Contingency Analyses with Warm Starter using Probabilistic Graphical Model

Shimiao Li,Amritanshu Pandey,Larry Pileggi

2024-03-20

Abstract:Cyberthreats are an increasingly common risk to the power grid and can thwart secure grid operations. We propose to extend contingency analysis to include cyberthreat evaluations. However, unlike the traditional N-1 or N-2 contingencies, cyberthreats (e.g., MadIoT) require simulating hard-to-solve N-k (with k >> 2) contingencies in a practical amount of time. Purely physics-based power flow solvers, while being accurate, are slow and may not solve N-k contingencies in a timely manner, whereas the emerging data-driven alternatives are fast but not sufficiently generalizable, interpretable, and scalable. To address these challenges, we propose a novel conditional Gaussian Random Field-based data-driven method that performs fast and accurate evaluation of cyberthreats. It achieves speedup of contingency analysis by warm-starting simulations, i.e., improving starting points, for the physical solvers. To improve the physical interpretability and generalizability, the proposed method incorporates domain knowledge by considering the graphical nature of the grid topology. To improve scalability, the method applies physics-informed regularization that reduces model complexity. Experiments validate that simulating MadIoT-induced attacks with our warm starter becomes approximately 5x faster on a realistic 2000-bus system.

Cryptography and Security,Systems and Control

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper primarily addresses the increasingly severe issue of cybersecurity threats in power systems and proposes a new method to improve the existing Contingency Analysis (CA) functionality. Specifically, the paper aims to solve the following key problems:

1. **Limitations of Traditional Contingency Analysis**:
- Current contingency analysis methods typically consider only single component failures (N-1) or two component failures (N-2), while modern cybersecurity threats (such as MadIoT attacks) can cause failures at multiple locations simultaneously (N-k, where k is much greater than 2). In such cases, traditional physical solvers, although accurate, are slow and struggle to complete large-scale N-k failure simulations in a short time.

2. **Shortcomings of Data-Driven Methods**:
- Existing data-driven methods, while fast, have limitations in generalization ability, interpretability, and scalability. For example, these methods may not handle changes in the power grid topology well, and their output results are difficult to correlate with the actual physical system.

3. **Improving the Speed and Accuracy of Contingency Analysis**:
- The paper proposes a data-driven method based on Conditional Gaussian Random Field (CGRF) that accelerates the convergence speed of physical solvers through warm-starting techniques. This method incorporates power grid topology information and applies physical information regularization techniques to enhance the model's generalization ability and interpretability.

4. **Addressing Modern Cybersecurity Threats**:
- In response to modern cybersecurity threats in power grids (such as MadIoT attacks), this method can effectively simulate the impacts of these attacks, helping grid operators identify potential risks in advance and take preventive measures.

### Summary

This paper aims to develop a novel data-driven method for quickly and accurately assessing cybersecurity threats in power systems, especially when dealing with large-scale N-k failures. By combining physical knowledge and machine learning techniques, this method significantly improves the speed of contingency analysis while ensuring the model's generalization ability and interpretability.