US 12,222,707 B2
Production schedule estimation method and system of semiconductor process
Chih-Neng Liu, Hsinchu (TW); Chih-Chuen Huang, Hsinchu (TW); Chia-Jen Fu, Hsinchu (TW); and Chih-Hsiang Chang, Hsinchu (TW)
Assigned to Powerchip Semiconductor Manufacturing Corporation, Hsinchu (TW)
Filed by Powerchip Semiconductor Manufacturing Corporation, Hsinchu (TW)
Filed on Sep. 30, 2021, as Appl. No. 17/491,502.
Claims priority of application No. 110131566 (TW), filed on Aug. 26, 2021.
Prior Publication US 2023/0066892 A1, Mar. 2, 2023
Int. Cl. G05B 19/418 (2006.01); G06N 20/00 (2019.01)
CPC G05B 19/41865 (2013.01) [G06N 20/00 (2019.01); G05B 2219/32252 (2013.01)] 12 Claims
OG exemplary drawing
 
1. A production schedule estimation method of a semiconductor process, comprising:
recursively executing:
obtaining a current-day work-in-process data, a cycle time data of a machine group, and a productivity data of the machine group;
inputting the current-day work-in-process data, the cycle time data of the machine group, and the productivity data of the machine group into a prediction model;
calculating a current-day cycle time data and a current-day move volume for each of a plurality of stations in the machine group through the prediction model;
calculating a current-day move data according to the current-day cycle time data and the current-day move volume for the each of the plurality of stations in the machine group, wherein the current-day move data comprises a type of a work-in-process and a current-day move volume after the work-in-process has been processed by the plurality of stations in the machine group on the current day;
calculating a next-day work-in-process data according to the current-day cycle time data and the current-day move volume for the each of the plurality of stations in the machine group, and executing the prediction model, so as to obtain a product production time and a quantity of a product produced;
summing a plurality of move volumes corresponding to each of the plurality of stations in a plurality of machine groups on different days generated according to above recursive execution to generate the quantity of the product produced and the product production time;
estimating a production schedule according to the product production time and the quantity of the product produced; and
controlling the machine group based on the estimated production schedule,
wherein the step of obtaining the cycle time data of the machine group and the productivity data of the machine group comprises:
obtaining a work-in-process historical data and a machine historical data;
performing a feature extraction on the work-in-process historical data and the machine historical data, so as to obtain a plurality of feature data; and
inputting the plurality of feature data into a machine learning module, to enable the machine learning module to perform a prediction to output the cycle time data of the machine group and the productivity data of the machine group.