Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Risk-aware Scheduling and Dispatch of Flexibility Events in Buildings

Paul Scharnhorst, Baptiste Schubnel, Rafael E. Carrillo, Pierre-Jean Alet, Colin N. Jones

TL;DR

The paper addresses coordinating multiple flexible building assets to provide grid services through load shifting. It introduces uncertainty-aware flexibility envelopes derived from data-driven virtual battery models to quantify feasible power shifts over a horizon. A MILP scheduling framework determines which assets activate and when, while a dispatch heuristic allocates the committed demand to individual assets. In simulations with Energym, the approach demonstrates scalable performance up to 2000 assets and improved peak reduction with tunable risk, while maintaining comfort and reporting meaningful trade-offs.

Abstract

Residential and commercial buildings, equipped with systems such as heat pumps (HPs), hot water tanks, or stationary energy storage, have a large potential to offer their consumption flexibility as grid services. In this work, we leverage this flexibility to react to consumption requests related to maximizing self-consumption and reducing peak loads. We employ a data-driven virtual storage modeling approach for flexibility prediction in the form of flexibility envelopes for individual buildings. The risk-awareness of this prediction is inherited by the proposed scheduling algorithm. A Mixed-integer Linear Program (MILP) is formulated to schedule the activation of a pool of buildings in order to best respond to an external aggregated consumption request. This aggregated request is then dispatched to the active individual buildings, based on the previously determined schedule. The effectiveness of the approach is demonstrated by coordinating up to 500 simulated buildings using the Energym Python library and observing about 1.5 times peak power reduction in comparison with a baseline approach while maintaining comfort more robustly. We demonstrate the scalability of the approach by solving problems with 2000 buildings in about 21 seconds, with solving times being approximately linear in the number of considered assets.

Risk-aware Scheduling and Dispatch of Flexibility Events in Buildings

TL;DR

The paper addresses coordinating multiple flexible building assets to provide grid services through load shifting. It introduces uncertainty-aware flexibility envelopes derived from data-driven virtual battery models to quantify feasible power shifts over a horizon. A MILP scheduling framework determines which assets activate and when, while a dispatch heuristic allocates the committed demand to individual assets. In simulations with Energym, the approach demonstrates scalable performance up to 2000 assets and improved peak reduction with tunable risk, while maintaining comfort and reporting meaningful trade-offs.

Abstract

Residential and commercial buildings, equipped with systems such as heat pumps (HPs), hot water tanks, or stationary energy storage, have a large potential to offer their consumption flexibility as grid services. In this work, we leverage this flexibility to react to consumption requests related to maximizing self-consumption and reducing peak loads. We employ a data-driven virtual storage modeling approach for flexibility prediction in the form of flexibility envelopes for individual buildings. The risk-awareness of this prediction is inherited by the proposed scheduling algorithm. A Mixed-integer Linear Program (MILP) is formulated to schedule the activation of a pool of buildings in order to best respond to an external aggregated consumption request. This aggregated request is then dispatched to the active individual buildings, based on the previously determined schedule. The effectiveness of the approach is demonstrated by coordinating up to 500 simulated buildings using the Energym Python library and observing about 1.5 times peak power reduction in comparison with a baseline approach while maintaining comfort more robustly. We demonstrate the scalability of the approach by solving problems with 2000 buildings in about 21 seconds, with solving times being approximately linear in the number of considered assets.
Paper Structure (18 sections, 2 theorems, 22 equations, 4 figures)

This paper contains 18 sections, 2 theorems, 22 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Flexibility estimation
  3. Characterization of flexibility
  4. Virtual battery modeling of thermal assets
  5. Scheduling and dispatch
  6. General scheduling problem
  7. Scheduling with request commitment
  8. Scheduling for self-consumption and peak reduction
  9. Dispatch algorithm
  10. Simulation setup
  11. Building models and requests
  12. Baseline approach
  13. Evaluation metrics
  14. Results
  15. Self-consumption experiments
  16. ...and 3 more sections

Key Result

Lemma 1

Let the sample sets of $a^+$ and $a^-$ be denoted by $\mathcal{P}^+$ and $\mathcal{P}^-$, $|\mathcal{P}^+ \times \mathcal{P}^-| = N$, and an uncertainty parameter $\alpha$ chosen as $\alpha=\frac{j}{N}$ for a $j\in\{1,\dots, N\}$. With the set of all $j$-point averages of parameter tuples denoted by then

Figures (4)

  • Figure 1: The three phases of our proposed approach.
  • Figure 2: Power increase (green dotted line) and decrease (red dotted line) potential for a duration of 3h with respect to a baseline consumption (blue line), computed with $\alpha = 1$.
  • Figure 3: Results of the self-consumption (a, b) and peak reduction (c, d) experiments with 100 buildings and a maximum activation time of 3h per building and impact of varying parameters in the peak reduction experiments with $\alpha=1$ (e, f).
  • Figure 4: Average solving times of the peak reduction scheduling problem for varying numbers of assets in three request cases.

Theorems & Definitions (14)

  • Definition 1: Relative Consumption Request
  • Definition 2: Flexibility envelope
  • Remark 1
  • Remark 2
  • Lemma 1
  • proof
  • Definition 3: Uncertainty-aware flexibility envelope
  • Remark 3
  • Corollary 1
  • Definition 4: Self-consumption request trajectory
  • ...and 4 more