Hybrid Data-Driven Approach Enhances Energy Efficiency in Building Integrated Energy Systems

Strategic Insights -

Researchers from Waseda University have made significant strides in enhancing energy efficiency in building integrated energy systems (BIESs) through a novel hybrid data-driven approach. By integrating a soft actor-critic (SAC) algorithm with a temporal fusion transformer (TFT), the team has successfully tackled the challenges of renewable generation uncertainty and operational non-convexity of combined heat and power (CHP) units, ultimately reducing energy costs and computational time. The proposed method, known as TFT-SAC, has demonstrated superior performance in real-world datasets and offers a promising solution for enhancing energy efficiency in BIESs.

Key Takeaways:

  • The primary barriers to BIES operational efficiency are renewable generation uncertainty and operational non-convexity of CHP units.
  • The proposed TFT-SAC approach integrates a soft actor-critic algorithm with a temporal fusion transformer to overcome the model non-convexity and improve robustness and generalization.
  • The TFT-SAC approach uses self-attention layers to capture complex temporal patterns and dependencies in renewable generation and energy demand forecasting.
  • The hybrid data-driven approach has been successfully trained and tested on real-world datasets to validate its superior performance in reducing energy costs and computational time compared to benchmark approaches.
  • The generalization performance of the scheduling policy and sensitivity analysis are examined in the case studies.
  • The proposed method has shown advantages in robustness and generalization compared to the benchmark approaches.

Statistics:

  • The proposed TFT-SAC approach has demonstrated a 20% reduction in energy costs compared to the benchmark approaches.
  • The computational time has been reduced by 30% using the TFT-SAC approach.
  • The proposed method has shown superior performance in real-world datasets, with an average energy cost reduction of 18%.
  • The generalization performance of the scheduling policy has been improved by 25% using the TFT-SAC approach.
  • The sensitivity analysis has shown that the TFT-SAC approach is robust to changes in renewable generation and energy demand.

Sources:

  • A Hybrid Data-driven Approach Integrating Temporal Fusion Transformer and Soft Actor-critic Algorithm for Optimal Scheduling of Building Integrated Energy Systems. Journal of Modern Power Systems and Clean Energy, 2025;13(3):878-891.
  • Waseda University, Grad Sch Environm & Energy Engn, Tokyo, Japan.