摘要
机场噪声预测对噪声的控制、机场周边的规划和航班计划的制定具有重要的指导作用。传统的机场噪声预测模型一般根据航空器NPD(噪声-推力-距离)曲线为基础预测噪声,缺少精确考虑特定外界条件对噪声传播的影响作用,导致其预测误差较大,但优点是预测较稳定。而基于机器学习的机场噪声预测模型虽综合考虑了特定外界环境对噪声的影响,但由于实际训练数据来源于实际监测点监测的历史数据,经常包涵错误信息,导致建立模型不准确等问题。针对上述问题,依靠实际监测点的噪声、气象数据,构建了朴素贝叶斯机场噪声修正预测模型,通过学习传统预测模型预测值相对于监测值的差值,修正传统预测模型由于客观外界因素造成的预测偏差,既保持传统模型预测稳定性,同时修正噪声关于外界环境造成的声音衰减。最后,通过对比实验可见,改进方法预测稳定性较高且具有一定的预测准确度。
The airport noise prediction plays an important role on the noise control. The traditional prediction model for airport noise prediction is based on the aircraft's noise-power-distance( NPD) curve. It has a big error because it is lack of consideration of the effect of meteorological conditions on noise propagation. Although the model based on machine learning considers the impact of the external environment,it is relys on historical data and predicts instability. This paper proposed the nave Bayesian network learning model for airport noise prediction correction which learns the difference between the prediction value of traditional model and the actual monitoring value. Finally,The experimental results show that this method has high prediction stability.
出处
《计算机仿真》
CSCD
北大核心
2015年第7期36-41,共6页
Computer Simulation
基金
国家自然科学基金重点项目(61139002)
国家"863计划"项目(2012AA063301)
国家科技支撑计划项目(2014BAJ04B02)
中国民用航空局科技项目(MHRD201006
MHRD201101)
中央高校基本科研业务费专项资金(3122013P013)
关键词
机场噪声预测
朴素贝叶斯
聚类
集成
Airport noise prediction
Naive bayesian
Clustering
Ensemble