Research on Noise Mode Prediction of Air Conditioner Based on Training Neural Network Method

Abstract

This research establishes forecasting models for air-conditioning noise through training-type neural networks, which can improve the accuracy of the forecast, reduce costs and time, optimize system design and improvement, improve user experience and reduce environmental impact. This study uses Matlab to build training neural networks and uses the Levenberg-Marquardt. The results showed that the neural networks built by the training methods and the data and model combination were excellent. Therefore, the neural network to establish the model for the forecast of air conditioning noise has a significant effect for future air conditioner companies in the research and improvement of the noise.