Real-time data quality control method of automatic weather station based on PCA and PSO-ELM

Abstract

The present invention discloses a real-time data quality control method of an automatic weather station based on the PCA and the PSO-ELM. The method comprises firstly collecting temperature data within a certain historical time scale of the observation station (an adjacent station) within a certain range from a to-be-detected station, and constituting an original collection signal; carrying out principal component analysis on the collection signal to achieve the purpose of redundancy; constructing a reconstruction model of the temperature according to the improved limit learning machine, and according to the reconstruction model of the temperature, obtaining an estimated value of the to-be-detected time through data in the adjacent station of the to-be-detected time; and finally, according to comparison between the estimated value and the actual observation value, carrying out conditional amendment so as to complete quality control of the temperature in the time. According to the method disclosed by the present invention, based on the basic quality control method, on one hand, fixation of the adjacent station selection in the current multi-station network quality control method is made up, and the dynamic adaptability of the algorithm is improved; and on the other hand, the improved limit learning machine improves the generalization performance of the network, and the accuracy of the reconstruction model is improved.

Description

technical field [0001] The invention relates to the field of quality control of real-time data collected by an automatic weather station, in particular a quality control method for real-time data collected by an automatic weather station. Background technique [0002] In recent years, the number of surface meteorological observation stations has been increasing. Ground automatic weather stations have distribution characteristics such as dense station distribution, large terrain differences, and harsh station environments, which determine that there are systematic errors, random errors, gross errors, and micro-meteorological errors in the observation data. , which will affect the quality of observation data of automatic weather stations, thereby affecting the accuracy of climate change, climate model research and short-term numerical weather prediction. Therefore, quality control of the collected data is required. The existing three-level quality control business in my count...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a method for quality control of surface air temperature observation data based on principal component analysis and improved extreme learning machine, which solves the problem that the current quality control method cannot eliminate random errors, systematic errors, and micro-meteorological errors, and improves the quality of automatic meteorological data. The quality of the real-time observation temperature of the station

Images