The Development Of Grain Condition Monitoring System And The Research On The Algorithm Of Storage Grain Temperature Prediction

On the basis of summarizing the current development situation of grain condition monitoring system,development process of STM32 hardware and software is firstly introduced.And then,implementation process and key technologies of host computer monitoring management system based on C#and SQL Server are illustrated.Based on the deep analysis of stored grain temperature variation rule,this thesis creatively puts forward the "hollowed-out and stratified" algorithm of temperature prediction.The main influencing factors of stored grain temperature in each layer is extracted by regression analysis with SPSS software,and the algorithm for predicting stored grain temperature by traditional GM(1,N)is developed.In allusion to the poor prediction accuracy of traditional GM model caused by the defect of background value selection,a new background value with an adjustment factor is taken.The genetic algorithm is proposed in this thesis to optimize the adjustable factor and an predicting algorithm program is designed to establish GA-GM model.By comparing the traditional GM model and the integral reconstruction GM model,the prediction of GA-GM model is verified to be the most stable and accurate.In order to further improve the predictive effect,time series prediction algorithm is adopted to optimize the forecast error of the GA-GM model.Software Eviews is used to analyze the feasibility of ARMA model error correction algorithm.To increase the applicability of the algorithm,ARIMA automatic modeling program is designed and put into combination with GA-GM model.It can be verified that the combined model further improves the prediction accuracy.A method of future temperature prediction of stored grain based on metabolic GM modeling thought is proposed.The results show that this method is accurate and reliable.Finally,the temperature prediction algorithm is invoked by the condition monitoring system,which realizes the combination of theoretical research and practical application.The functional test proves the grain condition monitoring system with function of stored grain temperature prediction realizes the accurate collection of granary environment data,the stable prediction of stored grain temperature and the reliable monitoring of grain condition,which is of great practical value.