摘要
家用负荷识别可提高用户对用电情况的认知度,优化用电模式,响应节能政策。提出一种分步识别的方法。识别前依据谱聚类方法得到负荷类别及其聚类中心,建立标准模板库,存储各负荷类别的特征量,特征量主要包括6项:负荷投入或者切除时刻变化的暂态有功功率、无功功率波形标幺值及各自幅值和稳态的有功功率、无功功率值。首先以综合负荷的功率变化为依据,提取负荷变化时刻及稳定运行后的功率确定特征量。第一步粗选,依据其投入或切除瞬时波形进行调整后与标准模板库波形进行匹配确定负荷所在大类;第二步精选,依据其瞬时波形幅值与稳态运行功率值与所在类别中的负荷相应特征量进行比较得到识别结果。该方法能精准确定负荷类别,简单可靠,可为家庭负荷建模提供数据支持。实例也验证了该方法的有效性。
Household load identification can improve users＇ awareness of electricity consumption, optimize the power mode, and make response to the energy saving policy. A step identification method was proposed in this paper. The load classification and clustering center were obtain by the spectral clustering method before identification, and a standard template library was established to store various load category features, including the standard value and am- plitude of transient active and reactive power waveforms, and the values of steady active and reactive power. First of all, based on the power change of comprehensive load, the characteristic quantities of power were extracted when the load changed and after stable operation began. The first step was rough election, in which, the classification of the load was determined on the basis of the waveform matching between the adjusted instantaneous waveform and that in the standard template library; and the second step was careful selection, in which the identification result was got through comparing instantaneous waveform amplitude and steady power value with the corresponding feature values of the same category. The method, which is simple and reliable, can determine the load category accurately, so it can provide data support for the family load modeling. Examples have verified the effectiveness of the proposed method.
出处
《电测与仪表》
北大核心
2015年第1期119-123,共5页
Electrical Measurement & Instrumentation
关键词
谱聚类
负荷识别
分步识别
相似度
支持率
spectral clustering, load identification, step identification, similarity, support rate
