| Popis: |
Doktorska naloga obravnava široko področje spremljanja stanja hidravličnih tekočin skozi njihov delovni življenjski cikel, pri čemer izpostavlja in podrobneje obravnava problem kvantitativnega ocenjevanja stanja mineralnih hidravličnih olj ter njihove preostale uporabne dobe. Pri uveljavljanju različnih metod in sistemov spremljanja stanja v industrijskem okolju je ključnega pomena njihova praktičnost in uporabnost, zato se doktorska naloga osredotoča na aplikativni razvoj inteligentnega sistema oz. modela za oceno stanja olja in njegove preostale življenjske oz. uporabne dobe, ki zmore samostojno, brez pomoči izkušenega strokovnjaka, preučiti fizikalno-kemične lastnosti hidravlične tekočine in podati informacijo o stopnji njene ustreznosti. Za oceno stanja olja, in njegove preostale uporabne dobe, je v disertaciji predlagan hibridni matematični model, zasnovan na osnovi predhodno pridobljenih podatkov pospešenega staranja olja. Ustreznost izdelanega hibridnega modela je najprej potrjena na bazi podatkov pospešenega staranja hidravlične tekočine, na podlagi katere je bil model tudi izdelan. V nadaljevanju prikazana implementacija modela na realnem industrijskem sistem pa prikazuje praktično uporabnost razvitega sistema, saj je izdelan model sposoben uspešno in dovolj natančno oceniti stanje in preostalo uporabno dobo olja že po prvi tretjini njegove življenjske dobe. The dissertation discusses the condition-monitoring of hydraulic fluids throughout their life-cycles. Particular emphasis is placed on quantitatively assessing,in detail, one mineral oil's condition and its remaining useful lifetime. Practicality and usefulness are vital when developing and deploying various methods for condition-monitoring systems within industrial environments. Therefore, this dissertation focuses on the development and application of an intelligent system for assessing an oil's condition and its remaining useful lifetime and the capability of independently, without the help of an expert, examining the physical-chemical properties of a hydraulic oil, thus providing information on its condition. For the assessment of an oil's condition and its remaining useful lifetime, this dissertation proposes a hybrid mathematical model based on gathered data from previously conducted oil-ageing tests. The adequacy of the developed hybrid model was first validated using the same data from the oil-ageing tests and was then further used when constructing the model. The following shows an implementation of the model within a real industrial system, which indicates the practical applicability of the developed system. This model is able to successfully and accurately assess the oil's condition and its remaining useful lifetime after the oil has only reached one third of its lifetime. |
