基于模型预测控制的生产线在线抽样检测方法OA

In the semiconductor manufacturing scenario,online detection is a necessary part of large-scale mass production of the production line.There is a coupling relationship between online detection and capacity release and yield guarantee.An online sampling inspection method for production lines based on model predictive control is proposed.A switching maximum plus algebraic modeling method for production capacity and a dynamic Bayesian network modeling method for product quality are proposed.The performance equilibrium model of production capacity,quality and inspection cost is constructed,and the maximum benefit of inspection strategy is taken as the objective function.A model predictive control framework dominated by discrete event model is proposed.Taking a chip large board level packaging production line as a case,simulation experiment is carried out.The results show that the production line efficiency is improved under the condition that the average pass rate gap is not large,and the feasibility and effectiveness of the model and algorithm are verified.The model proposed can enrich the theoretical method of modeling and evaluation of production line performance,and can be applied to the engineering problems of on-line detection and production optimization of flow production line.