| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 102 === Non-Intrusive load monitoring systems (NILMs) can determine the operating intervals and power consumptions of electrical loads in a system from measurements made at electrical service entry (ESE) point. In contrast to the conventional monitoring system, NILMs reduce sensor costs by installing sensors only at the ESE and identifies appliances by defining their power signature.
In this thesis, we proposed a new NILM method based on cascaded strategy to enhance the power signatures and is compared with the traditional methods, which use merely one characteristic, such as the real power for power signature. Moreover, we also proposed to use the linearized admittances to reduce the influence of voltage fluctuations on power signatures and the volume of database. For load identification, the inner product was proposed in this thesis. In contrast to the conventional identification tool, inner product can reduce the time for data training and obtain good results. Finally, as experimental results show, the proposed method yields better detection rate than the conventional methods.
|
|---|
