Application of AI Optimization in Mold Flow Analysis

Traditional optimization of injection-molded product quality relies heavily on Trial & Error or DOE methods, but their efficiency and flexibility are limited when dealing with multi-objective, constrained problems. Such problems often involve conflicting quality factors, including temperature, pressure, warpage, and flatness, requiring effective tools to support decision-making. This study investigates how AI-driven iterative optimization can be applied alongside molding simulation, exploring potential application strategies. In the early stage of product design, DOE is employed to identify key factors and overall trends, helping engineers understand the most influential parameters. In the later stage of local parameter refinement, where adjustments previously relied solely on DOE and often required extensive fine-tuning, AI now provides an additional option. Engineers can leverage AI-driven iterative search to explore the parameter space more dynamically and rapidly converge toward optimal solutions, complementing DOE results. This approach allows more flexible decision-making and efficient identification of optimal solutions. Three practical scenarios illustrate the application. In multi-objective optimization, AI automatically identifies multiple feasible solutions along the Pareto front, enabling engineers to quickly understand solution characteristics through visualization tools and select the most appropriate strategy. In more complex multi-objective, constrained scenarios—where engineers aim to minimize sprue pressure and warpage displacement, ensure geometric dimensions and functional tolerances (GD&T) are maintained, and consider material selection and cycle time—a two-step strategy (DOE followed by AI iteration) effectively captures overall trends and rapidly converges to optimal design solutions, demonstrating the complementary advantages of the tools. Additionally, when certain experimental conditions fail—such as short shots or otherwise infeasible setups—AI can dynamically flag these points and redirect the search toward alternative directions, maintaining exploration efficiency and demonstrating adaptive flexibility. Overall, the combined use of DOE and AI enhances efficiency in solving multi-objective, constrained optimization problems. Visualization tools, including iteration curves, Pareto front identification, and parallel coordinate plots, facilitate understanding of factor relationships and accelerate the discovery of optimal solutions. This approach significantly improves decision-making efficiency and flexibility in product and mold development, highlighting the practical value of AI-driven optimization for complex engineering applications.