# Towards Practical Physics-Informed ML Design and Evaluation for Power Grid

arXiv: https://arxiv.org/abs/2205.03673 · HTML: https://mlsanity.com/paper/2205.03673/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 39 references that are in the map.
- First posted on arXiv 7 May 2022.

## Builds on (most cited first, 17 of 17)

- [Hamiltonian Neural Networks](https://mlsanity.com/paper/1906.01563/) (2019)
- [Deep Lagrangian Networks: Using Physics as Model Prior for Deep Learning](https://mlsanity.com/paper/1907.04490/) (2019)
- [DC3: A learning method for optimization with hard constraints](https://mlsanity.com/paper/2104.12225/) (2021)
- [Generalization in Deep Learning](https://mlsanity.com/paper/1710.05468/) (2017)
- [Optimal Power Flow Using Graph Neural Networks](https://mlsanity.com/paper/1910.09658/) (2019)
- [Physics-Guided Deep Neural Networks for Power Flow Analysis](https://mlsanity.com/paper/2002.00097/) (2020)
- [Real-time Power System State Estimation and Forecasting via Deep Neural Networks](https://mlsanity.com/paper/1811.06146/) (2018)
- [DeepOPF: Deep Neural Network for DC Optimal Power Flow](https://mlsanity.com/paper/1905.04479/) (2019)
- [Fast Calculation of Probabilistic Power Flow: A Model-based Deep Learning Approach](https://mlsanity.com/paper/1906.06017/) (2019)
- [AI-Based Autonomous Line Flow Control via Topology Adjustment for Maximizing Time-Series ATCs](https://mlsanity.com/paper/1911.04263/) (2019)
- [State Estimation in Electric Power Systems Leveraging Graph Neural Networks](https://mlsanity.com/paper/2201.04056/) (2022)
- [Dynamic Graph-Based Anomaly Detection in the Electrical Grid](https://mlsanity.com/paper/2012.15006/) (2020)
- [Real-Time Contingency Analysis with Corrective Transmission Switching - Part I: Methodology](https://mlsanity.com/paper/1604.05570/) (2016)
- [Learning-Based Real-Time Event Identification Using Rich Real PMU Data](https://mlsanity.com/paper/2006.10121/) (2020)
- [Robust Power Flow and Three-Phase Power Flow Analyses](https://mlsanity.com/paper/1803.01211/) (2018)
- [Learning Latent Interactions for Event classification via Graph Neural Networks and PMU Data](https://mlsanity.com/paper/2010.01616/) (2020)
- [Circuit-theoretic Line Outage Distribution Factor](https://mlsanity.com/paper/2204.07684/) (2022)

## Built on by (most cited first, 1 of 1)

- [Power Grid Behavioral Patterns and Risks of Generalization in Applied Machine Learning](https://mlsanity.com/paper/2304.10702/) (2023)

## Abstract (arXiv)

When applied to a real-world safety critical system like the power grid, general machine learning methods suffer from expensive training, non-physical solutions, and limited interpretability. To address these challenges for power grids, many recent works have explored the inclusion of grid physics (i.e., domain expertise) into their method design, primarily through including system constraints and technical limits, reducing search space and defining meaningful features in latent space. Yet, there is no general methodology to evaluate the practicality of these approaches in power grid tasks, and limitations exist regarding scalability, generalization, interpretability, etc. This work formalizes a new concept of physical interpretability which assesses how a ML model makes predictions in a physically meaningful way and introduces an evaluation methodology that identifies a set of attributes that a practical method should satisfy. Inspired by the evaluation attributes, the paper further develops a novel contingency analysis warm starter for MadIoT cyberattack, based on a conditional Gaussian random field. This method serves as an instance of an ML model that can incorporate diverse domain knowledge and improve on these identified attributes. Experiments validate that the warm starter significantly boosts the efficiency of contingency analysis for MadIoT attack even with shallow NN architectures.