Summary: | 碩士 === 雲林科技大學 === 環境與安全工程系碩士班 === 96 === This study takes the maximum noise value from both day and night during every season and the five types of aviation noise monitored and recorded from a data base at a certain southern Airport. It sets the maximum noise events data as a Dependent variable parameter and meteorological data of day and night in every season and five type events as an Independent variable parameter. These include wind direction, wind speed, temperature, relative humidity, and atmospheric pressure. The Pearson Correlation Coefficient method and P-Value were use to find correlation between Dependent variable parameters and Independent variable parameters. it chooses the maximum correlation umber and through Artificial Neural Networks, develops a model for predicting the maximum noise level of both day and night in every season and the five types of aviation events. The result of research reveals that day’s noise amplitude at the airport was larger than the night amplitude in every season. It also showed that the noise correlated with temperature and atmospheric pressure. After processing day and night’s temperature and atmospheric pressure factors with related noise value, an artificial neural network method was applied for the prediction of day and night’s noise level. The results showed the airport noise level amplitude was too high causing prediction error to be too large. But night’s noise level amplitude was smoother; the prediction result was more precise. In the prediction model of maximum noise level of five types of aviation events, no matter in simulation or prediction, the result was more precise. This has proved that positive correlation exists between the aviation noise level and meteorological parameters. It is available to predict aviation maximum noise level through the meteorological parameters in the future, and improve noise monitor system in operation and provide reference for ruling association policy.
|