基于贝叶斯估计的气体污染源方位辨识

气体污染源的方位辨识是监控部分企业废气超标排放及治理空气污染的一项重要措施。针对气体污染源的方位辨识问题,设计了基于贝叶斯概率模型的污染源方位辨识的多传感器检测系统,该系统使用4个气体传感器组建检测系统,4个传感器按照90°间隔均匀等半径放置,系统可辨识16个方向,分辨角度为22.5°。首先对八个标准方向进行标定,确定八个方向的先验概率,再根据八个方向的先验概率和4个传感器检测到的实时样本值,由贝叶斯模型计算各方向后验概率,进而由各后验概率来确定气体污染源的方位。实验表明,基于贝叶斯模型的气体污染源方位辨识方法是可行的、有效的。

Identification of direction of air pollution source based on Bayesian estimation

Identification of the direction of air pollution source is the key for monitoring some enterprises whose discharges of exhaust gases exceed the standard and cleaning the air. Aiming at this problem,a multi-sensor system based on Bayesian mathematical model,which can identify the direction of air pollution source,is designed in this paper. Four gas sensors are used to form the detection system,which are fixed evenly on the same circle every 90 degrees. The system can identify 16 directions and the resolution of the identification is 22. 5 degrees. Firstly,eight standard directions are calibrated,and the prior probability of each direction should to be confirmed. Then the posteriori probability of each direction is calculated by Bayesian mathematical model according to the prior probability of each direction and real-time sample values of air pollution source detected by sensor. And then the direction of air pollution source is decided based on posteriori probabilities. The experiments indicate that the method to identify the direction of air pollution source based on Bayesian mathematical model is reasonable and effective.