🧑🎓
What exactly is a hysteresis loop? I see the simulator draws a loop, but why doesn't the magnetization just follow the same path up and down?
🎓
Basically, it's the "memory" of a magnetic material. When you apply and then remove a magnetic field, the material's internal magnetization lags behind and doesn't retrace its steps. This lag creates the loop you see. In practice, it's caused by tiny magnetic domains inside the material resisting change. Try moving the "Coercivity H" slider in the simulator—you'll see the loop get wider, meaning it takes a stronger reverse field to "reset" the magnetization.
🧑🎓
Wait, really? So the area inside that loop has a physical meaning? What happens if I change the "Field Amplitude H" control?
🎓
Great observation! Yes, the loop area represents energy loss per cycle—energy turned into heat as domains snap and realign. If you reduce the Field Amplitude so the material isn't fully saturated, you'll see a smaller, skinnier loop. A common case is in power transformers, where you want this area as small as possible to save energy. The simulator calculates this loss using the integral $W = \mu_0 \oint H\,dM$.
🧑🎓
Okay, so what are "Remanence" and "Coercivity" on the parameter panel? They sound like the key specs for a magnet.
🎓
Exactly! Remanence (Br) is how strong the magnet remains after you turn off the external field—its "stickiness." Coercivity (Hc) is how hard it is to demagnetize it—its "toughness." For instance, a fridge magnet needs high Br. A motor magnet needs high Hc to resist heat. Play with the "Saturation M" and "Coercivity H" sliders together. You'll see Br is the loop's height at zero field, and Hc is its width on the axis.
The simulator uses the Jiles-Atherton model to calculate the magnetization M. The core idea is separating the ideal, reversible anhysteretic magnetization from the irreversible, lossy process of domain wall motion. The effective field accounts for internal interactions.
$$H_e = H + \alpha M$$
$$M_{an}= M_s \left( \coth\left(\frac{H_e}{a}\right) - \frac{a}{H_e}\right)$$
Here, $H$ is the applied external field, $\alpha$ is a mean field parameter representing inter-domain coupling, $M_s$ is the saturation magnetization, and $a$ is a shape parameter. $M_{an}$ is the ideal magnetization if there were no hysteresis.
The actual magnetization M is found by solving a differential equation that pins the change in M to the change in the anhysteretic curve, with a term for irreversible losses. From M, we find the total magnetic flux density B and the energy loss per cycle W.
$$B = \mu_0 (H + M)$$
$$W = \mu_0 \oint H \, dM$$
$B$ is the flux density (what you'd measure with a gaussmeter), $\mu_0$ is the permeability of free space. The loss $W$ is the area enclosed by the M-H loop. In the simulator, adjusting parameters like coercivity directly changes the shape of this loop and thus the calculated loss.
Common Misconceptions and Points to Note
When you start using this simulator, there are a few common pitfalls. First is the oversimplified idea that "a material with a high coercivity Hc is a good magnet". While it's true that hard magnets have a large Hc, the definition of "good" varies completely depending on the application. For example, a magnetic circuit in a speaker requires a material with a reasonably high Hc that is resistant to temperature changes (hard to demagnetize). On the other hand, a motor core absolutely requires a soft iron material with a small Hc (a thin loop); if you used a hard magnet here, it would just overheat and not rotate. When selecting a material, work backwards from "what you want to achieve".
Next is the pitfall of parameter settings. While you can freely change parameters like "saturation magnetization" and "molecular field coefficient α" in the simulator, in real materials these parameters are not independent of each other. For instance, if you arbitrarily increase α too much for a ferrite, you might get non-physical loop shapes (e.g., overly squared). When identifying J-A model parameters in practical work, the golden rule is to fit them to measured B-H loop data and determine them as a set of parameters. For example, for a certain ferrite material, you get a balanced set of values like $M_s=3.2\times10^5$ [A/m], $a=50$ [A/m], $α=0.001$.
Finally, be careful not to assume that "the simulation loop area directly equals the heat generation". It's true that hysteresis loss W is proportional to the loop area, but you must not forget that in actual devices, "eddy current loss" is added to this. Especially in an alternating magnetic field, eddy current loss becomes more dominant as the frequency increases. In this tool, if you select "transformer core" and increase the frequency, the loop shape won't change (because it doesn't consider eddy currents), but in reality, the reason for laminating thin silicon steel sheets is precisely to suppress these eddy currents.
Related Engineering Fields
Understanding magnetic hysteresis is fundamental, even foundational, within CAE, particularly for electromagnetic field analysis (EM analysis). Specifically, it directly relates to "iron loss calculation", which accurately estimates core losses in motors and transformers. When analyzing a motor with FEM (Finite Element Method) software, the B-H curve and loss coefficients set for the materials are often represented by this very J-A model. Sloppy calculations can lead to large errors in predicting motor efficiency, causing significant setbacks in product development.
It is also deeply connected to the fields of power engineering and power electronics. In selecting cores for transformers and inductors used in switching power supplies, hysteresis loss in the operating frequency band dictates efficiency. For example, a DC-DC converter operating at 100kHz uses a ferrite core with a thin loop and excellent high-frequency characteristics. Conversely, large transformers for commercial frequencies (50/60Hz) use grain-oriented electrical steel. By switching materials in this simulator, you should be able to experience how the order of magnitude of the loss changes.
As another application area, consider non-destructive testing (NDT). Internal stress and fatigue degradation in magnetic materials subtly change the shape of the magnetic hysteresis loop. There are techniques that measure this to evaluate material condition. In the simulator, changing the "pinning coefficient k" changes the loop's width; in reality, this can correspond to an increase in internal defects or strain in the material. In other words, the B-H loop can also serve as a material's "health check report".
For Further Learning
As a first next step, I recommend learning about "dynamic hysteresis". This simulator models quasi-static changes (when the magnetic field is changed slowly), but in actual AC applications, you need to consider the loop broadening due to frequency (the influence of eddy currents) and the effect of the rate of change of the magnetic field itself on losses. Mathematically, a world awaits you with extended J-A models containing differential terms like $dM/dt$, or empirical formulas like the "Steinmetz equation" which separates and calculates losses.
If you want to delve a bit deeper into the mathematical background, take a look at numerical solutions for ordinary differential equations. The core of the J-A model is solving the differential equation for magnetization M: $$ \frac{dM}{dH} = \frac{M_{an} - M}{k \delta - α(M_{an}-M)} $$ (where $\delta$ is the sign indicating the increasing or decreasing direction of the magnetic field H). Behind the simulator's smooth animation, it calculates this equation frame by frame using numerical methods like the "Euler method" or "Runge-Kutta method". Writing a simple script yourself to reproduce this will make the essence of the model crystal clear.
Finally, learning from the material's perspective is also important. Knowing the basics of microscopic magnetism physics makes the meaning of the parameters much clearer. For example, "saturation magnetization $M_s$" corresponds to the sum of the magnetic moments of the atoms in the material, and the "molecular field coefficient α" corresponds to the strength of the exchange interaction that forms magnetic domains. When you select "ferrite" in the simulator, the loop becomes somewhat squared because of the underlying microscopic magnetization reversal mechanism (primarily magnetization rotation rather than domain wall motion). Being able to imagine the connection between the tool's parameters and physical phenomena should elevate your applied skills as a CAE engineer to the next level.