KCLNet: Physics-Informed Power Flow Prediction via Constraints Projections

2025-06-15

AI Summary

In smart grid applications, AI-based power flow forecasting often violates Kirchhoff’s Current Law (KCL), leading to physically inconsistent predictions. To address this, this paper proposes a physics-informed graph neural network (GNN) framework that enforces KCL as a hard constraint within the model architecture. Specifically, we introduce a differentiable hyperplane projection layer that guarantees exact KCL satisfaction—achieving zero KCL violation. The method jointly encodes power system topology and performs end-to-end optimization, balancing physical fidelity with computational efficiency. Extensive evaluation across multiple operational scenarios demonstrates that the approach attains numerical-solver-level accuracy (mean absolute error < 0.5%), maintains strict KCL compliance, and accelerates inference by two orders of magnitude (~100×) compared to conventional solvers. These advances significantly enhance real-time grid security assessment and decision-making capabilities.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

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
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In the modern context of power systems, rapid, scalable, and physically plausible power flow predictions are essential for ensuring the grid's safe and efficient operation. While traditional numerical methods have proven robust, they require extensive computation to maintain physical fidelity under dynamic or contingency conditions. In contrast, recent advancements in artificial intelligence (AI) have significantly improved computational speed; however, they often fail to enforce fundamental physical laws during real-world contingencies, resulting in physically implausible predictions. In this work, we introduce KCLNet, a physics-informed graph neural network that incorporates Kirchhoff's Current Law as a hard constraint via hyperplane projections. KCLNet attains competitive prediction accuracy while ensuring zero KCL violations, thereby delivering reliable and physically consistent power flow predictions critical to secure the operation of modern smart grids.

Problem

• Ensures physically plausible power flow predictions
• Reduces computation time for dynamic grid conditions
• Enforces Kirchhoff's Current Law in AI models

Innovation

• Physics-informed graph neural network
• Kirchhoff's Current Law hard constraint
• Hyperplane projections for zero violations