Maxwell 3D Transient Solution Based on A-Phi Formulation

By default, the Maxwell 3D transient solver is based on T-Omega (W) formulation, which is a powerful method for solving a wide range of low-frequency EM problems. However, there are some areas where the applicability of the method is limited – such as multiple (mixed) source excitations on a single conduction path and capacitive effect (displacement current). For simulation of such problems, the Maxwell 3D Transient A-Phi (F) formulation  is a more suitable solution type.

Comparison of T-Omega and A-Phi Solvers

T-Omega A-Phi

Solves second-order elements for magnetic B field

Solves second-order F for electrical E field, and solves first-order A for magnetic B field. (In order to account the difference in the order of elements solved, increasing the mesh density in A-Phi should help achieve the same B field results as T-Omega.)

Computational efficient for electric machines applications

Computational efficient and flexible for ECAD PCB and electronics applications

Does not support multi-terminals with mixed excitation types on the same conduction path

Supports multi-terminals with mixed excitation types on the same conduction path

Ignores displacement current

Can consider capacitive effects (displacement current)

Supports all advanced material modeling

Limited advanced material modeling capabilities

Easy handling of motion due to only scalar potential for the motion coupling

Does not support motion

For more information on the Transient A-Phi formulation, refer to the following sections:

Limitations

The following 3D Transient features is not supported/available for this release:

Boundaries and Excitations

Boundaries and Excitations supported are as follows:

Winding Excitations for A-Phi Formulation

Winding excitation is assigned as in the Transient solver (Refer to Assigning a Winding Setup for a Transient Solver for more information on winding assignments). For winding excitations, the coil terminals must be ordered according to the physical potential distribution. To do this:

  1. Right-click on the winding to open the context menu.
  2. Click on Order Coil Terminals to open the Order Coil Terminals dialog box.

    Example of the Order Coil Terminals dialog box

  3. Order the terminals by dragging them in the Order Coil Terminals dialog box.
Coil Order

The solver connects the coils according to the coil order specified via Order Coil Terminals. Voltage/current values are assigned to the first terminal, and the last terminal is grounded. To generate the correct potential plot, the terminals should be ordered according to the actual connection of the conductors.

A winding with coil terminals assigned to the external surfaces of the conductors are shown below. The coil terminals are labeled with numbers.

If a winding has more than three external terminals, as in this example, the solver makes the consecutive terminals (except first and last terminal) equipotential (or connected). For example, in Coil Order Case 1, terminals (4,5) and (6,7) have the same potentials . Coil Order Case 2 shows the effect of reordering the coil terminals. Even though Coil Order 1 and 2 have the same terminals, the different order of the terminals gives different potentials, but still the same B, H, and J fields.

For coil terminals defined inside the conductors, the solver defines two potentials for each terminal, and the same connection rule described above applies considering this double potential definition. See Rules/Limitations of Excitations on Conduction Paths for further information about winding excitations.

Voltage Excitations for A-Phi Formulation

For designs with known potentials, select and assign a voltage excitation on each face.

You can also define initial currents for Voltage Excitation by checking the Initial Current box. When initial current is defined, the solver treats it as the current excitation. The default value is 0 A. You should select at least one voltage excitation without enabling the initial currents on a solid conduction path to set a reference potential. (If the path is touching the odd symmetry boundary, where the potential is set to 0 V, or if displacement effect is considered in the problem domain, then this rule does not apply on this path.)

Voltage excitation supports three different excitation models:

Current Excitations for A-Phi Formulation

For designs with known current values, the user selects and assigns a current excitation on each face and also selects the type of conductor: solid or stranded. Current excitation supports three different excitation models for a Solid conductor type based on how the potential DOF is assigned on the face:

Rules/Limitations of Excitations on Conduction Paths

Winding Excitations

For the eddy current simulations with displacement effect off for all objects:

When the displacement effect is on, the restrictions above will not apply. You can define one external coil terminal per conduction path and needs to ensure that there is a reference potential defined in the design.

Current Excitations

Voltage Excitations

Independent/Dependent and Odd Symmetry Boundaries

Matrix Setup

Parametric matrix setup is supported for the A-Phi transient solver. Only one matrix setup can be defined for inductance, capacitance, or DC resistance calculations. For computation of the defined matrix, select Enable Inductance Matrix Computation, Enable Capacitance Matrix Computation, or Enable DC Resistance Matrix Computation in the 3D Design Settings window:

3D Design Settings window with the Matrix Computation tab selected

Next, select the signals and grounds for the matrix of interest. For the inductance matrix, the following are restrictions on ground selections:

It is highly recommended that ground selection be as in the nominal problem setup, especially for designs with multi-terminal excitations on a conduction path. Other solver types do the matrix computation after the solution of nominal problem so it is fully postprocessed, and you can select different signal/ground/return path configurations. However, the transient solver is computing the matrix at each timestep and the computation is highly dependent on how the current flows on a path, especially for designs with multiple excitations on the path.

Example of matrix setup for a design with multiple excitations

The capacitance matrix is computed using the electrostatic energy stored in a system under 1 V excitation for one selected signal and 0 V for all others. Based on the selection of signals and grounds, the solver computes the capacitance matrix of the system. Stranded conductors and floating conductors (no signal or ground assigned) are treated as equipotential surfaces. For the capacitance matrix, the following rule applies for the ground selection:

The resistance matrix is computed with the assumption of a DC conduction solution using the material values at time zero, so it is time independent. For stranded conductors, the solver internally calculates the stranded resistance values. The resistance value (the user typically inputs the stranded resistance value in addition to the source resistance) in the winding excitation setup is not included in the DC resistance matrix.

Note: When there is a layout component with RLC components, the inductance, capacitance, and resistance matrices are calculated with the assumption that R is connected, L is shorted, and C is open. Thus, the L matrix values are the total inductances of the conduction paths connected with R and L circuit elements. Also, the resistance matrix values include the lumped R circuit component values. Therefore, you should be careful while defining signal(s) and ground(s) on a conduction path.

Postprocessing Support

The A-Phi transient solver supports all the postprocessing features of the transient solver. Additionally, the following features are supported: