Approximate dynamic programming and artificial neural network control of electric vehicles: from motor drives to grid integration

dc.contributorHaskew, Tim A.
dc.contributorBalasubramanian, Bharat
dc.contributorHu, Fei
dc.contributorSong, Aijun
dc.contributor.advisorLi, Shuhui
dc.contributor.authorSun, Yang
dc.contributor.otherUniversity of Alabama Tuscaloosa
dc.date.accessioned2020-01-16T15:03:36Z
dc.date.available2020-01-16T15:03:36Z
dc.date.issued2019
dc.descriptionElectronic Thesis or Dissertationen_US
dc.description.abstractThe drive system of an electric vehicle (EV) includes two major parts- the powertrain and charging system. This dissertation investigates the implementation of the approximate dynamics programming (ADP) based artificial neural network (ANN) control on these two parts to increase the efficiency, stability and reliability of EVs. The major challenge of the powertrain control is to control the EV motor, which is usually an interior mounted permanent magnetic motor(IPM). By using the conventional vector controller, the IPM encounters high current distortion and speed oscillation especially when working in overmodulation area, due to the decoupling inaccuracy issue. The ADP-ANN controller resolves the decoupling issue and guarantees better speed and current tracking performance. For industrial implementation, the motor control algorithm is normally achieved by a digital signal processor (DSP), which has limited computational resources. As ADP-ANN has more complex structure than the conventional controller, whether it can be put into a DSP need to be tested. This dissertation optimized the ADP-ANN algortithm and make it successfully running in a TMS320F28335 DSP platform. To control a gird-connected solar based EV charging system, the dc-bus voltage stability of the solar inverter need to be maintained to acquire high charging efficiency and reduce the grid current distortion. This will become a challenge to conventional vector controller when the solar irradiation level changing rapidly. The implementation of the proposed controller allows the solar inverter improve the dc-bus voltage stability, energy capture efficiency, adaptivity, power conversion efficiency and power quality. Multiple EVs can be used to supply reactive power to the grid when connected with the charging system. But, a great challenge is that grid integration inverters would fight each other when operated autonomously in participating grid voltage control using the conventional control methods. The ADP-ANN control is able to properly handle the inverter constraints in achieving Voltage/Var control objectives at the grid edge and overcomes the challenges of conventional DER inverter control techniques.en_US
dc.format.extent141 p.
dc.format.mediumelectronic
dc.format.mimetypeapplication/pdf
dc.identifier.otheru0015_0000001_0003401
dc.identifier.otherSun_alatus_0004D_13894
dc.identifier.urihttp://ir.ua.edu/handle/123456789/6458
dc.languageEnglish
dc.language.isoen_US
dc.publisherUniversity of Alabama Libraries
dc.relation.hasversionborn digital
dc.relation.ispartofThe University of Alabama Electronic Theses and Dissertations
dc.relation.ispartofThe University of Alabama Libraries Digital Collections
dc.rightsAll rights reserved by the author unless otherwise indicated.en_US
dc.subjectElectrical engineering
dc.titleApproximate dynamic programming and artificial neural network control of electric vehicles: from motor drives to grid integrationen_US
dc.typethesis
dc.typetext
etdms.degree.departmentUniversity of Alabama. Department of Electrical and Computer Engineering
etdms.degree.disciplineElectrical and Computer Engineering
etdms.degree.grantorThe University of Alabama
etdms.degree.leveldoctoral
etdms.degree.namePh.D.
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
file_1.pdf
Size:
13.04 MB
Format:
Adobe Portable Document Format