Load Hiding to Preserve Privacy from Smart Meter Measurements Ryan Fraser Advisor: Dr. Kevin Tomsovic Ailin Asadinejad Final Project Presentation July 14, 2016 Knoxville, Tennessee
Smart Meters United States Smart Meter Use, 2014 Provide More information More accurate forecasting Fault detection Increased information to consumers Increased stability for renewables and energy storage 2
Non-Invasive Load Monitoring (NILM) Capable of Identifying Specific Appliances Loss of privacy Map activities and behaviors of homeowners 3
Previous Research Battery-based Load Monitoring Home Power Use Battery 1600 500 1400 1200 Power Demand (W) Power Demand (W) 0 1000 800 600 -500 400 200 0 -1000 0 5 10 15 20 0 5 10 15 20 hrs hrs Time Time Power Read by Meter 581 580.8 580.6 580.4 Power Demand (W) 580.2 580 579.8 579.6 579.4 579.2 579 0 5 10 15 20 4 hrs Time
Project Goal Design a system to improve individual household privacy Maintain the benefits provided by smart meters 5
Information Theory & Entropy Entropy is the measure of unpredictability of information � ) � � = − ∑ � � � log � �(� � ��� � � � = � � ⁄ � ��� : 6
Load Imitating Model of Television Power Demand Model of Lighting Power Demand Add false loads in order to increase system entropy 7
Data 8
Analyzed Loads 9
Formula � � � = 1 2� � � � + � − �(�) ���� when � � > 0 battery discharges battery charges � � < 0 Probability of Load Occurring = � � + �(�) 10
Results Imitation Original 1000 800 600 400 200 0 -200 -400 8 8.5 9 9.5 10 10.5 11 11.5 hrs
Results 12
Possible Further Work Design to work with non-time based loads Optimize with a battery Test with more realistic data Test against NILM 13
Acknowledgements This work was supported primarily by the ERC Program of the National Science Foundation and DOE under NSF Award Number EEC-1041877. Other US government and industrial sponsors of CURENT research are also gratefully acknowledged. 14
Questions and Answers 15
Recommend
More recommend
