Masters Degree Dissertations

Permanent URI for this collectionhttp://localhost:4000/handle/20.500.12504/93

Browse

Browsing Masters Degree Dissertations by Subject "Grey"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Support vector machine approach in selection of welding method for grey cast iron
    (Kyambogo University [unpublished work], 2023-09) Okure, Robert Omita
    Grey cast iron is the utmost common form of cast iron. It is used in applications where its high stiffness, machinability, vibration dulling, high heat capacity and great thermal conductivity are of advantage, such as in automobile components especially for internal combustion engine cylinder blocks, flywheels, gearbox cases, manifolds, disk brake rotors and cookware. It therefore contributes a lot to the socio-economic and technological standards of any country. With its unique properties such as; easy castability, vibration reduction, high stiffness and high thermal conductivity, repair of grey cast iron steel requires careful consideration when selecting the correct welding method from the many techniques available. This study is focused on developing a tool for selecting appropriate welding methods for grey cast iron. Two common welding methods; Oxygen-acetylene Gas welding and Shielded Metal Arc Welding, were evaluated in respect of the critical factors that affect weld quality like: type of joint, filler material and the carbon composition of the material. It was observed that Arc welding a butt joint with a cast iron electrode gave very high tensile strength of 116.04N/m2 hence very good joining. On the other hand, gas welding a lap joint with Mild steel gave the least Tensile strength of 31.22N/m2 signifying a poor-quality weld. A support Vector machine (tool) was then developed to choose the most applicable method of welding grey cast iron, basing on the three attributes; joint, filler material and carbon composition. Results showed that a Non-linear Support vector machine was the most appropriate, as it was able to identify the most appropriate welding method in all cases. The non-linearity in the attributes was identified to come from the filler material, which was associated with microstructure cracking. Results from the Support Vector Machine were compared to that from the popular TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) method which was able to classify only 75% of the welding methods correctly. A non-linear support vector machine model can be applied to help welders to identify the appropriate welding technique.