A bottom-up customizable Markov model for household appliance electric consumption
Résumé
Residence load is an important part of the total power consumption. This paper develops a bottom-up non-intrusive method to build a predictive model for the residence load using a Markov model. Taking the fridge as the experimental subject, the total load is decomposed into three subsystems. The applicability of this Markov model is proved by the simulation in the Matlab TM platform. To do this, a pair of Markov chain is then created to predict the load power consumption for historical days. Then the average power of simulation and measurement are compared, as well as the state transitions along time. The results clearly shows that Markov chains can simulate the behavior of the residence load in a good way, but considering the influencing factors of the environment, this Markov model should be improved to obtain a better accuracy.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |