finally the PF (which needs THD). However, these results are not consistent with the other methods mentioned.

I apologize if I am overlooking something basic or if there is a more appropriate approach to this problem.

Thank you very much for your time and any assistance you can provide.

[GLUO]

Hello,

Solid loss is for solid winding or loss on conductors with eddy effect turned on. Core loss results is from post processing and it is not included in your input power. Please use integration of 1/T*i(t)*v(t) over time period T for the real power.

GL

[pepecarte22]

Hello GLUO, thank you very much for your fast response. One question, how do you recommend me to do this integral inside ANSYS, and since there is a phase shift between current and voltage, what is the best way to define the interval of this integral? Also, wouldn't the average of the product between I(t) and V(t) also result in the same value of real power?

[pepecarte22]

I plotted V(t)I(t) as a function of time, then added markers on the two extremities of a period in steady state. After that, in the expression box, I used 1/(t2 - t1) * integ(Voltage(t)*Current(t)), with the range delimited to t1 and t2, where t1 and t2 are the values of time from the markers (so t2 - t1 is the period). The value I obtained is basically the same as avg(Voltage(t)*Current(t)).