Knowledge-based injection molding optimization framework under real manufacturing constraints

Injection molding parameter selection requires the simultaneous balancing of product quality, cycle time, and process stability. The inconsistencies among shop-floor, polymer datasheet, plastic simulation, and Subject Matter Expert (SME) recommended parameters may lead to uncertainty and impractical solutions in real manufacturing. Existing optimization and simulation-based approaches fail to account for this issue, thereby limiting industrial applicability. Consequently, this study proposes a knowledge-based integrated framework for robust and industry-applicable parameter optimization. The methodology employs Grey Wolf Optimization (GWO) as the primary optimization tool to perform global exploration and local exploitation of process parameters within defined bounds and ideal targets, as well as under constraint-driven conditions without predefined ideal values based on surface-finish requirements. Taguchi analysis and regression modeling are utilized to identify important parameters and provide statistical validation. The results demonstrate that melt temperature is the primary factor governing sink marks and cooling time, while mold temperature primarily controls volumetric shrinkage, confirming thermally driven process behavior. Although both parameter ranges, namely SME-guided theoretical ranges defined prior to production and SME-adjusted practical ranges used at the shop floor, originate from SME knowledge, they differ due to presumptions versus real process conditions. By incorporating these practical SME constraints into the GWO search, the framework achieves near-zero deviation, including 0% deviation in injection pressure, $$-3.45\%$$ in mold temperature, $$+0.18\%$$ in melt temperature, $$-8.33\%$$ in hold time, $$-11.11\%$$ in cooling time, and $$-6.94\%$$ in cycle time, from shop-floor conditions while effectively balancing quality and productivity, thereby enabling stable and industry-ready parameter selection. Further independent shop-floor validation showed close agreement with the GWO recommendations and achieved acceptable quality within three gate-seal trials.