State Estimation under Malicious Data Attacks in Distribution System
The electric distribution network is gradually making transition from a passive to an active and smart network, where it can improve the performance and flexibility of operation. It can improve and maintain the quality of service, reduce costs and increase the capacity of grid to host Distributed Generation. This smart distribution network will make efficient use of sensors and Automated Metering Infrastructure (AMI) in terms of measurement and communication architecture. To enable these smart functionalities of the network the states of the unbalanced distribution system need to be observed properly. The distribution management systems (DMS) will play an important role in control and operation of the smart distribution systems. Hence, the development of the Distribution System State Estimation (DSSE) has become an essential part of the distribution network operation. The state estimator provides a real-time estimate of system states from the measurements obtained from meters and sensors in the remote terminal unit (RTUs). But this smartness of the system carries with it an inherent difficulty. With the proliferation of remote management and control of this smart system, security plays an important role, because the vastness of the system and the convenience of remote management can be exploited by adversaries or hackers for nefarious purposes. An adversary can influence the estimates by tampering with some of the meter data, the price information and control commands. The attacker can inject false data by targeting either some chosen or random state variables. Since most of the bad data detection algorithms depend on the residues of measurements, the attacker can take advantage of the fact that some of the measurements can pass the bad data detection test even if it is tampered. Usually, the measurement errors are assumed to be independently and identically distributed and obey normal distribution. So, the weighted least squares (WLS) estimator is used to find the best estimates of the system states. This project aims to identify the gross measurement data as a part of state estimation process and thus helps to keep the system secured and reliable under threats from adversaries.
- Ankur Majumdar
- Bikash Pal