Development of Manufacturing Simulation Model for Semiconductor Fabrication

This research presents the development of simulation modeling for WIP management in semiconductor fabrication. Manufacturing simulation modeling is needed for productivity optimization analysis due to the complex process flows involved more than 35 percent re-entrance processing steps more than 15 times at same equipment. Furthermore, semiconductor fabrication required to produce high product mixed with total processing steps varies from 300 to 800 steps and cycle time between 30 to 70 days. Besides the complexity, expansive wafer cost that potentially impact the company profits margin once miss due date is another motivation to explore options to experiment any analysis using simulation modeling. In this paper, the simulation model is developed using existing commercial software platform AutoSched AP, with customized integration with Manufacturing Execution Systems (MES) and Advanced Productivity Family (APF) for data collections used to configure the model parameters and data source. Model parameters such as processing steps cycle time, equipment performance, handling time, efficiency of operator are collected through this customization. Once the parameters are validated, few customizations are made to ensure the prior model is executed. The accuracy for the simulation model is validated with the actual output per day for all equipments. The comparison analysis from result of the simulation model compared to actual for achieved 95 percent accuracy for 30 days. This model later was used to perform various what if analysis to understand impacts on cycle time and overall output. By using this simulation model, complex manufacturing environment like semiconductor fabrication (fab) now have alternative source of validation for any new requirements impact analysis