QUALITY FORECASTING METHOD

PROBLEM TO BE SOLVED: To provide a quality forecasting method capable of forecasting the normal/defective quality of a casting with high accuracy by evaluating the correlation of the normal/defective quality of the casting and casting conditions from various perspectives. SOLUTION: A plurality of casting conditions for every shot in die casting are collected and the normal/defective deciding results of the casting by operator are inputted. A multiple regression model having casting conditions as each explanation variable and the normal/defective deciding results as objective variable is built (STEP 30). Based on the multiple regression model, the probability (P1) that the casting is non-defective and the probability (P2) that the casting is defective are calculated, and the degree of coincidence of the forecasting deciding result and the normal/defective deciding results based on these probabilities is calculated as a forecasting accuracy (STEP 40). The deciding threshold for the magnitude (¾P1-P2¾) of the probability difference between the probability that the casting is non-defective (P1) and the probability that the casting is defective (P2) is set so that the forecasting accuracy is a predetermined value or above (STEP 50). COPYRIGHT: (C)2010,JPO&INPIT