Parametri za detekciju i lociranje puknuća cijevi vodoopskrbnog sustava

doctoral thesis

• Title (english): The parameters of detection and location of the pipe burst of the water supply system
• Author: Kristijan Iličić
• Mentor: Mario Šavar (mentor)
• Committee member: Dubravko Majetić (predsjednik povjerenstva)
• Committee member: Mario Šavar (član povjerenstva)
• Committee member: Vesna Alar (član povjerenstva)
• Committee member: Ivo Džijan (član povjerenstva)
• Committee member: Igor Sutlović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2012-01-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering; Process Energy Engineering
• Universal decimal classification (UDC): 628 - Public health engineering. Water. Sanitation. Illuminating engineering

Abstract

U radu je istražena mogućnost detekcije i lociranja puknuća cijevi unutar izoliranog dijela sustava vodoopskrbe, tj. kontrolirane zone na temelju poznavanja vrijednosti pet parametra. Ti parametri bili su po jedna veličina protoka i tlaka mjerene na ulazu u sustav, te tri veličine tlaka mjerene na različitim lokacijama unutar sustava. Izmjerene vrijednosti svih pet parametra su prikupljane i bilježene kontinuirano u vremenu. Za istraživanje su korišteni podaci o vrijednostima veličina protoka i tlaka po satima i danima u vremenskom razdoblju od jedne godine, na način da su hidrauličkim modelom simulirana hidraulička stanja sa i bez puknuća cjevovoda s različitim protocima curenja. Ovi podaci su korišteni kao ulazi u neuronske mreže, dok je primarno kao izlazna veličina razmatrano stanje puknuća, a sekundarno su, ukoliko je puknuće nastalo, izlazne veličine bile prostorne koordinate njegove lokacije. \Nakon provedenih simulacija različitih stanja sa i bez puknuća unutar vodoopskrbne mreže ustanovljeno je da se pomoću dobro modeliranih neuronskih mreža odabrani parametri mogu iskoristiti za dobivanje informacije o pojavi puknuća cijevi. \Neuronske mreže modelirane u svrhu lociranja puknuća cijevi pokazale su da s povećanjem protoka iz cijevi nastalog uslijed njenog puknuća, pogreška kod određivanja lokacije puknuća linearno pada. Međutim, nije ostvaren dovoljan nagib tog pravca kod kojeg bi se, za neki realno očekivani protok puknuća, pogreška lociranja puknuća smanjila na prihvatljivu vrijednost.

Abstract (english)

The work explores the possibility of detection and location of the pipe burst inside an isolated part of the water supply system, i.e. district metered area, based on the known values of five parameters. These parameters were one variable of flow, one variable of pressure at the inlet of the system and three variables of pressure measured at different locations within the system. The measured values of all five parameters were collected and recorded continuously in time of one year. For the research, we used data on the values of flow and pressure for hours and days in the period of one year, in a manner that the hydraulic model simulated hydraulic conditions with and without pipe burst with different leakage flows. These data were used as inputs in neural networks; while the primarily considered output value was the state of burst, and the secondary, if the bursts occurred, output values were the spatial coordinates of burst location. After the simulation of a number of different conditions with and without a pipe burst in the water supply network, it was found that, managing well trained neural networks, selected parameters could be used to provide information of occurrence of pipe burst. Neural networks modeled for the purpose of locating pipe burst showed that, when flow formed due to pipe burst increases, error in determining the location of burst drops. However, we could not obtain sufficient decrease of this flow-error correlation line in order to achieve acceptable pipe burst location error for particular realistic pipe burst flow.

• Keywords: vodoopskrbni sustav; kontrolirana zona; vodoopskrbni cjevovod; puknuće cjevovoda; hidraulički model; umjetne neuronske mreže
• Keywords (english): water supply system; district metered area; water pipeline network; pipe burst; hydraulic model; artificial neural networks
• Language: croatian
• URN:NBN: urn:nbn:hr:235:513032
• Study programme: Title: Mechanical Engineering and Naval Architecture; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, područje tehničkih znanosti (doktor/doktorica znanosti, područje tehničkih znanosti)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-05-14 16:57:51