BİLDİRİLER

BİLDİRİ DETAY

Ahmet Gürkan YÜKSEK, Tahsin BOYRAZ, Abdullah KARTAL, Ahmet AKKUŞ
PREDICTION OF WEAR PROPERTIES OF CAO AND MGO DOPED STABILIZED ZIRCONIA CERAMICS WITH GAUSSIAN PROCESS REGRESSION (GPR)
 
In this study, the wear properties of CaO/MgO added stabilized zirconia ceramic composites produced by powder metallurgy method were examined and then experimental wear results were analysed and modelled using Gaussian Process Regression (GPR). CaO and MgO stabilized zirconia ceramics were prepared by the traditional ceramic production method. CaO/MgO added stabilized zirconia ceramics were fabricated by using a combined method of ball milling, cold pressing (CP) and sintering. CaO and MgO in different amounts (0-8 %mole) were mixed with zirconia (ZrO2). These mixtures were prepared by mechanical alloying method and used zirconia ball mill to homogenize the blend with acetone as medium. The powders were dried in oven at 110 ºC for 24 hours before mixing. After drying, powders were compacted to cylindrical preforms with a diameter of 12.7 mm by uniaxial pressing at 300 MPa. The green compacts were sintered at 1600 oC for 3 h a heating rate of 5oC min-1. The characterization studies of the produced samples were carried out and the wear experimental results obtained were converted into data suitable for modelling with Gaussian process regression. In the continuation of the study, Gaussian process regression, one of the machine learning methods, was preferred to predict the wear amount of zirconia. Wear load, wear time, sintering temperature and sintering time data were used as Gaussian process regression input variables. Wear value were taken as output variables of Gaussian process regression. As a result, the training results and test results were compared with the actual values to control the network performance. A good agreement was observed between the experimental and Gaussian process regression model results. After the Gaussian process regression estimation, confirmation tests were performed to confirm the experimental results. ORCID NO: 0009-0009-7048-9095

Anahtar Kelimeler: Ceramic, Zirconia, Wear, Gaussian Process Regression



 


Keywords: