Electromagnetic Field Simulation of Switched Reluctance Motors (SRM)

Exactly. The rotor is a simple salient pole iron structure with no magnets or coils. The advantages are rare-earth-free, low cost, and high-temperature tolerance. However, the major challenges are large torque ripple and noise. For example, recent Nidec SRMs for washing machines have achieved a 60% noise reduction using AI-based control.

Simply put, it utilizes the "force that attracts iron to a magnet." When current flows through the stator coil to create an electromagnet, the iron salient poles of the rotor are attracted to it. This is reluctance torque. It exploits the property that the rotor moves to the position where the inductance of the magnetic circuit is maximized.

Exactly. In a typical 8/6 structure (8 stator poles, 6 rotor poles), energizing phases A → B → C → D in sequence causes the rotor's salient poles to be successively attracted to the stator's salient poles, resulting in rotation. However, the phase switching timing and current waveform are critical; getting them wrong can result in no torque or even reverse rotation.

Fundamentally different. PMSM torque is "current × magnet flux," but SRM torque comes from the rate of change of inductance. In the linear region (no magnetic saturation), the instantaneous torque is:

Yes, in the saturation region, inductance also depends on current, so it becomes $L(\theta, I)$. In this case, accurate torque calculation requires derivation from co-energy:

At low currents, it's close to a clean trapezoidal wave. It reaches a maximum $L_{\max}$ at the position where the rotor salient pole faces the stator salient pole (aligned position) and a minimum $L_{\min}$ at the shifted position (unaligned).

However, when the current increases, the iron core saturates near the aligned position, causing $L_{\max}$ to drop significantly. On the other hand, $L_{\min}$ is dominated by the air gap magnetic circuit and hardly changes. In other words, saturation reduces $dL/d\theta$, making it harder to generate torque. This is the biggest challenge in SRM design.

It varies greatly depending on the electrical steel sheet used. For example, common materials in SRMs:

• 35A300 (Japanese thin sheet): Saturation flux density $B_s \approx 1.7$ T. Low iron loss but saturates relatively early.
• 50A470: $B_s \approx 1.8$ T. Low cost but slightly higher iron loss.
• 10JNEX900 (JFE's 6.5% Si steel): $B_s \approx 1.5$ T but has extremely low high-frequency iron loss. Suitable for high-speed SRMs.

In practice, the B-H curve starts to flatten around 1.5–1.8 T, and $\mu_r$ (relative permeability) plummets from thousands to tens. In FEM, accurately inputting this nonlinear B-H curve is key to precision.

It's loud. Subjectively, it produces a metallic sound like "grrr" or "bzz." The mechanism is clear: radial electromagnetic forces deform the stator into an ellipse.

In an SRM, the radial force changes abruptly with phase energization ON/OFF. This force excites the stator iron core's natural vibration modes (especially the elliptical mode = breathing mode). When the excitation force frequency components match the stator's natural frequencies, resonance occurs, causing large vibration and noise. What needs to be done with FEM:

• Obtain the radial force time waveform via electromagnetic field analysis
• Analyze the excitation force spectrum via FFT
• Determine the stator's natural frequencies via structural modal analysis
• Design to avoid matching excitation and natural frequencies (resonance)

Exactly. Therefore, SRM NVH analysis becomes a multi-stage workflow: "Electromagnetic FEM → Radial force extraction → Structural FEM (modal analysis) → Acoustic analysis." Tools like JMAG or Maxwell have functions to export force data to structural analysis, which is commonly utilized.