Abstract: Existing privacy preserving mechanisms for smart meter are still faced with the problem of insufficient protection for consumer power consumption mode, the temporal perturbation to destroy the raw power waveform is proposed, and a time-disturbed smart meter privacy protection model based on the data availability is derived，which can achieve the compromise between data availability and privacy security. Besides, a non-intrusive load monitoring algorithm as a testing method to evaluate the security performance of the scheme has been carried out. The experimental results show that the temporal perturbation based privacy-preserving method can effectively reduce the recognition accuracy of electrical appliances, achieve better suppression efficiency compared with random perturbation and rechargeable battery, keep the aggregation error of multi-user stabilized around 10%, and perform well in billing measurement.

Key words: smart meter, private preserving, data perturbation, Non-Intrusive Load Monitoring（NILM）

摘要： 针对目前智能电表隐私保护方法存在对用户用电模式保护力度不足的问题，提出采用时延扰动来破坏数据波形，在智能电表数据可用性的基础上推导基于时间扰动的智能电表隐私保护模型，通过扰动智能电表数据发布时间来实现数据安全性与可用性的折中，并利用非侵入式负载监测算法对隐私安全性进行检测。实验结果表明，基于时间扰动的智能电表隐私保护方法能够有效地抑制电器切换事件的识别准确率，相比于随机扰动和充电电池方法有更好的抑制效率，多用户的聚合误差稳定在10%左右，同时在计费误差上有着优异的表现。

关键词: 智能电表, 隐私保护, 数据扰动, 非侵入式负载监测（NILM）