Research On Storm Recognition And Forecast Model Of "Train Effect"

In recent years，disastrous weather occur constantly in our country. It causesgreat losses upon our national economy, agriculture, livelihoods and zoology. How dowe timely reports for the weather and prevent disasters to a large extent needs furtherresearch and discussion. Tools and means about forecast and recognition need to beprovided to monitor and analysis the weather phenomenon.The features of “Train Effect” were obtained by historical samples in the paper.Those monitored “Train Effect” were recognized by contrasting and analyzing realtime data, and forecasting and extrapolation were gotten.The major work included:(1) Historical samples were observed, and the features and the entiretydevelopment law of “Train Effect” were grasped. Through these preparing, thefeatures of “Train Effect” were extracted manually, and then the data was entered intothe database. Rules were excavated by the tool about rough sets and the model wasformed.(2) After preprocessing the real-time, the method of fitting line and clusteringabout monomer was proposed, and several processes of “Train Effect” were obtained.“Train Effect” was judged by comparing with the models, and target area wasconfirmed.(3) Through entering into several data continuously, time-series was updated andgenealogy of cloud cluster was gotten. The method of extrapolating and forecastingabout “Train Effect” was proposed by motion inertia of monomer and whole cloudband.In conclusion, the real-time surveillance and recognition of “Train Effect” wasrealized, and the development tendency, position and strength were forecasted byusing the above method.