Občutljivostna analiza in optimizacija geometrije sinhronskega motorja s trajnimi magneti

STRGULEC, JURIJ (author), Čorović, Selma (mentor)

V magistrskem delu so predstavljeni rezultati občutljivostne analize in optimizacije z nadomestnimi modeli sinhronskega motorja s površinsko nameščenimi trajnimi magneti (SSTM). V prvem delu magistrske naloge je opravljena podrobna analiza numeričnega modela izhodiščne oziroma referenčne različice dizajna stroja nazivne moči 1100 W, ki je namenjen za vgradnjo v avtomobilski krmilni mehanizem. Opravljena je bila občutljivostna analiza in optimizacija geometrije stroja, remanentne gostote magnetnega pretoka trajnih magnetov in gostote električnega toka navitja z namenom znižanja samodržnega navora in stroškovne cene glede na njegovo izhodiščno različico. Delo v okviru magistrske naloge je bilo opravljeno v sodelovanju s podjetjem iz avtomobilske industrije na empirično preverjenem simulacijskem modelu motorja. Numerični model motorja je bil zgrajen v programu Ansys Maxwell, optimizacija in občutljivostna analiza pa sta bili izvedeni v programu Ansys Optislang. Numerični izračuni so bili izvedeni na simetričnem dvodimenzionalnem modelu stroja v Ansys Maxwell programu, pri čemer proizvodne tolerance niso bile upoštevane. Teoretični del magistrskega dela opisuje koncept optimizacije z nadomestnimi modeli, ki so bili pridobljeni kot rezultat predhodno opravljene občutljivostne analize. Predstavljeni sta tudi večkriterijska optimizacija in optimizacijski evolucijski algoritmi, saj je optimizacija z nadomestnimi modeli stroja opravljena z več kriteriji in z omenjenim algoritmom v programu Ansys Optislang. Matematični algoritem programa Ansys Optislang, z namenom iskanja optimalnega dizajna stroja, sistematično spreminja vrednosti vhodnih konstrukcijskih parametrov stroja znotraj meja vnaprej določenega optimizacijskega območja. Optimizacijsko območje smo določili oziroma omejili z vrednostmi parametrov referenčnega stroja in analitično določenimi konstrukcijskimi parametri stroja. Potrebno je poudariti, da je optimizacijsko območje enako območju občutljivostne analize. Pomemben prispevek pričujoče magistrske naloge je tudi nadgradnja procesa optimizacije z dodatnim modulom, ki omogoča ustrezno kontrolo postopka navijanja statorskega navitja in natančno določitev medfazne upornosti navitja. Omenjeni modul je bil ciljno razvit v programskem jeziku Python v okviru magistrske naloge. Optimiziran dizajn motorja izraža 37 % nižjo prisotnost samodržnega navora ter 5 % cenejšo zasnovo od referenčnega dizajna stroja, kar je spodbudno za izvedbo nadaljnje tolerančne analize, razmagnetilnega testa trajnih magnetov, izdelave prototipov in drugih testov, ki so potrebni za vpeljavo novega dizajna stroja v proces proizvodnje.

• Language: Slovenian
• Keywords: občutljivostna analiza, optimizacija z nadomestnimi modeli, metoda končnih elementov, numerično modeliranje, sinhronski motor s površinsko nameščenimi trajnimi magneti
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FE - Faculty of Electrical Engineering
• Year: 2022
• PID: 20.500.12556/RUL-141344
• COBISS.SI-ID: 124541699

Note

Prešernova nagrada, FE UL, 2023

• Publication date in RUL: 28.09.2022

Secondary language

• Language: English
• Title: Sensitivity analysis and optimisation of a permanent magnet synchronous motor geometry

Abstract

This master thesis presents the results of sensitivity analysis and optimization method based on surrogate models of a synchronous motor with surface-mounted magnets (PMSM). In the first part of the thesis, the analysis of the output characteristics of the reference design motor has been performed on the finite element based numerical model with rated power of 1100 W, which is integrated into an automotive electric steering mechanism. The main goal of the thesis was to minimize the presence of the cogging torque and the total cost of the reference motor design via sensitivity analysis and the motor geometry optimization. In addition, the remanent magnetic flux density of the magnets and the electric current density have also been included into the optimization process in order to improve the motor characteristics and to reduce its total cost, while still complying with the manufacturing requirements. The master thesis was performed in collaboration with a company from the automotive industry on the empirically validated 2D numerical model of the motor with permanent magnets. The tolerances of the manufacturing procedure have not been considered in the model. The numerical model has been built using the commercial software Ansys Maxwell, while the sensitivity analysis and the optimization procedure have been performed in Ansys Optislang software environment. The theoretical background of the optimization based on surrogate models and basics of the sensitivity analysis, including the multi-objective optimization procedure with the evolutionary algorithms, based on which the Ansys Optislang optimization algorithm works, are also described in the thesis. The mathematical algorithm of the Ansys Optislang systematically varies the values of the input design parameters of the motors within the full potential optimization range which was determined in advance. The boundaries of optimization range (which is equal to the sensitivity analysis range) have been determined with the dimensions of the reference motor design and the analytically calculated motor design values. The important contribution of this master thesis is also the software tool, written in the Python programming language, to accurately calculate the stator winding resistance and to evaluate the feasibility of the stator windings insertion into the stator slots of the proposed motor design by the Ansys Optislang optimization procedure. The final results of the sensitivity analysis and the optimization procedure indicated that the cogging torque and the total costs of the PMSM motor can be successfully reduced by 37 %and 5 %, respectively. These results are important also for the prototype manufacturing and also for future development of the PMSM motors which will take into account also the manufacturing tolerances, results of the permanent magnets demagnetization tests and other test which are important for the production process of newly designed machines.

• Keywords: sensitivity analysis, surrogate-based optimization, finite element method, numerical modelling, surface-mounted permanent magnet motors