Abstract
The aim of this study, is modelling wear behaviour of Ni3Al+B intermetallic compound, which is used commonly in applications where high temperatures required, by using artificial neural networks. The study was carried out in two stages; experiment and artificial neural network modelling. Firstly, four different load of 50, 75, 100 and 125N at four different temperatures of 25 (room temperature), 100, 300 and 450°C respectively, the experiments were accomplished. Then, the system was modeled as a three-layer feed forward neural network structure, formed with the help of datas obtained from the experimental results. The 75 percent of experimental results was used for education and remaining 25 percent was used for test. Consequently, the proximity of values obtained from experiments and artificial neural networks reveals the success of this study.
Keywords
Abstract
Bu çaly?manyn amacy, yüksek sycaklyk uygulamalarynda syklykla kullanylan Ni3Al+B metaller arasy bile?i?inin (intermetalik) a?ynma davrany?yny yapay sinir a?lary (YSA) kullanarak modellemektir. Çaly?ma deneysel ve yapay sinir a?y modellemesi olmak üzere iki a?amada gerçekle?tirilmi?tir. Ylk olarak 50, 75, 100 ve 125 N olmak üzere 4 farkly yük ve oda sycakly?y, 100, 300 ve 450°C olmak üzere 4 farkly sycaklykta a?ynma deneyleri gerçekle?tirilmi?tir. Daha sonra 3 katmanly ileri beslemeli bir YSA yapysy olu?turularak deneysel sonuçlardan elde edilen veriler yardymyyla sistem modellenmi?tir. Deneysel verilerin %75'i e?itim, geri kalan %25'i ise test amaçly kullanylmy?tyr. Çaly?manyn sonucunda model ile gerçek de?erlerin birbirine olan yakynly?y olu?turulan modelin ba?arysyny ortaya koymu?tur.
Keywords
Details
| Primary Language | Turkish |
|---|---|
| Journal Section | Computer Engineering |
| Authors | |
| Publication Date | February 1, 2011 |
| Published in Issue | Year 2011 Volume: 6 Issue: 1 |