![]() | Only 14 pages are availabe for public view |
Abstract Electrochemical Machining (ECM) has established itself as one of the major alternatives to traditional methods of machining difficult to cut materials and generating complex contours, without inducing residual stress and tool wear. Owing to the complexity of ECM process, it is very difficult to study the effect of various predominant process cutting conditions on resulting process performance measuring parameters and also, predict their values. To decrease this difficulty, many researchers have so far concentrated on the process improvement in ECM as will be seen in literature review. However, this review showed that no effort has been put on the development of multi input- multi output models to correlate the effect of various machining parameters, on the predominant electrochemical machining criteria. Keeping this is in consideration, the present thesis has attempted to develop a new multi input- multi output model using artificial neural networks (ANN). As an efficient approach to predict the values of resulting process performance measuring parameters such as material removal rate and surface roughness. And also, study the effect of variation of these cutting conditions on performance measuring parameters. The proposed model was trained using experimental data available from a previous experimental work conducted and will be discussed in later chapters. The network was built and trained using MATLAB Neural Networks Toolbox. To verify the accuracy and generalization of the proposed model, a new set of experimental data that haven’t been used during training phase, were introduced to the network as a new input. ANOVA test was performed in order to measure the degree of fitness between experimental data and ANN predicted data. And also, to determine the degree of contribution of cutting conditions considered on material removal rate and surface roughness respectively. The ANOVA test was conducted using MiniTab software. |