Določanje napolnjenosti baterij z uporabo umetnega nevronskega omrežja

magistrsko delo

Rok Rečnik (Author), Gorazd Štumberger (Mentor), Primož Sukič (Co-mentor)

Abstract

Glavni cilj magistrske naloge je določitev stanja napolnjenosti baterij s pomočjo umetnega nevronskega omrežja. Določitev stanja napolnjenosti (SOC) baterij predstavlja velik izziv, saj je SOC težko natančno določiti. V delu je bil obravnavan tip izredno zmogljivih Toshibinih litij-ionskih baterij s titanovim oksidom (LTO). Na podlagi pregledane literature je bila izbrana metoda določanja SOC z umetnim nevronskim omrežjem. S pomočjo namensko izdelanega testerja baterij so bile opravljene meritve toka, napetosti in temperature baterije. Meritve so bile izvedene s pomočjo merilnega sistema Dewesoft Sirius HS. Podatki so bili obdelani v programskem okolju Matlab, kjer se je tudi kreiralo in naučilo umetno nevronsko omrežje. Testi nevronskega omrežja so pokazali, da je sposobno napovedovanja SOC. S pomočjo programa v Simulinku so bili izvedeni testi za napovedovanje SOC v realnem času.

Keywords

umetna nevronska omrežja;določanje stanja napolnjenosti baterij;litij-ionske baterije;magistrske naloge;

Data

Language:	Slovenian
Year of publishing:	2019
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[R. Rečnik]
UDC:	621.354.32.08(043.2)
COBISS:	22774550
Views:	868
Downloads:	148
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	State of charge estimation for batteries using artificial neural network
Secondary abstract:	The main goal of this Master's thesis is estimating the state of charge of batteries with the help of an artificial neural network. Estimating the state of charge (SOC) of batteries is a big challenge, because it is hard to estimate SOC precisely. The main focus was working with Toshiba's lithium-ion batteries that contain titanium oxide (LTO). With regards to available literature, a method for estimating SOC by the use of artificial neural networks was chosen. Battery current, voltage and temperature measurements were done with the help of a battery tester designed specifically for this purpose. The measurements were recorded using a measuring system called Dewesoft Sirius HS. The data was processed in the programming environment Matlab. Using the same programming environment a neural network was built and trained. Tests of the trained neural network showed that it was capable of assesing the SOC. Tests on determining SOC in real time were done with the help of the program Simulink.
Secondary keywords:	artificial neural network;state of charge estimation;lithium-ion battery;LTO;
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Elektrotehnika
Pages:	XVIII, 135 str.
ID:	11237514