基於位置感知非侵入式負載監測之建築能源管理系統

Abstract

非侵入式負載監控系統藉由分析迴路之總電壓、總電流、總功耗、電流波型來了解耗電情形，利用每一電器的電器特性來判別電器狀態以及其耗電。建築能源管理系統將此演算法應用在建築物中，然而我們發現此方案存在幾項問題。首先，電力傳輸線存在微量電阻使得電器的特徵值改變，導致電器辨識率下降。另外，建築物中往往有在不同位置的同種電器如冷器、電燈，使得傳統的演算法很難分辨出來。因此，在本研究中提出了一套完整的系統來解決上述問題，使用者可以透過行動裝置的輔助，在最少的人為操作完成訓練電器以及建立電力線拓撲資訊，接著利用此電力線拓撲資訊來改善原有的電器辨識與能源拆解演算法，最後使用者可以透過此能源管理系統一目瞭然地得知用電資訊。

Nonintrusive Load Monitoring (NILM) system is an innovative approach for monitoring electrical loads by analyzing the aggregated signatures such as voltage, current, active power, and reactive power from the main power supply. One of the products based on NILM is Building Energy Management System (BEMS), which helps user to understand the energy consumption of their building. However, there are two major problems if NILM is applied to BEMS directly. First, the extra power consumption of electric power transmission lines of building cannot be ignored and it influences the accuracy of the energy disaggregation. Second, NILM cannot distinguish two appliance instances if they have the same signatures. In this paper, we propose a location-based NILM technology to solve the problems mentioned above. Together with a smart training tool and a visualized app, we demonstrate that location-based NILM can considerably enhance the functionalities of BEMS. User can establish the topology model with smart training tool in minimal manual operations, and they can easily understand the improved disaggregation result with the visualized app.