Application Of Artificial Neural Network For Defect Prediction In Ceramic Tiles Manufacturing
The potential of using artificial neural network (ANN) in prediction the defects in tile manufacturing process have been presented in this study. This study offers a method capable of predicting the influence of each stage of manufacturing the ceramic tiles on the quality of the produced tiles. The tile data to be considered and collected for the preparation, pressing, and firing stages are as follows; tile thickness (THK), dry mechanical resistance (DMR) were collected in the preparation stage, water absorption (WA), Shrinkage (SHR) were collected in the pressing stage while firing mechanical resistance (FMR), and tiles temperature (TEMP) were collected in the firing stage. All the previously mentioned factors and data were used as input data while concavity defect (CD) and surface defects (SD) in tile manufacturing process were the output of the process. The proposed model is proved experimentally by using different sets of inputs..
Index Terms - Artificial neural network, Tile manufacturing, Ceramic, Surface defects.