Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Nonparametric Kernel Regression for Coordinated Energy Storage Peak Shaving with Stacked Services

Emily Logan, Ning Qi, Bolun Xu

Abstract

Developing effective control strategies for behind-the-meter energy storage to coordinate peak shaving and stacked services is essential for reducing electricity costs and extending battery lifetime in commercial buildings. This work proposes an end-to-end, two-stage framework for coordinating peak shaving and energy arbitrage with a theoretical decomposition guarantee. In the first stage, a non-parametric kernel regression model constructs state-of-charge trajectory bounds from historical data that satisfy peak-shaving requirements. The second stage utilizes the remaining capacity for energy arbitrage via a transfer learning method. Case studies using New York City commercial building demand data show that our method achieves a 1.3 times improvement in performance over the state-of-the-art forecast-based method, achieving cost savings and effective peak management without relying on predictions.

Nonparametric Kernel Regression for Coordinated Energy Storage Peak Shaving with Stacked Services

Abstract

Developing effective control strategies for behind-the-meter energy storage to coordinate peak shaving and stacked services is essential for reducing electricity costs and extending battery lifetime in commercial buildings. This work proposes an end-to-end, two-stage framework for coordinating peak shaving and energy arbitrage with a theoretical decomposition guarantee. In the first stage, a non-parametric kernel regression model constructs state-of-charge trajectory bounds from historical data that satisfy peak-shaving requirements. The second stage utilizes the remaining capacity for energy arbitrage via a transfer learning method. Case studies using New York City commercial building demand data show that our method achieves a 1.3 times improvement in performance over the state-of-the-art forecast-based method, achieving cost savings and effective peak management without relying on predictions.
Paper Structure (10 sections, 1 theorem, 15 equations, 4 figures, 1 algorithm)

This paper contains 10 sections, 1 theorem, 15 equations, 4 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Problem Formulation and Decomposition
  3. Peak Shaving and Arbitrage Formulation
  4. Decomposition of Stacked Services
  5. Learning and Control Method
  6. Kernel Regression for SoC Reserve Prediction
  7. Real-Time Control
  8. Parameter Search
  9. Simulation and Results
  10. Conclusion

Key Result

Proposition 1

Given that $\delta$ is sufficiently small and $\kappa \gg \lambda_t, c$, solving the two-stage optimization in PS_Only_Formulation and Arb_Formulation_Stage2 is equivalent to solving the combined peak shaving and arbitrage problem in PS_and_Arb_Formulation.

Figures (4)

  • Figure 1: Building electricity demand from 2019–2024. Line colors indicate years. Yearly standard deviations are: 2019: 0.21 MW, 2020: 0.16 MW, 2021: 0.16 MW, 2022: 0.16 MW, 2023: 0.18 MW, and 2024: 0.19 MW.
  • Figure 2: Comparison of the kernel-regression predicted and deterministic minimum SoC reserve trajectories for a 1 MW battery.
  • Figure 3: Peak demand (MW) by month for perfect foresight, kernel regression-based, and demand forecasting control algorithms for a 0.6 MW battery.
  • Figure 4: Total annual cost savings (%) vs. battery size for perfect foresight, kernel regression-based, and demand forecasting control algorithms.

Theorems & Definitions (2)

  • Proposition 1
  • proof