A Novel Non-Intrusive Load Identification Algorithm Based On Hybrid Neural Network And Ensemble Learning

Energy is an important material basis for the existence and development of human society progress.With the development of global economy,energy has become a bottleneck restricting the development of human society.How efficient use of energy and energy conservation has attracted more and more attention.As the main part of energy consumption,how to save energy scientifically and efficiently has become most people concerned problem.The popular method of power saving is to understand the user's behavior habit by analyzing state of appliances,so as to make relevant policies to achieve the goal of saving power.In this paper,a novel load identification algorithm is proposed to solve the problem of non-intrusive load identification.Non-intrusive load identification as a low-cost feasible program gets more and more attention from researchers.Non-intrusive load identification algorithms are divided into two categories: event detection and non-event detection.The non-event detection algorithm has better experimental results.In this paper,a novel non-event detection algorithm is proposed,which is based on the hybrid neural network and ensemble learning.The hybrid neural network has two parts: using Long Short Term Memory(LSTM)to automatically extract time sequence feature of data,using Artificial Neural Network as classifier.And using ensemble learning improve the experimental result after introducing harmonic feature which is high dimensional and unevenly distributed data.Compared with the traditional method,the new algorithm has achieved good results in both recognition accuracy and running time.