Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
167 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Data-driven Predictive Control for Unlocking Building Energy Flexibility: A Review (2007.14866v1)

Published 29 Jul 2020 in eess.SY and cs.SY

Abstract: Managing supply and demand in the electricity grid is becoming more challenging due to the increasing penetration of variable renewable energy sources. As significant end-use consumers, and through better grid integration, buildings are expected to play an expanding role in the future smart grid. Predictive control allows buildings to better harness available energy flexibility from the building passive thermal mass. However, due to the heterogeneous nature of the building stock, developing computationally tractable control-oriented models, which adequately represent the complex and nonlinear thermal-dynamics of individual buildings, is proving to be a major hurdle. Data-driven predictive control, coupled with the "Internet of Things", holds the promise for a scalable and transferrable approach,with data-driven models replacing traditional physics-based models. This review examines recent work utilising data-driven predictive control for demand side management application with a special focus on the nexus of model development and control integration, which to date, previous reviews have not addressed. Further topics examined include the practical requirements for harnessing passive thermal mass and the issue of feature selection. Current research gaps are outlined and future research pathways are suggested to identify the most promising data-driven predictive control techniques for grid integration of buildings.

Citations (181)

Summary

  • The paper presents a comprehensive review of data-driven predictive control strategies that leverage machine learning over traditional physics-based models.
  • It categorizes modeling techniques into black-box, reduced order, and hybrid models, with state-space representations noted for their accuracy and efficiency.
  • The study highlights challenges such as benchmarking, scalability, and multi-energy integration as key areas for future research.

Overview of "Data-driven Predictive Control for Unlocking Building Energy Flexibility: A Review"

The paper "Data-driven Predictive Control for Unlocking Building Energy Flexibility: A Review" offers a comprehensive survey of various data-driven predictive control (DDPC) strategies applied in building energy management systems, specifically focusing on their role in demand-side management (DSM). The paper highlights the increasing importance of predictive control in buildings due to the growing penetration of renewable energy sources and the consequent challenges in electricity grid stability. Buildings, as major energy consumers, have the potential to contribute significantly to the smart grid's flexibility. This review synthesizes literature from numerous studies, underscoring the modeling and algorithmic advancements made in DDPC for harnessing building energy flexibility.

Key Findings

  • Data-driven Predictive Control Growth: The review summarizes findings from 115 studies demonstrating a substantial growth in research focused on data-driven predictive control techniques in recent years. These methods eschew traditional physics-based models, instead leveraging real-time data and machine learning to predict and optimize building energy management.
  • Modeling Techniques: The paper classifies modeling approaches into three categories: black-box models (e.g., neural networks), reduced order models (e.g., state-space models), and hybrid models. Reduced order models, particularly state-space representations, dominate the field due to their balance between accuracy and computational efficiency.
  • Variations in Control Systems: Model Predictive Control (MPC) is the prevalent strategy, with a growing interest in reinforcement learning (RL) for its adaptive capabilities. Recent innovations include approximate MPC using machine learning to strip implementation complexity and achieve real-time control efficiencies.
  • Applications Beyond Thermal Mass: While most studies focus on utilizing building passive thermal mass for energy flexibility, there is increasing interest in integrating active thermal storage, electric storage systems, EVs, and on-site generation to create more comprehensive energy management strategies.
  • Impact of Forecasting: The review notes widespread use of weather forecasting in predictive modeling, essential for optimizing energy use in buildings. There is a call to improve forecasting techniques for variables like occupancy and dynamic pricing further to enhance the efficiency of DDPC systems.
  • Empirical Validation: Many approaches remain validated through simulations rather than real-world implementations, leading to questions about scalability and transferability across different building types and operational conditions.

Implications

The implications of this research are profound in the pursuit of more energy-efficient and grid-responsive buildings. With DDPC, buildings can optimize their energy usage dynamically, ultimately reducing costs, improving occupant comfort, and contributing to grid stability. The scalability of these models across different building types and their integration into wider energy systems remain areas for future investigation.

Future Research Directions

To fully realize the potential of DDPC in smart buildings, several challenges must be addressed:

  • Feature Selection and Data Quality: Developing robust methodologies for selecting relevant data features is crucial, particularly given the variability in sensor data quality and richness.
  • Benchmarking and Standardization: As highlighted, the lack of standard evaluation benchmarks for comparing DDPC approaches limits the ability to gauge performance across varied operational environments.
  • Scalability and Transferability: There is a need for methodologies that ensure DDPC systems can be applied consistently across different buildings, fostering widespread adoption in the industry.
  • Integration of Multi-energy Vectors: Future frameworks must consider the simultaneous management of multiple energy resources (thermal, electric, automotive) within buildings to optimize overall system performance.

Overall, the paper underscores the significance of data-driven innovations in transforming building energy management, promoting a trajectory toward a more integrated and responsive energy ecosystem.