Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Chine → 2017 → Research and Analysis of Anfis Forecasting Model of Crop Reference Evapotranspiration in Arid Area

Université du Xinjiang (2017)

Research and Analysis of Anfis Forecasting Model of Crop Reference Evapotranspiration in Arid Area

李志磊;

Titre : Research and Analysis of Anfis Forecasting Model of Crop Reference Evapotranspiration in Arid Area

Auteur : 李志磊;

Grade : Master’s Theses 2017

Université : Université du Xinjiang

Résumé
Water saving irrigation is a new technology in agriculture,and its core is to transfer water directly to the roots of crops according to the needs of crops.It is a problem to be solved urgently to judge the crop water requirement accurately.There are two methods to diagnose crop water requirement : direct calculation method and indirect method based on the reference crop evapotranspiration(ET0),this study is the indirect method,the accuracy of ET0 calculation results will directly have an effect on the diagnosis results.There are many methods to calculate the ET0,but it is generally limited by the shortage of meteorological data.With the improvement of intelligent learning,the prediction algorithm overcomes the disadvantages of calculating ET0 when the weather data is short.The meteorological information library and standard ET0 library of 11 sites in the arid area of Xinjiang were established in this paper,Hargreaves-Samani based on the temperature(HS),radiation(P-T)and Priestley-Taylor based on the empirical formula Irmark-Allen(Allen)were compared,the best ET0 calculation model of different meteorological conditions were chosen.Taking PM56 as the standard,the HS method,the radiation method Priestley-Taylor and the empirical formula Irmark-Allen are modified to find the best model of different sites.The best model is the modified HS method for Urumqi,Ruoqiang and Hotan site ;The best model for the site of Urumqi Mushi,Dabancheng,Turpan,Korla,Kashi,Mengaiti,Hami is the revised Priestley-Taylor.In this paper,the adaptive neuro fuzzy inference system(ANFIS)is used as the prediction model of ET0,The simulation results show that the prediction results of the proposed algorithm are closer to the calculated values of PM than those of the modified model.Taking into account the accuracy of the model is affected by the noise in the meteorological environment,Calman filter is used to remove the crop environment information noise.The simulation results show that the modified model can improve the accuracy of the prediction model by comparing with the ANFIS model.In order to apply the algorithm to the actual planting of crops,using Delphi and Matlab mixed programming,the test module of ET0 prediction algorithm is designed,and the irrigation decision module with multiple models is established

Mots clés : Reference Evapotranspiration; Penman Formula; ANFIS Prediction Model; Calman Filt

Présentation (CNKI)

Page publiée le 6 février 2018