Measuring techniques to determine hydrophobicity of polymer insulators

Date
2020-12
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: High voltage insulators play an important role in power grid networks. Maintaining highvoltage insulators is essential in creating a reliable grid. Various factors such as pollutioncan influence the hydrophobicity characteristics of an insulator, affecting its ability to in-sulate. Therefore, it is imperative that techniques should exist to determine the insulator'shydrophobicity properties. Several techniques already exist, all with some shortcomings.Two techniques were developed in the quest to create a field ready hydrophobicity measure-ment method. Machine learning was used to do image analysis of the insulators. Severalconvolutional neural networks were investigated and tested. A good correlation was foundfrom the convolutions neural network results and the insulator's surface hydrophobicity.A 3D laser surface constructor was developed to accurately scan an insulator's surface, cre-ating a 3D computer model of the water formations present on its surface. Analysis wasdone on the 3D model, displaying the correlations between the insulator's surface hydropho-bicity and features extracted from the 3D model. Promising results were achieved with thismethod. Both these methods can be used in the field to obtain the surface hydrophobicityof an insulator.
AFRIKAANSE OPSOMMING: Hoogspanningsisolators speel ’n belangrike rol in ons nationale krag netwerk. Die instand-houding van hoogspanningsisolators is noodsaaklik om ’n betroubare krag toevoer netwerkte skep. Verskeie faktore, soos besoedeling, kan die hidrofobisiteitseienskappe van ’n isolatorbeà ̄nvloed. Die gevolg is nadelige effek op die isolators se isoleringsvermoë. Daarom isdit noodsaaklik dat daar tegnieke moet bestaan om die hidrofobisiteitseienskappe van dieisolator te bepaal. Daar bestaan reeds verskeie tegnieke, almal met enkele tekortkominge.Twee tegnieke was ontwikkel om ’n veldgereedheidsmetode om hidrofobisiteit graduering tebepaal. Masjienleer is gebruik om beeldanalise van die isolators te doen. Verskeie indruk-wekkende neurale netwerke is ondersoek en getoets. ’n Goeie korrelasie was gevind tussendie resultate van die neurale netwerk en die oppervlakhidrofobisiteit van die isolator.’n 3D-laseroppervlakte-konstruktor was ontwikkel om die oppervlakte van ’n isolator ak-kuraat te skandeer. Die water formasies op die oppervlakte was dan ontrek vanuit die3D-rekenaarmodel. Analise was gedoen op die 3D-model. Korrelasies tussen die isolator seoppervlaktehidrofobisiteit en karakterastieke was toe gevind uit die 3D-model. Belowenderesultate was behaal met hierdie metode. Albei hierdie metodes kan in die veld toegepasword om die oppervlakhidrofobisiteit van ’n isolator te verkry.
Description
Thesis (MEng)--Stellenbosch University, 2020.
Keywords
Polymer insulators, UCTD, Hydrophobicity, Electric insulators and insulation, Machine learning
Citation