Overview
Evaluating motor performance for a large number of design proposals necessitates a certain calculation cost, but many multiples of design proposals can be evaluated at a high speed by using reduced order models in JMAG.
An application example would be situations dealing with multi-objective and multi-constraint optimization issues where feasible solutions are difficult to obtain. There exists an approach that involves searching for solutions effectively by dividing searches for genetic algorithms into multiple stages. The design variable range is narrowed down with high-speed calculations that use reduced order models in optimization calculations in the first stage, with optimization performed in the second stage using high-accuracy FEA for the narrowed-down range of design variables The amount of time for the entire optimization is shortened as a result.
This document describes the method for executing single analysis and optimization using a reduced order model.
Keywords
reduced order model, ROM, efficiency map analysis, rotating machine, optimization, genetic algorithm, GA, multi-objective, multi-constraint, multi-stage
