智慧電錶(Smart Meter)在智慧電網(Smart Grid)中是一個不可缺少的部份。智慧電錶會收集日常使用的功率數據,如此一來電力公司能夠取得客戶的消費模式來設定電價時段或預測負載需求。透過客戶的用電量數據統計,電力公司期盼提高電力輸送品質和效率。但是這將會透露出客戶用電隱私的問題。在這項研究中,我們提出一個新穎的用戶日常行為干擾演算法(User Daily Behavior Disturbance, UDBD)來提供智慧電錶隱私保護的機制。利用客戶以前的用電數據,UDBD讓目前的用電數據產生些微的變化。實驗結果證實在識別電器的機率(Identifiable Appliance Rate, IAR)、正規化錯誤(Normalised Error in Assigned Power, NRE)和均方根誤差(Root Mean Square Error, RMSE)的項目中UDBD都優於高斯隨機雜訊(Gaussian random noise)的方法。UDBD是一個實際的智慧電錶隱私保護工具,它能夠確保隱私保護和降低數據失真的目標。
The smart meter is an essential component in a smart grid environment. After collecting daily use power usage data from smart meters, the power company is able to find normal customer consumption patterns in response to set electricity price over time or predict load demands. By applying the knowledge of customer power usage statistics, the power companies look forward to improving the quality and efficiency of the power delivery. However, using electrical usage data bears privacy issues. In this study, we present a novel user daily behavior disturbance algorithm (UDBD) to provide a privacy preserving mechanism for smart meters. Utilizing the user previous power usage data, UDBD makes slight changes for current power usage as needed. Experimental results show that UDBD outperforms the Gaussian random noise approach in terms of identifiable appliance rate (IAR), normalised error in assigned power (NRE) and root mean square error (RMSE). UDBD is a practical smart meter privacy preserving tool which is able to reach the goal of ensuring privacy protection and reducing the total data distortion.