1. Household load forecasting model based on Markov state transition
- Author
-
Shao, Chuanyong, Zhaobin, DU, CHAUVEAU, Eric, Lidan, CHEN, South China University of Technology [Guangzhou] (SCUT), ESEO-AGE (AGE), ESEO-Tech, Université Bretagne Loire (UBL)-Université Bretagne Loire (UBL), Institut de Recherche en Electrotechnique et Electronique de Nantes Atlantique EA4642 (IREENA), Institut Universitaire de Technologie Saint-Nazaire (IUT Saint-Nazaire), Université de Nantes (UN)-Université de Nantes (UN)-Ecole Polytechnique de l'Université de Nantes (EPUN), Université de Nantes (UN)-Institut Universitaire de Technologie - La Roche-sur-Yon (IUT La Roche-sur-Yon), and Université de Nantes (UN)-Université de Nantes (UN)
- Subjects
[SPI]Engineering Sciences [physics] ,[SPI.AUTO]Engineering Sciences [physics]/Automatic - Abstract
International audience; Abstract：The development of smart grid improves the emphasis on family load forecasting. Based on the theory of state transition, a Monte Carlo Markov Chain load forecasting model of single equipment based on the selection of similar days is proposed and the bottom⁃up analysis method to obtain the comprehensive load level of a single family is used. For the temperature control equipment, Pearson correlation coefficient is used to study the correlation between the ambient temperature and the operation cycle of the equipment, and the hidden Markov model is used to predictthe operation state of the compressor of the temperature control equipment according to the external environment information of the day. With the operation state predicted, the average power in different time periods is calculated to reflect the user load level. The simulation results show that the predicted error of the Monte Carlo Markov chain model based on the similar day selection is about 2% ~8% for the daily load expectation of different equipment, while the predicted accuracy of the hidden Markov model for the temperature control equipment is about 70%.
- Published
- 2021