Affiliations: Department of Industrial Engineering, Faculty of Engineering, Prince of Songkla University, Songkhla, Thailand | Industrial Engineering Program, Sirindhorn International Institute of Technology, Thammasat University, Pathum-Thani, Thailand
Note: [] Corresponding author. Supapan Chaiprapat, Department of Industrial Engineering, Prince of Songkla University, Songkhla, Thailand. Tel.: +66 7428 7025; Fax: +66 7455 8829; E-mail: [email protected]
Abstract: Wood is a natural derivative, it possesses randomly distributed inherent defects all over its mass. This complicates the cost estimation process; data collected showed that even planks with similar defect pattern would have different percentage of material loss. Such uncertain loss was caused by changes in cutting parameters. In this study, a Fuzzy Inference (FI) method was employed to predict wood loss in a rubber wooden toy manufacturing cutting process. Notable variables are: length of cut, and area of cut. Prediction accuracy from the FI method was compared with that of other alternative methods. Common practice assumes a constant defective proportion, resulting in inaccurate cost estimation. A regression equation allows the loss to be varied by the cutting parameters; but only one parameter was found to be significant. Experimental results show that the FI method greatly outperformed regression and conventional methods. These findings emphasize the influence of cutting parameters on product cost. Accurate cost estimation enables better planning for efficient pricing strategies and enhances business competitiveness.
