Detection of Anomalies in Smart Meter Data: A Density-Based Approach

Detection of Anomalies in Smart Meter Data: A Density-Based Approach

Smart grid is the next generation of power grid that provides two-way communication, both in sending and receiving information and in power transfer, among its programs, and using advanced technologies and features such as flexibility, ensuring reliability, affordability, reducing carbon footprints, reinforcing global competiveness and etc. Along with such advantages that give the system administrators and electricity customers the convenience and speed to do business, the security of such a system is far more intrusive. One of the important aspects of maintaining security is on the consumption side, because maintaining the privacy of customers is important and neglecting that will cause an irreparable financial and social losses. Hence, in this paper, we tried to use the OPTICS density-based technique to diagnose abnormalities in information and intelligent data of customers instantly and compare the results of different scenarios. To improve the efficiency of the methodology, we use the index called LOF. Which is actually a factor in detecting the unusual nature of the data in the density-based methods, and will do this based on the score given to it. In other words, it is not binary but gives a score based on which the disturbance of the data can be measured. In order to carry out these simulations, we used London s intelligent metering data in January 2013, which was sent to the control center every 30 minutes.

Smart grid is the next generation of power grid that provides two-way communication, both in sending and receiving information and in power transfer, among its programs, and using advanced technologies and features such as flexibility, ensuring reliability, affordability, reducing carbon footprints, reinforcing global competiveness and etc. Along with such advantages that give the system administrators and electricity customers the convenience and speed to do business, the security of such a system is far more intrusive. One of the important aspects of maintaining security is on the consumption side, because maintaining the privacy of customers is important and neglecting that will cause an irreparable financial and social losses. Hence, in this paper, we tried to use the OPTICS density-based technique to diagnose abnormalities in information and intelligent data of customers instantly and compare the results of different scenarios. To improve the efficiency of the methodology, we use the index called LOF. Which is actually a factor in detecting the unusual nature of the data in the density-based methods, and will do this based on the score given to it. In other words, it is not binary but gives a score based on which the disturbance of the data can be measured. In order to carry out these simulations, we used London s intelligent metering data in January 2013, which was sent to the control center every 30 minutes.