Automatic Weather Station Of Six Elements Based On ARM

The data from automatic weather stations can't be used if it isn't subjected to quality control,this has been reached a consensus by meteorologists.At present,research of quality control methods is in constant development in China.With the building of more and more regional automatic weather stations,the emphasis of research turns to spatial methods,while the research for single station is limited.Existing methods of single station are not only very single,but also imperfect for the flow,this makes the data that appears in extremes can't be judged accurately,and shows the shortage of quality control ability.With the limit of geographical position,environmental condition,economic development and so on,spatial methods can't be used adequately in applications.Meanwhile,with the development of technology for automatic weather stations,it is worth reconsidering that whether the existing methods can meet the demands of processing data in real time and quickly.The quality control system for real-time data is researched and developed aimed at above questions.The system takes Visual Studio 2008 as development tool,SQL Server 2008 as database,C#as development language which achieves the functions of real time monitoring,management and so on.According to climatology characteristics of each element and properties of automatic weather stations,quality control methods which fit the single station are researched and the flow of quality control is improved.The system achieves the demands of rapidity,instantaneity and stability for processing large volumes of data.Meanwhile,support vector machine based on genetic algorithm optimization is applied to quality control of substantial drop in temperature,a model of quality control is built,which effectively reduces the possibility of false detections in extreme weather and makes up the shortage of quality control capacity when the automatic weather station running stand-alone.According to the noise in the data,fuzzy support vector machine based on root-mean-square membership function is proposed,which improves generality and applicability of the model and ensures the effect of quality control.Finally,there are some questions and shortages in the process of system implementation which need to be further refined and improved.