Summary

On this page, you can find some exercises for getting familiar with EMDtool. You can find all the required files in the Examples folder of your EMDtool download. Make sure you have unzipped the file.

General instructions

Before starting, it is recommended to clear your Matlab path and variables with clear and restoredefaultpath.
Begin by adding the EMDtool folder to your Matlab search path, for instance by running the example_init_and_readme.m script.
If you get stuck at any point, do check out the documentation either by typing e.g. help SPM1 or help MagneticsProblem.solve_static in the Matlab command prompt, or by checking the online documentation (hint: the search bar on top of the documentation page can be a handy tool).

Exercises

Simple surface-PM machine

This exercise demonstrates some simple yet powerful analysis options using a simple surface-PM machine geometry template.

Begin by running the Geometries/toy_SPM.m script to initialize the model. Please make sure to read through the script to see what is actually done.
Run Analysis scripts/calculate_no_load_flux_density.m to calculate and visualize the no-load flux density distribution along with the airgap flux density.
Run Analysis scripts/stepping_simulation_static.m to run a sweep of magneto-static solutions over a range of rotor angles.
- In the absence of strong damping by eddy-currents or other dynamics, this type of analysis is an extremely powerful too.
- Read through the analysis script to see if you understand all the steps.
- Pay particular attention to the summary structure returned by the motor.results_summary method. This structure contains a large amount of useful data in compact format.
Try different options:
- The Halbach and/or six-phase options included in the geometry script.
- Changing the rpm, pole angle, current density, and other simulation settings.

IPM machine with hairpin winding

Next, we will see some transient analysis and winding eddy-current losses.

Initialize the model by running Geometries/toy_winding_model_example.m. Please make sure that the winding_model parameter on line 61 or so is set to "hairpin".
Run torque_curve_vs_pole_angle.m using 10 Arms/mm^2 current density. Note the pole angle at which the torque peaks.
- This computes the torque curve, using a single rotor position only. Thus, the accuracy of the results depends on the amplitude and phase of the torque ripple, and the rotor angle used.
Again run the static stepping script, setting rpm to 3000, and the pole angle to the one you noted just seconds ago. Also, set the current density to the same as at previous step.
- After running the script, note the phase voltage d- and q-components in summary.phase_circuit_data.U_phase_dq
- Does the torque agree? Why / why not?
Next, run Analysis scripts/stepping_simulation_transient_simpler. Make sure that:
- The rpm is the same you used before with the static analysis.
- Voltage supply is used: lines 33- are uncommented.
- The supply voltage vector Udq is set to the phase voltage vector you obtained with static analysis.
See how the results differ from static analysis. Why?
Pay particular attention to the winding losses computed and visualized by the [w, wdata] = phase_circuit.losses(stepping_solution, 'plotting_on', true, 'Jlim', 3*Jrms*[-1 1]) line.
Next, let’s re-run the analysis using PWM supply. Near line 33, comment away the line source = VoltageSource(motor); and replace it with modulator = SVPWM_Modulator(400, 6e3); source = VoltageSource(motor, modulator);. This sets the model (or the phase winding of the stator, to be exact) to be fed with space-vector PWM controlled voltage, at a switching frequency of 6 kHz and a DC-link voltage of 400 V.
- It’s a good idea to increase the number of time steps per period to 400 or so, to accurately catch the switching phenomena.
- Pay particular attention to the iron losses, torque ripple and waveform, and phase circuit AC losses.
For extra ugliness, repeat the analysis with only 1 kHz switching frequency to see what happens.

Understanding templates

This exercise helps you to get started with creating your own templates.

First, to get an idea of the end-result:
1. Open and run the Geometries/super_simple_solid_rotor_IM.m script to initialize a motor model with the SimpleSolidRotor template.
2. Run Analysis scripts/torque_curve_vs_slip to visualize the torque-slip behaviour and the flux density distribution at (approximately) maximum torque.
Next, type open SimpleSolidRotor on the Matlab command prompt to quickly open the main template file.
- Note that the main file does not contain much functionality - all the magic happens in the create_geometry method in its external file.
- Open this file next, easiest done via the Matlab GUI by clicking ‘Open’
Read through the file to see if you understand it.
Next, try modifying the file. You can try, for example:
- Visualizing some of the created points using e.g. O.plot('text to plot', 'rx').
  - NOTE In the main geometry script Geometries/super_simple_solid_rotor_IM.m, you’ll have to move the rotor creation line rotor = SimpleSolidRotor(dim); after the figure(1); clf; ... line.
- Manually creating an Arc object for the airgap arc, and then adding it to the core_surface:
```
  airgap_arc = Arc(Pout_cw, O, Pout_ccw);
  airgap_arc.name = 'n_ag';
  core_surface.add_curve(airgap_arc);
```
- Replacing the single airgap arc by two arcs of shorter span. For this purpose, you’ll first have to create a suitable midpoint by e.g. Pout_mid = Pout_cw.mirror(angle_pole/2);. This actually improves the template and makes it also work for two-pole geometries.