基于隐马尔可夫模型的用户用电状态识别

The Pattern Recognition of Residential Power Consumption Based on HMM

针对生活用电个体用电量小、波动性强、难以预测识别的问题,提出一种利用隐马尔可夫模型对不同场景的用电状态识别方法。从单用户单电器、单用户多电器和多用户多电器逐渐深入分析,研究学习和解码的主要算法,并求解关联概率矩阵。最后,应用该方法对私人电器和公用电器两个场景进行用电活动自识别,准确率达90%以上。

Aiming at the problems of small electricity consumption, strong volatility and unpredictable identification of household electricity consumers, this paper discusses the different scenarios of power use the Hidden Markov Model （HMM） modeling method of pattern recognition, from single-user single-appliance, single-user multi-appliances, multi-users multi-appliances gradually in-depth analysis. Secondly, the main algorithms of learning and decoding are introduced, and the concept of association probability matrix is introduced. Finally, this method is applied to the self-identification of power consumption in two scenarios of private and public appliances, and the accuracy is more than 90%.