WOMEN SCIENTIST SCHEME ENGINEERING AND TECHNOLOGY
- 1. Women Scientists Scheme-A (WOS-A) Engineering & Technology Reference No. : SR/WOS-A/ET-84/2018 PI’s Name and Institution Address : Dr. V. Agnes Idhaya Selvi, Kalasalingam Academy of Research and Education, Krishnankoil, Tamil Nadu. Mentor’s Name and Host Institute : Dr. D. Devaraj, Kalasalingam Academy of Research and Education, Krishnankoil, Tamil Nadu. A Deep Learning based Cyber Security Assessment system for Smart Grid
- 2. Rationale for the proposed work The Wide Area Monitoring Systems (WAMS) enables real time collection of power system data by installing phasor measurement units (PMUs) at specific locations on the grid. PMUs are deployed at substations, connected to a Local Area Network (LAN). Phasor Data Concentrator (PDC) is used to collect the data from multiple PMUs. As the WAM networks heavily depends on the cyber infrastructure for transferring critical data, the system will be exposed to cyber risks from intruders. The intruders can attack the system in many ways such as breaking the intermediate fire walls of intra networks and hijacking the server. By breaking the fire wall, the false data can be injected, which provide a false information to the entire power network. The false data injection attack causes the violation of operating parameters, resulting in contingency occurring in power system which leads the power system to the insecure condition. 2
- 3. Objectives 3 • To develop a cyber physical test system for analyzing the cyber- attacks in smart grid. • Create the PDC to process streaming time-series data in real- time. Here the PMU data are synchronized with GPS. • Development of deep learning technique to analyze the real- time measurement data from the PMU and detect the data corruption. • To mitigate the False Data Injection attack in PMU data and to propose preventive action against the cyber attack. • Study the impact of cyber-attack on the power system voltage and transient stability.
- 4. Methodology Phase I: Development of Smart Grid test bed Cyber physical test bed for power system will be setup using Real Time Digital Simulator (RTDS) with Hardware In the Loop Simulation. Through hardware interface facility, necessary hardware setup for the proposed work are interconnected with physical system. The main protocol IEEE C37.118 is used in this CPS. Implementation of power system bus network in RTDS, which consist of generator, load, PMU and PDC with hardware relay and breaker circuit. The location of hardware relay can be chosen, in where the possibility of cyber attack may be happen. The false data injection attack can be simulated by python coding. 4
- 6. Contd.., Phase II: Design and Implementation of PDC The RTDS generated PMU Data set (30 to 60 samples per second) are send to PDC with time stamps through GTNET card. The GTNET can accept the PMU data in IEEE C37.118 standard and provide output up to 8 PMUs with symmetrical component information using TCP connection. The PDC is designed by the Grid Protection Alliance (GPA) in which the Measured data aligned with GPS-time and it can be further accessed by the control centre. This PDC data base can now accessed for two purpose, one is for Deep Learning network for detecting the False Data Injection. The other purpose, it can be stored in historian data base for future reference. 6
- 7. Contd.., Phase III: Deep Learning - technique for false data Injection attacks detection in smart Grid In this work, a new Deep Learning techniques is to be adopted to analyze the sequential PMU data in real time and detect the existence of information corruption. The Deep Learning algorithm will be used to model temporal data by treating the previously observed data as additional input and implementing autoregressive (AR) data modeling scheme. The preventive action can be send by the WAMS to the hardware relay which trip the main circuit breaker to prevent the consequence of false data information. The threats detected are categorized according to specific security goals set for the Smart Grid environment and their impact on the overall system security is evaluated. 7
- 8. Work Plan 8 S. N O Activity I Yr. II Yr. III Yr. 1-6 7-12 13 -18 19 - 24 25 - 30 31 –36 A1 Development of Cyber Physical test bed for power system. A2 Generating PMU data by RTDS and fetch the data to PDC. A3 The study of possible Cyber security issues on the developed CPS system A4 Development and implementation of Deep Learning based model for cyber-attack identification A5 Mitigation of false data injection attack by Deep Learning method and preventive action against the cyber attack A6 Study the impact of a real-time cyber-attack on the power system in terms of voltage and transient stability.
- 9. The possible cyber attacks on smart grid environment will be identified. Able to study the cyber security issues in smart grid and possibilities of avoiding cascading failure in smart grid. Possibilities of utilizing available Information Technology infrastructure for establishing communication between the power system customer and control center in secured way. Able to Identify the Efficient Deep Learning algorithm for cyber- attack detection in the PMU data. Will be recognized various cyber security assessment in order to take preventive and corrective action to avoid blackouts. 9 Expected outcome
- 10. Budget Details No. Title 1st Yr. 2nd Yr. 3rd Yr. Total (Rs.) 1 Fellowship 6,60,000 6,60,000 6,60,000 19,80,000 2 Consumables 25,000 20,000 15,000 60000 3 Travel 10,000 5,000 5,000 20,000 4 Contingencies 5,000 5,000 10,000 20,000 5 Institute Overhead 8,000 8000 8,000 24,000 6 Other cost 8,000 8,000 8,000 24,000 7 Equipment 8,15,000 - - 8,15,000 8 Total 15,31,000 7,06,000 7, 06, 000 29,43,000