Back to search

National Technical University of Athens (NTUA)

Σχεδίαση και βελτιστοποίηση λειτουργίας κινητήρων μονίμων μαγνητών: ανάπτυξη αλγορίθμων εκμάθησης και θεώρηση του κύκλου φόρτισης

Abstract

dc:description

The present work undertakes the development of composite multicriteria optimization techniques in order to determine permanent magnet electrical machine geometries and energy management profiles for electric vehicles. Within this research framework, hybrid swarm intelligence strategies have been developed that accomplish faster and more stable solution, while they enable effective multiple operating points consideration as well as manufacturing uncertainties, by means of machine learning algorithms. Permanent magnet geometries involving internal and external rotor topologies have been examined, under low and high-speed operating conditions, with fractional slot concentrated windings, through the development of automated parametric models based on finite element analysis. Dynamic simulation models of electric vehicles operation have been developed, respectively, that consider the speed and torque controllers and compute the consumed energy through the efficiency map of the propulsion system. All the developed methods have been tested experimentally, and their advantages have been validated. More specifically, initially, a multicriteria optimization routine was developed for the design of an outer rotor 16-pole traction machine with surface permanent magnets for a prototype light electric vehicle. It is based on meta-heuristic swarm intelligence algorithms and works best in cases of a single or a few operating points. The proposed strategy exhibited faster and more stable convergence with respect to well-known and established algorithms of the same class, based on test function suite and motor application assessments. In a next stage, a composite multiple criteria optimization methodology has been developed that enables the computation of the complete efficiency map of the internal rotor two-pole machine and consequently the total consumed energy of an electric vehicle during the driving cycle with the use of a dynamic model. The strategy key developments lie on a finite element surrogate modeling via a machine learning algorithm and on an iterative training scheme that minimizes the prediction error. Efforts were made to track the trends of the optimum solutions in the evaluation and search spaces and the procedure was experimentally tested via a prototype machine construction and measurements. In a next step, a new robustness criterion has been expressed, that considers multiple criteria during optimization in an objective nature. The resulting methodology has been assessed through test functions and has been applied in the robust optimal design of a high-speed double pole machine by using a machine learning routine so that feasible computational complexity is achieved. Afterwards, modeling and experimental validation of thermal and mechanical phenomena that need to be considered in electrical machine high-speed operation has been undertaken. Thermal, rotor dynamic and structural analysis has been performed for the two-pole permanent magnet motor to determine the maximum achievable rotating speed in terms of geometrical characteristics of the structure and the materials used. The correlation between the various geometric dimensions of the rotor with the maximum rotating speed have also been investigated. Finally, an optimization methodology of the energy consumption and trip time of a purely electric vehicle has been suggested to determine appropriate speed and torque limit profiles tolerating limited overloading of the traction motor. The proposed strategy is step-wise, and an itinerary segmentation technique based on the road elevation profile has been introduced. Dynamic model simulations and experimental results under real driving conditions highlight the significant benefits with respect to the classic steady speed driving technique.

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2017

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Krasopoulos, Christos
  • Κρασόπουλος, Χρήστος

Subjects

dc:subject × 42

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/42192

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Krasopoulos, Christos; Κρασόπουλος, Χρήστος. Σχεδίαση και βελτιστοποίηση λειτουργίας κινητήρων μονίμων μαγνητών: ανάπτυξη αλγορίθμων εκμάθησης και θεώρηση του κύκλου φόρτισης. National Technical University of Athens (NTUA), 2017. http://hdl.handle.net/10442/hedi/42192