Intrusion detection using neural networks for smart grid networks

Smart Grids have the potential to create a revolution in the energy industry. Apart from financial and social impacts, Smart Grids are a necessity for sustainable development and reduction of dependence on non renewable energy resources. However, the operation of smart grid is going to entirely diff...

Full description

Saved in:
Bibliographic Details
Main Author: Kumar Ankit
Other Authors: Ma Maode
Format: Theses and Dissertations
Language:English
Published: 2019
Subjects:
Online Access:http://hdl.handle.net/10356/78653
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Smart Grids have the potential to create a revolution in the energy industry. Apart from financial and social impacts, Smart Grids are a necessity for sustainable development and reduction of dependence on non renewable energy resources. However, the operation of smart grid is going to entirely different from the traditional grids. With the requirement of bidirectional communication links and increased reliance on information and communication technology, the smart grids are certainly vulnerable to security threats. Moreover,as it has been demonstrated by attacks like Stuxnet that any security breach in cyber-physical systems like the smart grids catering to the critical sectors like energy can have massive social, economic and technological impacts and can take the organization decades to recover. The smart grid networks characteristics such as heterogeneity, delay constraints, bandwidth, scalability, and others make it challenging deploying uniform security approaches all over the networks segments. One approach to provide a second line of defence for the smart grid networks. In this work various cyber security requirements are analysed and security threats are reviewed. Based on the guidelines a scalable online intrusion detection system is designed to act as the second line of defence for the smart grid. The design is then implemented on python using tensorflow. The design is then trained and tested with the NSL KDD dataset and is then compared with other relevant implementations.