Investigation of wire erosion operation parameters of Co based shape memory alloys

Nowadays, in parallel with the developing technology, the traditional materials used in the industry are being replaced by the newly developed functional materials. Among the functional materials, smart materials are regarded as the foremost; they can change their shape during presence of an external force and change back to their original shape when the external force is removed. The recently developed Co-Ni-Al alloys, when compared to other shape-remembered alloys have advantages such as having the high ductility for shaping along the thermal applications, making it possible to regulate the conversion temperature of the alloy-forming components as well as the Curie temperature and being inexpensive, thus the alloys are started to be used in the aerospace industry, the automotive industry and the medical sectors. When shape-remembered alloys are shaped, in order to avoid loss of properties the heat effect should be limited, thus the method of wire erosion cutting is extensively used. In this thesis, the production of CoNiAl alloy, which is one of the popular shape remembered alloys of recent years, was made by vacuum casting. After the optical, XRD and DSC analyzes of the produced samples were done, they were subjected to the wire erosion cutting machine with different parameters. The effects of wire erosion cutting parameters (pulse duration, voltage and current) on surface roughness, abrasion resistance and the recrystallization layer formed as a result of surface cutting are investigated. In addition, mathematical modeling of the effects of wire erosion cutting parameters on surface roughness was done by using genetic programming which is one of the genetic algorithm based methods. As a result of the studies, it was shown that the cutting parameters of wire erosion affected the Ra value of surface roughness in the range of 2.06-3.18 µm and the re-crystallizing layer thickness in the range of 2.96-9.13 µm. Furthermore, as the result of the GEP the correlation coefficients for the training and test data were determined as 0.764 and 0.689, and it was observed found that the model values were mostly clustered around the target values. KeyWords: Co based shape memory alloy, Wire EDM, Machinability, GEP