The Study Of Massive Atmosphere Particulates Components Online Analysis System Based On Realtime Storage

In recent years,the domestic fog haze weather,a wide range,a long time,a serious threat to the health of the people,haze governance has aroused great concern of the government and society.Because of the different air pollution in different cities,and affected by the geographical location,weather conditions,industrial composition,urban pattern and other factors,so the management of environmental pollution must be on the source of urban pollution and qualitative quantitative scientific research.Atmospheric particulate matter monitoring and analysis is an important means to understand the quality of air,and the traditional means of atmospheric particulate analysis mainly rely on the overall analysis of particulate matter technology,artificial identification of particulate matter and source analysis,these technical means have obvious defects:(1)The traditional analytical method of particulate matter can not reflect the internal characteristics of particulate matter;(2)The traditional relational database does not apply to this scenario;(3)the traditional manual analysis means time-consuming,high labor costs,low accuracy,in the face of large amounts of data can not do anything,We urgently need an automatic particle analysis technique.In this paper,an on-line analysis system of massive atmospheric particulate matter based on real-time storage technology is designed.The system consists of two subsystems,which are massive data storage subsystem RyDB and online analytical subsystem.The underlying data storage system,RYDB,is a KV NoSQL database which uses the levelDB storage engine to support master-slave replication and cluster deployment.The upper-line on-line analysis system uses adaptive resonance theory(ART)network clustering and logical regression classification and other data mining technology to achieve the classification of particulate data statistics,source analysis and other functions.Experiments show that the data storage system RyDB has excellent performance,and can read and write up to 100,000 times per second in the test environment.It has the characteristics of high throughput and low delay.The experiment of the on-line analysis system of particles shows that the system has strong timeliness,32 million particles can be analyzed within two hours,the accuracy of particle classification is 80% or more,to meet the system requirements.