Synthetic Power Flow Data Generation Using Physics-Informed Denoising Diffusion Probabilistic Models

TL;DR

Real-world power-flow data are often limited by privacy and operational constraints. The authors develop a physics-informed denoising diffusion probabilistic model (DDPM) framework to synthesize feasible power-flow data, incorporating auxiliary training to learn a linear diffusion schedule and a physics-informed loss to enforce power balance and bounds. Empirical evaluation on IEEE 14-bus and 30-bus systems shows significant gains in feasibility, diversity, and statistical fidelity over physics-informed GANs and baseline DDPMs, with residual imbalances around 0.013–0.017 p.u. The approach enables privacy-preserving, data-driven power-system research and scalable synthetic data generation for training and validation of grid analytics, while highlighting directions for speed, distributed training, scenario conditioning, and integration with optimization modules.

Abstract

Many data-driven modules in smart grid rely on access to high-quality power flow data; however, real-world data are often limited due to privacy and operational constraints. This paper presents a physics-informed generative framework based on Denoising Diffusion Probabilistic Models (DDPMs) for synthesizing feasible power flow data. By incorporating auxiliary training and physics-informed loss functions, the proposed method ensures that the generated data exhibit both statistical fidelity and adherence to power system feasibility. We evaluate the approach on the IEEE 14-bus and 30-bus benchmark systems, demonstrating its ability to capture key distributional properties and generalize to out-of-distribution scenarios. Comparative results show that the proposed model outperforms three baseline models in terms of feasibility, diversity, and accuracy of statistical features. This work highlights the potential of integrating generative modelling into data-driven power system applications.

Figures (5)

• Figure 1: Average Power Imbalance in Forward Process on IEEE 14-bus System
• Figure 2: The Framework for Auxiliary Training
• Figure 3: Performance of Proposed Method on IEEE 14-bus System
• Figure 4: The Diversity of Synthetic Data from Proposed DDPMs
• Figure 5: The Diversity of Synthetic Data from GAN