Physics-Guided Memory Network for Building Energy Modeling

2025-08-05

AI Summary

Building energy consumption forecasting suffers from low accuracy, strong dependence on physical model parameters, and high modeling costs—particularly for newly constructed buildings or scenarios with missing historical data. Method: This paper proposes a memory-augmented neural network framework that synergistically integrates physical priors with deep learning. It innovatively incorporates learnable memory units and an empirical bias correction module, employs parallel projection layers to handle incomplete inputs, and embeds physics-based simulation outputs directly into the network architecture—yielding a hybrid model with mathematically provable properties. Contribution/Results: The method significantly enhances generalization capability and extrapolation stability under data-scarce conditions. Extensive evaluations across diverse real-world scenarios demonstrate high-accuracy hourly energy forecasting. It exhibits robustness and practicality in zero-shot settings (no historical data), dynamic building modifications, and under environmental disturbances.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Learning to Search
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Accurate energy consumption forecasting is essential for efficient resource management and sustainability in the building sector. Deep learning models are highly successful but struggle with limited historical data and become unusable when historical data are unavailable, such as in newly constructed buildings. On the other hand, physics-based models, such as EnergyPlus, simulate energy consumption without relying on historical data but require extensive building parameter specifications and considerable time to model a building. This paper introduces a Physics-Guided Memory Network (PgMN), a neural network that integrates predictions from deep learning and physics-based models to address their limitations. PgMN comprises a Parallel Projection Layers to process incomplete inputs, a Memory Unit to account for persistent biases, and a Memory Experience Module to optimally extend forecasts beyond their input range and produce output. Theoretical evaluation shows that components of PgMN are mathematically valid for performing their respective tasks. The PgMN was evaluated on short-term energy forecasting at an hourly resolution, critical for operational decision-making in smart grid and smart building systems. Experimental validation shows accuracy and applicability of PgMN in diverse scenarios such as newly constructed buildings, missing data, sparse historical data, and dynamic infrastructure changes. This paper provides a promising solution for energy consumption forecasting in dynamic building environments, enhancing model applicability in scenarios where historical data are limited or unavailable or when physics-based models are inadequate.

Problem

• Accurate energy forecasting with limited historical data
• Integrating deep learning and physics-based model predictions
• Enhancing model applicability in dynamic building environments

Innovation

• Integrates deep learning with physics-based models
• Uses Parallel Projection Layers for incomplete inputs
• Memory Unit and Module enhance forecast accuracy