doctoral thesis
INTELLIGENT ESTIMATION IN DYNAMIC POSITIONING SYSTEMS OF MARINE VESSELS

Marko Valčić (2015)
University of Rijeka
Faculty of Maritime Studies, Rijeka
Metadata
TitleInteligentna estimacija u sustavima za dinamičko pozicioniranje plovnih objekata
AuthorMarko Valčić
Mentor(s)Vinko Tomas
Sadko Mandžuka
Abstract
Sustavi za dinamičko pozicioniranje plovnih objekata koriste se za održavanje njihove pozicije, smjera napredovanja i brzine, održavanje unaprijed definirane putanje gibanja, potpomognuto sidrenje i sl. Da bi se ove operacije uopće mogle provoditi, nužno je, između ostalog, omogućiti i određivanje precizne estimacije niskofrekventne pozicije, smjera napredovanja i brzine plovnog objekta, te estimaciju vjetrovnog i sporopromjenjivog opterećenja koje uzrokuju ostali vanjski poremećaji. U realnim sustavima za pozicioniranje plovnih objekata funkciju observera, tj. estimatora, ima neka od inačica Kalmanovog filtra koji ima već dugu tradiciju u brodskim sustavima upravljanja. U radu su analizirani klasični koncepti na kojima su temeljeni postojeći sustavi za dinamičko pozicioniranje te su istražene značajke dinamičkog pozicioniranja plovnih objekata s teoretske i praktične strane, posebno u dijelu koji se odnosi na problematiku filtriranja, identifikacije, estimacije i predikcije. Uočene su brojne prednosti, ali i nedostaci postojećih rješenja koji se mogu otkloniti primjenom novijih računalnih tehnologija kao što su algoritmi strojnog učenja i računalne inteligencije. Iz navedenih razloga, predložene su i konstruirane strukture statičkih, dinamičkih i hibridnih inteligentnih identifikatora i estimatora za potrebe identifikacije i estimacije u sustavima za dinamičko pozicioniranje. Od posebnog značaja su predloženi hibridni sustavi inteligentnih identifikatora i estimatora s proširenim Kalmanovim filtrom te inteligentni identifikatori za fuziju senzorskih informacija i rekonstrukciju signala u prekidu. Predloženi inteligentni identifikatori i estimatori su verificirani na realnim mjerenjima DP Log arhive dizaličara i cjevopolagača Saipem 7000 tijekom postupka polaganja cijevi na Projektu Ormen Lange (Norveška, 2006.).
Keywordsdynamic positioning systems Kalman filter neural networks intelligent identification intelligent estimation intelligent prediction
Parallel title (English)INTELLIGENT ESTIMATION IN DYNAMIC POSITIONING SYSTEMS OF MARINE VESSELS
Committee MembersSerđo Kos (committee chairperson)
Vinko Tomas (committee member)
Sadko Mandžuka (committee member)
GranterUniversity of Rijeka
Faculty of Maritime Studies, Rijeka
PlaceRijeka
StateCroatia
Scientific field, discipline, subdisciplineTECHNICAL SCIENCES
Traffic and Transport Technology
Maritime and River Traffic
UDK004
GENERALLY
Computer science and technology. Computing. Data processing
629
APPLIED SCIENCES. MEDICINE. TECHNOLOGY
Transport vehicle engineering
Study programme typeuniversity
Study levelpostgraduate
Study programmePostgraduate (doctoral) university study programme - Maritime Studies
Academic title abbreviationdr. sc.
Genredoctoral thesis
Language Croatian
Defense date2015-01-08
Promoted date2016-04-28
Parallel abstract (English)
Dynamic positioning (DP) systems are used for maintaining position, heading and speed of the vessels, but also a predefined motion path, position mooring, etc. To ensure performing of these operations, it is necessary, among other things, to determine an accurate estimation of low-frequency position, heading and speed of the vessel. Additionally, it is necessary to ensure the estimation of wind and slowly-varying loads caused by other environmental disturbances. In actual DP systems, the vessel observer is usually an extended Kalman filter (EKF) which is traditionally used in marine control systems. In this doctoral thesis the classical base concepts of the existing commercial DP systems are analysed. Furthermore, the characteristics of DP systems are analysed both from the theoretical and practical point of view, especially in the part which is closely related to filtering, identification, estimation and prediction. Numerous advantages of existing solutions are identified, but also the several disadvantages which can be eliminated by using modern computational technologies such as machine learning and computational intelligence algorithms are pointed out. For these reasons, structures based on static, dynamic and hybrid intelligent identifiers and estimators have been proposed for the purpose of intelligent identification and estimation in DP systems. Proposed hybrid system of intelligent identifiers and estimators combined with EKF, as well as the intelligent identifiers for the sensor fusion and reconstruction of lost signals, are of particular interest. Intelligent identifiers and estimators are further adjusted, tested, and verified with real measurements from the DP Log archive of the heavy-lift and J-lay pipe vessel Saipem 7000.
Parallel keywords (Croatian)dinamičko pozicioniranje plovnih objekata Kalmanov filtar neuronske mreže inteligentna identifikacija inteligentna estimacija inteligentna predikcija
Versionaccepted version
Resource typetext
Access conditionAccess restricted to students and staff of home institution
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
Noteaccepted version
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:187:080541
CommitterDolores Markotić