Game-Theoretic and Machine-Learning Techniques for Cyber-Physical Security and Resilience in Smart Grid

Authors

Longfei WeiFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Electrical Engineering

First Advisor's Name

Arif Sarwat

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Malek Adjouadi

Second Advisor's Committee Title

Committee Member

Third Advisor's Name

Armando Barreto

Third Advisor's Committee Title

Committee Member

Fourth Advisor's Name

Jean Andrian

Fourth Advisor's Committee Title

Committee Member

Fifth Advisor's Name

Walid Saad

Fifth Advisor's Committee Title

Committee Member

Keywords

Cyber-Physical Security, Game Theory, Machine Learning, Reliability, Resilience, Smart Grid

Date of Defense

10-29-2018

Abstract

The smart grid is the next-generation electrical infrastructure utilizing Information and Communication Technologies (ICTs), whose architecture is evolving from a utility-centric structure to a distributed Cyber-Physical System (CPS) integrated with a large-scale of renewable energy resources. However, meeting reliability objectives in the smart grid becomes increasingly challenging owing to the high penetration of renewable resources and changing weather conditions. Moreover, the cyber-physical attack targeted at the smart grid has become a major threat because millions of electronic devices interconnected via communication networks expose unprecedented vulnerabilities, thereby increasing the potential attack surface. This dissertation is aimed at developing novel game-theoretic and machine-learning techniques for addressing the reliability and security issues residing at multiple layers of the smart grid, including power distribution system reliability forecasting, risk assessment of cyber-physical attacks targeted at the grid, and cyber attack detection in the Advanced Metering Infrastructure (AMI) and renewable resources.

This dissertation first comprehensively investigates the combined effect of various weather parameters on the reliability performance of the smart grid, and proposes a multilayer perceptron (MLP)-based framework to forecast the daily number of power interruptions in the distribution system using time series of common weather data. Regarding evaluating the risk of cyber-physical attacks faced by the smart grid, a stochastic budget allocation game is proposed to analyze the strategic interactions between a malicious attacker and the grid defender. A reinforcement learning algorithm is developed to enable the two players to reach a game equilibrium, where the optimal budget allocation strategies of the two players, in terms of attacking/protecting the critical elements of the grid, can be obtained. In addition, the risk of the cyber-physical attack can be derived based on the successful attack probability to various grid elements.

Furthermore, this dissertation develops a multimodal data-driven framework for the cyber attack detection in the power distribution system integrated with renewable resources. This approach introduces the spare feature learning into an ensemble classifier for improving the detection efficiency, and implements the spatiotemporal correlation analysis for differentiating the attacked renewable energy measurements from fault scenarios. Numerical results based on the IEEE 34-bus system show that the proposed framework achieves the most accurate detection of cyber attacks reported in the literature. To address the electricity theft in the AMI, a Distributed Intelligent Framework for Electricity Theft Detection (DIFETD) is proposed, which is equipped with Benford’s analysis for initial diagnostics on large smart meter data. A Stackelberg game between utility and multiple electricity thieves is then formulated to model the electricity theft actions. Finally, a Likelihood Ratio Test (LRT) is utilized to detect potentially fraudulent meters.

Identifier

FIDC007052

ORCID

https://orcid.org/0000-0002-4927-9706

Comments

In the submitted dissertation, all Chapters are unpublished except Chapters 5 and 7. Chapter 5 is published in 2016 IEEE industry applications society annual meeting, and Chapter 7 is published in 2017 Resilience Week.

Creative Commons License

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Recommended Citation

Wei, Longfei, "Game-Theoretic and Machine-Learning Techniques for Cyber-Physical Security and Resilience in Smart Grid" (2018). FIU Electronic Theses and Dissertations. 3850.
https://digitalcommons.fiu.edu/etd/3850