Development of artificial neural network model to correlate the cutting and process parameters in high speed machining (HSM)
Shaik. Reshma Begum, Koppula. Supriya
The three major manufacturing sectors where High Speed Machining is widely applied are: Aerospace, Die and Mold manufacturing and automotive sectors. The High Speed Machining implies both high velocity machining involving high spindle speeds and at high feed rates with significantly increased spindle speeds. High Speed Machining in fact reduces total machining time as well as reduces bench time. Though high speed machining facilitates faster material removal rates, the process can be made efficient only if optimal combination of machining parameters is selected and tool path is to be optimized. Much research and development work has been carried out in the area of High speed machining of aluminum alloys, titanium alloys, steels and Super alloys. A major factor in the adoption of high speed machining has been the desire to improve tolerances in cutting operations. With high speed machining, most of the heat generated in cutting is removed by the chip, so the tool and the work piece remains close to ambient temperature Apart from that high speed machining leads to reduction in cutting forces that means low power consumption. The study is concerned with the effect of various parameters (cutting speed, feed, depth of cut etc.) on tool life, surface roughness and cutting temperature during HSM process. The data of different experiment are collected and then they are represented in the suitable Taguchi’s table for three different work piece and tool combination (AISI-4140 Steel & Al203 + TiCN Mixed Ceramic, AISI-1117 Steel & Cemented Carbide, Inconel 718 & Al203 + TiCN Mixed Ceramic). After obtaining the data in organized form an ANN (artificial neural network) model was developed to have a clear representation of the data collected. The ANN model will give the user an Optimized value of surface roughness, flank wear and cutting temperature for the given input parameters. As neural network has to train with an optimized amount of data, an L 27 array is used. With the given data the neural network is trained and then the validation of the model is performed i.e. for a given set of input parameters we obtained the value of output and this value is compared with the actual output to ensure the validity of the ANN model. The result from the validation show that the error fell within the range of 0% to 3%. Hence it shows that the artificial neural network model is efficient to correlate the cutting and process parameters and can be integrated into an intelligent manufacturing system for solving complex machining optimization problems.