The Research And Application Of Airport Noise Annoyance Model

With the rapid development of civil aviation, the problem of noise pollution has become more and more serious. Noise exposure arising from air traffic has been the main cause of conflicts between airports and nearby communities in most major cities around the world.Airport noise annoyance can better reflect the mental and psychological impact which was because of airport noise, can effectively evaluate airport noise influence, and play an important role in the airport noise pollution treatment and airport noise control.This paper focus on research and construct a model of airport noise annoyance, and the main work is as follows:In order to filter the key factors influence airport noise annoyance, this paper design airport noise annoyance questionnaire. On the basis of questionnaire data collection and preprocessing, based on frequency analysis method and description analysis method, paper makes clear the key factors, which are noise level, noise occurrence period of day, and noise impact zones.Because the noise annoyance is a fuzzy concept, the paper builds the airport noise annoyance model based on fuzzy logic, which through fuzzy key factors, and formulate noise annoyance fuzzy rule. Different from previous qualitative assessment of crowd annoyance caused by airport noise, this model evaluate the airport noise influence in a quantitative way.Based on a large real busy international airport noise monitoring data, the experimental results show the model is more intuitive, and more convenient for people to understand.As the fuzzy of key factors lacks of acknowledged standards, the fuzzy operation is with certain degree subjectivity. The neural-fuzzy system is integration of fuzzy logic and neural network, and be better deal with the uncertain problems and model learning. So this paper adjusts the parameters of fuzzy membership function, optimizes the airport noise annoyance model based on neural-fuzzy system. In addition, in view of the existing problems which are large amount of calculation, slow convergence speed, and low learning efficiency when using the conventional neural-fuzzy learning algorithm based on gradient, this paper proposes amixed learning method of airport noise annoyance model. Experiments show that mixed method not only can optimize model, and can improve the convergence speed, shorten the learning time, and reduce the calculation in the process of solving the error. Based on a large busy airport noise monitoring data, the experiments show that this model can effectively assess the airport noise effects in different noise impact zones.