KDE-GRU load interval prediction method considering interval optimization

• Main Authors: ZHANG JIAAN, CAI LEI, LI JIAN, WANG ZHENZHEN, GAO XIUQING, ZHANG TIANMING, FAN RUIQING, GUO LINGXU, ZHAO BEITAO, LI QUANYOU, HAO YAN, CHEN JIAN, LIU FANG, LI FENGXIAN, HAO YI
• Format: Patent
• Language: Chinese; English
• Published: 21.03.2023
• Subjects: CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTINGELECTRIC POWER; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER; ELECTRICITY; GENERATION; SYSTEMS FOR STORING ELECTRIC ENERGY

The invention discloses a KDE-GRU load interval prediction method considering interval optimization. The KDE-GRU load interval prediction method comprises the following steps: 1, obtaining historical load data of a certain region; 2, calculating a load power difference value at adjacent moments; 3, simultaneously decomposing the load variation delta Pt and the basic load Pt into two groups of K load components by adopting an MEMD algorithm; 4, establishing a delta P1 interval prediction model based on a KDE method; 5, establishing a reconstruction component delta P2 point prediction model based on the GRU network; 6, superposing the prediction result [delta P1, min, delta P1, max] of the high-frequency component delta P1 interval and the prediction result delta P '2t of the reconstruction component delta P2 point to obtain a prediction interval of the load variation; 7, establishing a prediction interval optimization model based on the interval optimization coefficient; according to the prediction method, on