Model geometry, circuits, and windings can all be confirmed within a single screen. Analyzing phenomena is now possible by arranging the physical quantities of causal relationships.

Model View in Analysis Parameter View

The dashboard can arrange the calculation results and quickly determine if there are any issues. In this example, the IPM motor is the design target, and the flux line, magnetic flux density, demagnetization ratio, and torque are displayed on the dashboard. Although there appears to be no problems with torque, it can be observed that the magnet is demagnetized.

Arbitrary winding methods, such as coil pitch at irregular intervals, can now be specified.

Modeling can prove difficult for circumstances where there is an odd number of layers because the coil geometry per each layer is not a single, unified geometry. By specifying winding methods with tables, however, coil end geometry where the number of layers is an odd number can be created with ease.

Calculations have been sped up even further for models that are at a scale of elements ranging in numbers from several hundred million to tens of millions of elements.

Scalability of parallel solver

It can only be expected that shorter calculation times are desired as model scales and the number of cases increase. In Ver.20.0, the number of cores that can be used in parallel computing has been expanded from 512 to 1,024. A speed up more than 12.7 times faster with 1,024 parallels—compared to 32 parallels—was achieved as a result of measuring multiple models.

Speeds up many calculations, including circuit calculation, in a stable manner.

Speed up ratio for total calculation time

This is an example of acceleration in SMP8 parallel calculation using the GPU as a linear solver. The vertical axis shows the speed improvement ratio with respect to non-parallel calculation. An average speed improvement ratio of 9.4 times has been obtained.
System requirements 

Improved accuracy in estimating motor efficiency in the initial stages of design can be achieved.

Speed priority mode does not account for AC copper loss, which may lead to copper loss being underestimated. In Ver.20.0, a method for high-speed calculations by approximating increases in copper loss (AC copper loss) that occurs as a result of current bias within the wires has been introduced. Diagrams displayed show the minute differences in efficiency for efficiency maps that use accuracy priority mode accounting for the current density within wires. It is understood that minute differences are smaller in accounting for AC loss in speed priority mode.

Vibration can be estimated at a high accuracy in consideration of the interaction between electromagnetic force and structural deformation.

One-way coupled analysis is viable in circumstances where it can be assumed that electromagnetic force is not affected by displacement and stress. Meanwhile, two-way coupled analysis estimates electromagnetic force and vibration at a high accuracy by accounting for the changes in gap width and magnetization characteristics that occur due to structural deformation.

Topology optimization using NGnet supports coupled analysis. This expands the range of applicable applications, such as induction heating.

Coil topology and work temperature distribution

This shows an investigation into heating coils that can reduce the inconsistencies in work surface temperature. The left design proposal sees occurrences of inconsistent temperatures with changes made to the gap in accordance with work thickness, whereas the work surface is heated in uniform for the design proposal on the right.