Design Optimization of Air Duct for Noise Reduction Using Gaussian Process Regression Algorithm

In the context of Battery Electric Vehicles (BEVs), airborne noise from Heating, Ventilation and Air Conditioning (HVAC) ducts becomes a prominent concern in the view of passenger comfort. The automotive industry traditionally leverages Computational Fluid Dynamic (CFD) simulation to refine HVAC duct design and physical testing to validate acoustic performance. Optimization of the duct geometry using CFD simulation is a time-consuming process as various design configurations of the duct have to be studied for best acoustic performance. To address this issue effectively, the proposed a novel methodology uses Gaussian Process Regression (GPR) to minimize duct noise. Present solution demonstrates the power of machine learning (ML) algorithms in selecting the optimal duct configuration to minimize noise. Utilizing both real test data and CFD results, GPR achieves remarkable accuracy in design validation, especially for HVAC air ducts. The adoption of GPR-based ML algorithms significantly enhances the accuracy and cost-effectiveness of air duct design. This approach accelerates the development process, ensuring quicker design optimization. Present article focus on an efficient solution for rapidly achieving optimized acoustic designs, surpassing conventional validation methods. ML techniques, particularly GPR provides an optimal solution for improving the passenger comfort in the car.