Inputs

Standard inputs ––––--

Load resistance

The phase resistance of the 3-phase star connected load “Load resistance” (Phase resistance of the 3-phase load) load must be set. The default value is 1 Ω.

Load inductance

The phase inductance of the 3-phase star connected load “Load inductance” (Phase inductance of the 3-phase load) load must be set. The default value is 1.0E-4 H.

Speed

The imposed “Speed” of the machine must be set.

Ripple torque analysis

The “Ripple torque analysis” (Additional analysis on ripple torque period: Yes / No) allows to compute or not the value of the ripple torque and to display the corresponding torque versus the angular position over the ripple torque period.

The default value is “No”.

Note:

This choice influences the accuracy of results and the computation time. The peak-peak ripple torque is calculated.

This additional computation needs additional computation time.
In case of “Yes”, the ripple torque is computed.

Then, the flux density in regions and the magnet demagnetization rate are evaluated through the ripple torque computation.
In case of “No”, the ripple torque is not computed.

Then, the flux density in regions and the magnet demagnetization rate are evaluated by considering the Park’s model computation.

Advanced inputs ––––--

No. comp. / ripple period

The number of computations per ripple torque period is considered to perform a “Ripple torque analysis”.

The user input “No. comp. / ripple period” (Number of computations per ripple torque period) influences the accuracy of results (computation of the peak-peak ripple torque) and the computation time.

The default value is equal to 30. The minimum allowed value is 25. The default value provides a good compromise between the accuracy of results and computation time.

Figure 1. Definition of the number of computations per ripple torque period

Rotor initial position

By default, the “Rotor initial position” is set to “Auto”

(except in the test Characterization / Cogging where it is a user input whose default value is 0).

When the “Rotor initial position mode” is set to “Auto”, the initial position of the rotor is automatically defined by an internal process of FluxMotor.

The resulting relative angular position corresponds to the alignment between the axis of the stator phase 1 (reference phase) and the direct axis of the rotor north pole.

When the “Rotor initial position” is set to “User input” (i.e. toggle button on the right), the initial position of the rotor considered for computation must be set by the user in the field « Rotor initial position ». The default value is equal to 0. The range of possible values is [-360, 360].

Notice: The relative angular position between the axis of the stator phase 1 (reference phase) and the direct axis of the rotor north pole must be controlled to perform the tests. See the picture below. That allows defining the working point of the machine.

Figure 2. Definition of rotor initial position – Rules for direction

The winding axis of the reference phase is defined from the phase shift of the first electrical harmonic of the magneto motive force (M.M.F.).

By convention, a field pole "North” corresponds to a magnetic flux density vector oriented towards exterior of the rotor.

Note: In the test Characterization / Cogging there is a user input whose default value is 0, no Auto mode available.

Skew model – No. of layers

When the rotor magnets or the stator slots are skewed, the number of layers used in Altair Flux Skew environment to model the machine can be modified: “Skew model - No. of layers” (Number of layers for modelling the skewing in Flux Skew environment).

Note: When there is magnet step skew topology, the number of layers is defined at the design level.

Mesh order

To get the results, Finite Element Modelling computations are performed.

The geometry of the machine is meshed.

Two levels of meshing can be considered: First order and second order.

This parameter influences the accuracy of results and the computation time.

By default, second order mesh is used.

Airgap mesh coefficient

The advanced user input “Airgap mesh coefficient” is a coefficient which adjusts the size of mesh elements inside the airgap. When the value of “Airgap mesh coefficient” decreases, the mesh elements get smaller, leading to a higher mesh density inside the airgap, increasing the computation accuracy.

The imposed Mesh Point (size of mesh elements touching points of the geometry), inside the Altair Flux software, is described as:

MeshPoint = (airgap) x (airgap mesh coefficient)

Airgap mesh coefficient is set to 1.5 by default.

The variation range of values for this parameter is [0.05; 2].

0.05 giving a very high mesh density and 2 giving a very coarse mesh density.

CAUTION: Be aware, a very high mesh density does not always mean a better result quality. However, this always leads to a huge number of nodes in the corresponding finite element model. So, it means a need for huge numerical memory and increases the computation time considerably.

The impact of the airgap mesh coefficient on resultant meshing is illustrated bellow:

Figure 3. Airgap mesh coefficient = 0.45 = default value

Figure 6. Airgap mesh coefficient = 0.1 (zoomed view)

Convergence criteria on current

The advanced user input “Converg. Criteria on current” (Convergence criteria on current) is a percentage driving the convergence of the computation.

During the computation, the process iterates to find the working point (current in DQ area) occurring in the generator feeding the load. The convergence process is completed when the variation of current between two iterations gets lower than the ratio “Converg. Criteria on current” set in input.

Convergence criteria on current is set to 0.5% by default.

The variation range of values for this percentage is ]0;10].

A percentage close to zero gives more accurate results, but a longer computation time. A high percentage can make the convergence shorter but decreases the accuracy of the results. The default value of 0.5% gives a good compromise between accuracy and computation time on most of the computations, but smaller value can be required to increase the computation accuracy on some working points, especially working points close to the open circuit behavior.