The advancement of intelligent manufacturing is driving the evolution of polymer processing into a new developmental stage. Artificial intelligence technologies

particularly machine learning (ML)

have become key enablers for overcoming the limitations of traditional empirical methods and achieving intelligent optimization of processing workflows. In response to the current gap between ML education and real engineering applications in polymer materials curricula

this study introduces a teaching innovation that integrates data-driven intelligence with professional practice through an undergraduate hollow blow molding experiment. Specifically

a hybrid approach combining a BP neural network with a genetic algorithm was applied to optimize five critical process parameters in hollow blow molding: die temperature

mold closing speed

airflow rate

blowing time

and cooling time. A predictive model was developed to identify the optimal parameter set for maximizing the vertical compressive strength of molded parts. Experimental validation confirmed the effectiveness of this approach

establishing a complete instructional loop encompassing component fabrication

data collection

model building

algorithmic optimization

and result verification. The redesigned course not only strengthens students’ mastery of conventional polymer processing techniques but also guides them through the entire workflow of applying ML to real-world process optimization. This initiative effectively bridges ML theory with practical manufacturing scenarios

while deepening students’ understanding of the “process-structure-property” relationship in polymer science and fostering computational thinking framed around “data-model-decision”. The project offers a replicable model for cultivating talent capable of supporting the transformation and upgrading of the intelligent polymer manufacturing industry.