Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Aggregated demand flexibility prediction of residential thermostatically controlled loads and participation in electricity balance markets

Alejandro Martín-Crespo, Enrique Baeyens, Sergio Saludes-Rodil, Fernando Frechoso-Escudero

TL;DR

The paper addresses the challenge of predicting how much demand flexibility a TCL-based aggregation (virtual battery, VB) can reliably provide for grid balancing markets. It introduces a probabilistic power-capability function ${\Phi(x)}$ and develops MC&ESB, a method that combines Monte Carlo simulation with a bisection extremum search to estimate the maximum flexible power ${|x|}$ with guaranteed probability. Through three SEBM-focused case studies, the authors demonstrate the ability to quantify available flexibility, generate market-ready bids, and validate controller performance in delivering the requested power. The work enables robust demand-side participation by providing rigorous probabilistic guarantees for VB flexibility and offers a practical framework for aggregators aiming to bid in ancillary services using residential TCLs.

Abstract

The aggregate demand flexibility of a set of thermostatically controlled residential loads (TCLs) can be represented by a virtual battery (VB) in order to manage their participation in the electricity markets. For this purpose, it is necessary to know in advance and with a high level of reliability the maximum power that can be supplied by the aggregation of TCLs. A probability function of the power that can be supplied by a VB is introduced. This probability function is used to predict the demand flexibility using a new experimental probabilistic method based on a combination of Monte Carlo simulation and extremum search by bisection algorithm (MC&ESB). As a result, the maximum flexibility power that a VB can provide with a certain guaranteed probability is obtained. The performance and validity of the proposed method are demonstrated in three different case studies where a VB bids its aggregate power in the Spanish electricity balancing markets (SEBM).

Aggregated demand flexibility prediction of residential thermostatically controlled loads and participation in electricity balance markets

TL;DR

The paper addresses the challenge of predicting how much demand flexibility a TCL-based aggregation (virtual battery, VB) can reliably provide for grid balancing markets. It introduces a probabilistic power-capability function and develops MC&ESB, a method that combines Monte Carlo simulation with a bisection extremum search to estimate the maximum flexible power with guaranteed probability. Through three SEBM-focused case studies, the authors demonstrate the ability to quantify available flexibility, generate market-ready bids, and validate controller performance in delivering the requested power. The work enables robust demand-side participation by providing rigorous probabilistic guarantees for VB flexibility and offers a practical framework for aggregators aiming to bid in ancillary services using residential TCLs.

Abstract

The aggregate demand flexibility of a set of thermostatically controlled residential loads (TCLs) can be represented by a virtual battery (VB) in order to manage their participation in the electricity markets. For this purpose, it is necessary to know in advance and with a high level of reliability the maximum power that can be supplied by the aggregation of TCLs. A probability function of the power that can be supplied by a VB is introduced. This probability function is used to predict the demand flexibility using a new experimental probabilistic method based on a combination of Monte Carlo simulation and extremum search by bisection algorithm (MC&ESB). As a result, the maximum flexibility power that a VB can provide with a certain guaranteed probability is obtained. The performance and validity of the proposed method are demonstrated in three different case studies where a VB bids its aggregate power in the Spanish electricity balancing markets (SEBM).
Paper Structure (14 sections, 1 theorem, 14 equations, 5 figures, 3 tables)

This paper contains 14 sections, 1 theorem, 14 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Materials and methods
  3. The TCL model
  4. The VB model
  5. The demand flexibility control
  6. Demand-side participation in SEBM
  7. Results
  8. The VB power supply probability function
  9. The MC&ESB method for estimating a VB power supply probability function
  10. Discussion
  11. Case study 1: Estimation of VB supply probability function
  12. Case study 2: Demand flexibility prediction
  13. Case study 3: Demand flexibility control
  14. Conclusion

Key Result

Theorem 1

Let $\epsilon$ and $\delta$ be scalars in the open unit interval $(0,1)$ then whenever

Figures (5)

  • Figure 1: VB controller diagram martin2021flexibility.
  • Figure 2: MC&ESB method diagram martin2021flexibility.
  • Figure 3: Estimated demand flexibility probability function $\Phi(x)$ and its confidence bounds for the VB of case study 1
  • Figure 4: VB deviation power. At minute 30 the demand flexibility control starts to actuate.
  • Figure 5: Absolute error of the VB controller.

Theorems & Definitions (1)

  • Theorem 1