基于变权欧氏距离的学习型智能插座设计

使用现有智能插座实现电器用电量的统计和电器待机功耗的消除,存在灵活性差、适用范围有限等问题,为此提出一种基于变权欧氏距离的学习型智能插座.先采用有限状态机方式提取有功功率和功率因素作为待测电器的特征值,再用变权欧氏距离从样本库中找出待测电器的类型,引入学习功能解决电器多样性问题,并用记忆原理处理陈旧的电器样本,在识别电器的基础上,统计各个电器的用电量和消除电器的待机功耗.测试结果表明,同一类型电器之间的相似度达80%以上,可识别出宿舍中大多数电器的类型,验证了该方法的有效性和可行性.

Design of a learning smart socket based on Euclidean distance with varying weights

There exist some issues, such as limited flexibility and applicability, on using conventional smart sockets to implement the statistic of power consumption of electrical appliance and the elimination of standby power of electrical appliance. To address the issues, a new kind of learning smart socket based on Euclidean distance with varying weights is proposed in this paper. Firstly, the active power and power factor were extracted as eigenvalues of the measured electrical appliance through the finite state machine; the type of measured electrical appliance was then found from the sample library using the Euclidean distance; the problem of electrical appliance diversity was solved by introducing a learning function; the old samples were removed using the memory principle; the statistic of power consumption of electrical appliance and the elimination of standby power of electrical appliance were realized based on the recognition. The test result show that, the similarity among the same type of electrical appliance is more than 80%, and it can recognize most types of electrical appliances in dormitories. Therefore, the validity and feasibility of the presented method is thus demonstrated.