A prediction model for the lost wax process through fuzzy-based artificial neural network

The application of investment casting process is rapidly increasing, specifically for near net shape manufacturing of complex and small engineering components. The process begins with making of wax patterns, thereafter employing a precision mould, dewaxing, pouring molten alloy and knocking the shell, followed by minor finishing operations. This study is about predicting the quality of responses of the wax patterns namely linear shrinkage and surface roughness using fuzzy-based artificial neural network. The process parameters considered are injection temperature, injection pressure and holding time, and experiments have been performed as per Taguchi’s L18 orthogonal array. As optimum parameter levels were different for both the responses, fuzzy logic reasoning has been used to combine all the objectives and transform the experimental results into single performance index known as multi-response performance index. Later, modelling of the process has been done using artificial neural network with experimental process parameters as inputs and multi-response performance index as output obtained from fuzzy modelling. Further, experiments have been conducted at random combination of parameter levels to validate the developed model, and it has been found that the actual results agreed well with that of the predicted value on the basis of mean absolute percentage error and correlation plots.

[1]  Jae-Seob Kwak,et al.  Application of Taguchi and response surface methodologies for geometric error in surface grinding process , 2005 .

[2]  Pradeep Kumar Jha,et al.  Influence of injection process parameters on dimensional stability of wax patterns made by the lost wax process using Taguchi approach , 2013 .

[3]  A. Clegg,et al.  Precision Casting Processes , 1991 .

[4]  Senay Taşcıoğlu,et al.  A novel alternative to the additives in investment casting pattern wax compositions , 2003 .

[5]  Phillip J. Ross,et al.  Taguchi Techniques For Quality Engineering: Loss Function, Orthogonal Experiments, Parameter And Tolerance Design , 1988 .

[6]  Sameh Habib,et al.  Study of the parameters in electrical discharge machining through response surface methodology approach , 2009 .

[7]  Sadegh Rahmati,et al.  Dimensional accuracy analysis of wax patterns created by RTV silicone rubber molding using the Taguchi approach , 2007 .

[8]  Hyungsuck Cho,et al.  A fuzzy logic controller for the molten steel level control of strip casting processes , 2005 .

[9]  Genichi Taguchi,et al.  Introduction to quality engineering.... , 2014 .

[10]  T. Luck,et al.  Application of Plant Proteins as Thermoplastics , 1998 .

[11]  Senay Taşcıoğlu,et al.  Conversion of an Investment Casting Sprue Wax to a Pattern Wax by Chemical Agents , 2003 .

[12]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[13]  S. Mahapatra,et al.  A modeling approach for prediction of erosion behavior of glass fiber–polyester composites , 2008 .

[14]  Guo H. Huang,et al.  Application of soft computing models to hourly weather analysis in southern Saskatchewan, Canada , 2005, Eng. Appl. Artif. Intell..

[15]  A. Behravesh,et al.  An experimental investigation on dimensional stability of injected wax patterns of gas turbine blades , 2007 .

[16]  Angelos P. Markopoulos,et al.  Artificial neural network models for the prediction of surface roughness in electrical discharge machining , 2008, J. Intell. Manuf..

[17]  Rawson L. Wood,et al.  Investment castings for engineers , 1952 .

[18]  Y. Unigovsky Regression analysis of the mechanical properties—Composition dependencies for cast low- and medium-carbon steels , 2000 .