LoaDiff Generates Synthetic Electricity Consumption Time Series
September 11, 2026
LoaDiff uses a diffusion-based generative model to produce year-long, sub-hourly smart-meter load curves. The model supports conditioning on static household attributes and dynamic variables like temperature, providing a privacy-preserving alternative to restricted electricity datasets.
HOW THIS AFFECTS YOU
●
builderThis provides a way to generate high-fidelity synthetic data for energy analytics applications.
●
researcherYou can leverage this for training load forecasting models without accessing sensitive smart-meter data.