Interactive Design-of-Experiments: Optimizing a Cooling System

Area of Science:

• Thermodynamics and Heat Transfer
• Computational Engineering
• Artificial Intelligence in Engineering

Background:

• Cooling system optimization is critical for applications like electric vehicle cabin and battery thermal management.
• Multi-objective optimization in complex, non-linear systems is challenging due to large parameter spaces and long simulation times.
• Existing methods often require semi-automatic, iterative approaches to handle simulation constraints and unmodeled factors.

Purpose of the Study:

• To present an interactive visual optimization approach for complex cooling systems.
• To integrate deep learning (DL) for parameter estimation and numerical simulation for system characteristic computation.
• To enhance user comprehension and accelerate optimization convergence through novel visualization techniques.

Main Methods:

• An iterative, guided optimization process steered by the user via a p-h diagram.
• Utilizing a deep learning model to estimate system parameters based on target characteristics.
• Employing numerical simulations to compute system characteristics for parameter ensembles.
• Integrating a dual, interactive visual representation alongside the standard p-h diagram.

Main Results:

• The interactive visual approach facilitates user comprehension of the cooling system.
• The method leads to faster convergence in the optimization process.
• The dual visual representation effectively complements the p-h diagram, transferring its semantic meaning.
• Multiple sets of intermediate solutions are generated and visually related, aiding the iterative process.

Conclusions:

• The proposed interactive visual optimization approach enhances understanding and efficiency in cooling system design.
• The integration of DL models and advanced visualizations offers a powerful tool for complex engineering optimization.
• This method shows significant promise for applications in the automotive industry, particularly for electric vehicles.