Design And Implementation Of An Air Monitoring System Based On Fine Particulate Matter Tracing And Forecasting

In recent years,the environmental problems brought by the economic development have become prominent increasingly.And air quality has become the focus of public attention.The fine particulate matter(PM2.5)can penetrate deeply into the human alveoli and the blood circulation system,causing a great harm to the human health.Facing the severe fine particulate matter pollution,all levels government departments have already taken comprehensive treatment measures actively,and certain good results have been achieved.However,it is a long process to manage the haze.The relevant departments not only need a long-term governance plan,but also need the scientific and quantitative prevention and real-time monitoring mechanisms to tackle it on the surface and at the root.The purpose of this project is to provide a scheme designed for air monitoring systems.According to the forecasting and tracing of fine particulate matter,we can timely and effectively find out the trends of fine particulate matter changes and the distribution of pollution sources,which can provide strong support for relevant departments to make scientific decision and pursue the investigation of the responsibility.And the accurate prevention guidance can be provided to the public timely.First,by analyzing aerodynamic characteristics of fine particulate matter and development requirements of the monitoring system,a forecasting method of spatial distribution and a short-term change situation of fine particulate matter based on Gaussian processes regression(GPR)are proposed.This method uses the improved particle swarm optimization(PSO)algorithm based on differential evolution operator(DE)to optimize the hyper parameters of covariance function in the Gaussian processes regression,which can improve the accuracy of the hyper parametric adaptive acquisition and achieve the accurate tracking of the unstable changes in fine particulate matter.Second,by analyzing the characteristics of fine particulate matter tracing and system data characteristics,this paper proposes a multi-pollutant dynamic backtracking method based on glowworm swarm optimization(GSO)for fine particulate matter.This method is aimed at the problem of low peak discovery rate and poor convergence efficiency of GSO.It improves the adaptive step size and self-exploration mechanism.It also improves the optimization efficiency and accuracy of the algorithm,and it realizes the accurate positioning of the source of multi-polluted fine particulate matter.Finally,in order to achieve the decoupling of system data flow and modular development,this paper designs a three-tier system architecture contains data sensing layer,data service layer and data application layer.The data sensing layer completes the delivery of upper-layer data through the 433 M wireless data transmission module of SRM1276 based on the broadcast mechanism.The data service layer completes the secure and stable interaction of data through the Kafka distributed messaging system.The data application layer extracts the feature of the sensing data based on fine particulate matter forecasting and tracing method and completes dynamic visualization of feature data through the Baidu map API.