Machine learning method predicting thermal performance of conformal cooling systems.

Area of Science:

• Manufacturing Engineering
• Materials Science
• Computational Engineering

Background:

• Conformal cooling systems significantly improve injection molding efficiency and product quality.
• Current optimization relies on labor-intensive simulations and human expertise, limiting design exploration.
• Automated methods for channel creation exist but lack efficient thermal performance optimization.

Purpose of the Study:

• To develop an innovative machine learning (ML) method for assessing conformal cooling system thermal performance.
• To enable precise prediction of thermal behavior without extensive manual simulation.
• To empower designers in optimizing conformal cooling design parameters efficiently.

Main Methods:

• A hybrid ML approach combining a non-linear regression model and a neural network was employed.
• A logarithmic regression model was used to describe temperature profiles.
• A neural network was trained to predict the coefficients of the logarithmic regression model.

Main Results:

• The proposed ML methodology accurately assesses and predicts the thermal performance of conformal cooling systems.
• The method significantly reduces the need for time-consuming thermal and fluid simulations.
• It provides designers with a rapid and effective tool for evaluating thermal efficiency.

Conclusions:

• The ML-based approach offers a precise and efficient alternative to traditional methods for conformal cooling design optimization.
• This innovation streamlines the design process, enabling faster iteration and improved mold performance.
• The methodology facilitates better thermal management in injection molding through data-driven design insights.