A Flow-Based Model for Conditional and Probabilistic Electricity Consumption Profile Generation and Prediction

2024-05-03

arXiv.org

AI Summary

To address the need for conditional and unconditional generation and probabilistic forecasting of residential load profiles (RLPs) in low-carbon distribution networks, this paper proposes the Fully Convolutional Probability Flow (FCPFlow) model. Methodologically, FCPFlow introduces invertible linear and normalization layers, enabling end-to-end embedding of continuous conditioning variables—such as temperature and annual electricity consumption—for the first time. This design ensures strong cross-dataset generalization while effectively capturing complex temporal dependencies. Leveraging an invertible neural network architecture, FCPFlow provides exact probability density estimation, supporting uncertainty-aware, high-fidelity load generation and forecasting. Evaluated on multiple real-world residential load datasets, FCPFlow consistently outperforms conventional statistical methods and state-of-the-art deep generative models. Its interpretability, scalability, and probabilistic rigor make it a practical tool for distribution system planning and operation under increasing decarbonization pressures.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Natural Language Processing: Generation

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Residential Load Profile (RLP) generation and prediction are critical for the operation and planning of distribution networks, especially as diverse low-carbon technologies (e.g., photovoltaic and electric vehicles) are increasingly adopted. This paper introduces a novel flow-based generative model, termed Full Convolutional Profile Flow (FCPFlow), which is uniquely designed for both conditional and unconditional RLP generation, and for probabilistic load forecasting. By introducing two new layers--the invertible linear layer and the invertible normalization layer--the proposed FCPFlow architecture shows three main advantages compared to traditional statistical and contemporary deep generative models: 1) it is well-suited for RLP generation under continuous conditions, such as varying weather and annual electricity consumption, 2) it demonstrates superior scalability in different datasets compared to traditional statistical models, and 3) it also demonstrates better modeling capabilities in capturing the complex correlation of RLPs compared with deep generative models.

Problem

• Generates residential electricity consumption profiles under varying conditions
• Predicts probabilistic load forecasts for distribution network planning
• Models complex correlations in load data using deep generative architecture

Innovation

• Flow-based generative model for electricity consumption profiles
• Introduces invertible linear and normalization layers
• Enables conditional generation and probabilistic load forecasting