Adaptive Informed Deep Neural Networks for Power Flow Analysis

2024-12-03

arXiv.org

AI Summary

To address the challenge of efficiently modeling the nonlinear dynamics of large-scale modern power systems, this paper proposes PINN4PF, a physics-informed neural network for end-to-end power flow analysis. Methodologically, it employs a dual-head feedforward network with adaptive activation functions to separately model the nonlinear responses of active and reactive power injections; additionally, it introduces a topology-aware hybrid loss function that explicitly embeds grid structural constraints and physical governing equations as prior knowledge. Experimental evaluation across benchmark systems ranging from 4 to 2,224 buses demonstrates that PINN4PF significantly outperforms linear regression and black-box multilayer perceptrons—achieving up to two orders of magnitude improvement in generalization capability, noise robustness, and accuracy of derived quantities (e.g., branch currents and powers). To the best of our knowledge, this is the first approach to synergistically enhance both physical consistency and data-driven learning under high-dimensional topological coupling.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

This study introduces PINN4PF, an end-to-end deep learning architecture for power flow (PF) analysis that effectively captures the nonlinear dynamics of large-scale modern power systems. The proposed neural network (NN) architecture consists of two important advancements in the training pipeline: (A) a double-head feed-forward NN that aligns with PF analysis, including an activation function that adjusts to the net active and reactive power injections patterns, and (B) a physics-based loss function that partially incorporates power system topology information. The effectiveness of the proposed architecture is illustrated through 4-bus, 15-bus, 290-bus, and 2224-bus test systems and is evaluated against two baselines: a linear regression model (LR) and a black-box NN (MLP). The comparison is based on (i) generalization ability, (ii) robustness, (iii) impact of training dataset size on generalization ability, (iv) accuracy in approximating derived PF quantities (specifically line current, line active power, and line reactive power), and (v) scalability. Results demonstrate that PINN4PF outperforms both baselines across all test systems by up to two orders of magnitude not only in terms of direct criteria, e.g., generalization ability, but also in terms of approximating derived physical quantities.

Problem

• Develops deep learning for power flow nonlinear dynamics
• Proposes physics-based NN with adaptive activation functions
• Enhances accuracy and scalability in power system analysis

Innovation

• Double-head feed-forward NN for PF analysis
• Adaptive activation function for power injections
• Physics-based loss with topology information