A Computer-Aided Design Process for Optimizing the Size of Inverter-Fed Permanent Magnet Motors

A process for designing an inverter-fed motor while taking into account the performance of the inverter is proposed. To draw a comparison, the classical approach to machine sizing is discussed and its shortcomings in achieving a more optimal solution to the motor-drive system design problem are pointed out. The results prove that designing the motor as an independent system may give less optimal solutions compared to the case where an integrated system is considered. Therefore, the machines and the inverters should be designed simultaneously. An inverter-fed interior permanent magnet motor is investigated, and the corresponding design guidelines are presented. These guidelines are based on the knowledge extracted from an inclusively modeled, sampled, and simulated design space. The reliability of the design, the inverter efficiency, and the priorities of the transient, rated and flux weakening operations are assigned as the selection criteria. Moreover, various user-package interfaces are proposed to show how the designer should be guided towards a more realistic design choice. Finally, four different design cases are presented and compared to the output of the classical machine sizing process. The proposed approach serves as a proposal for the next generation of processes implemented with the aid of a high-performance cloud-computing service.

I. Introduction

The classical machine design process (CMDP) is the basis of existing commercial design packages (see Fig. 1) [1]. In this process, the rated specifications are determined by the designer who applies the analytical formulas of classical machine sizing (CMS) to the rated quantities in order to produce an initial design. Afterward, major refinements are performed using finite element analysis (FEA) [2]. The first issue is answering the question of “What is the best way of refining the design?” which requires high amounts of technical knowledge which an inexperienced designer may lack. Therefore, devising a more intelligent method of dealing with this issue is of high importance. The second issue is that analyzing the motor behavior in operating conditions other than the rated one is beyond the scope of CMS. The problem is that the existing analytical solutions obtained for the rated operation do not necessarily provide an optimal design structure for other operating conditions. That is why a detailed FEA is used to refine the model [3]– [6]. Using CMS, the motor is designed as a separate system and the process has nothing to do with the inverter and its effects on the machine sizing. The FEA-based refinement step usually uses a current-fed model. This step is performed by applying the same current magnitude and advance angle to multiple samples in the design space in search of an optimal solution [7], [8]. However, a more suitable approach is to apply different current magnitudes and advance angles for different motor operating conditions. Fig. 1.

Classical machine design process (CMDP) [1].