Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A multi-dimensional unsupervised machine learning framework for clustering residential heat load profiles

Vasilis Michalakopoulos, Elissaios Sarmas, Viktor Daropoulos, Giannis Kazdaridis, Stratos Keranidis, Vangelis Marinakis, Dimitris Askounis

TL;DR

This paper introduces an unsupervised machine learning framework for clustering residential heating load profiles, focusing on natural gas space heating and hot water preparation boilers, and finds that DTW is the most suitable metric, uncovering strong correlations between boiler usage, heat demand, and temperature.

Abstract

Central to achieving the energy transition, heating systems provide essential space heating and hot water in residential and industrial environments. A major challenge lies in effectively profiling large clusters of buildings to improve demand estimation and enable efficient Demand Response (DR) schemes. This paper addresses this challenge by introducing an unsupervised machine learning framework for clustering residential heating load profiles, focusing on natural gas space heating and hot water preparation boilers. The profiles are analyzed across five dimensions: boiler usage, heating demand, weather conditions, building characteristics, and user behavior. We apply three distance metrics: Euclidean Distance (ED), Dynamic Time Warping (DTW), and Derivative Dynamic Time Warping (DDTW), and evaluate their performance using established clustering indices. The proposed method is assessed considering 29 residential buildings in Greece equipped with smart meters throughout a calendar heating season (i.e., 210 days). Results indicate that DTW is the most suitable metric, uncovering strong correlations between boiler usage, heat demand, and temperature, while ED highlights broader interrelations across dimensions and DDTW proves less effective, resulting in weaker clusters. These findings offer key insights into heating load behavior, establishing a solid foundation for developing more targeted and effective DR programs.

A multi-dimensional unsupervised machine learning framework for clustering residential heat load profiles

TL;DR

This paper introduces an unsupervised machine learning framework for clustering residential heating load profiles, focusing on natural gas space heating and hot water preparation boilers, and finds that DTW is the most suitable metric, uncovering strong correlations between boiler usage, heat demand, and temperature.

Abstract

Central to achieving the energy transition, heating systems provide essential space heating and hot water in residential and industrial environments. A major challenge lies in effectively profiling large clusters of buildings to improve demand estimation and enable efficient Demand Response (DR) schemes. This paper addresses this challenge by introducing an unsupervised machine learning framework for clustering residential heating load profiles, focusing on natural gas space heating and hot water preparation boilers. The profiles are analyzed across five dimensions: boiler usage, heating demand, weather conditions, building characteristics, and user behavior. We apply three distance metrics: Euclidean Distance (ED), Dynamic Time Warping (DTW), and Derivative Dynamic Time Warping (DDTW), and evaluate their performance using established clustering indices. The proposed method is assessed considering 29 residential buildings in Greece equipped with smart meters throughout a calendar heating season (i.e., 210 days). Results indicate that DTW is the most suitable metric, uncovering strong correlations between boiler usage, heat demand, and temperature, while ED highlights broader interrelations across dimensions and DDTW proves less effective, resulting in weaker clusters. These findings offer key insights into heating load behavior, establishing a solid foundation for developing more targeted and effective DR programs.

Paper Structure

This paper contains 33 sections, 10 equations, 23 figures, 12 tables.

Table of Contents

  1. Introduction
  2. Problem Setting & Literature Review
  3. Problem Setting
  4. Related Work
  5. Methodology
  6. Heating load profiles
  7. Heat demand dimension
  8. Temperature dimension
  9. Building dimension
  10. Boiler dimension
  11. User dimension
  12. Distance Measures
  13. Euclidean distance (ED)
  14. Dynamic time warping (DTW) distance
  15. Derivative dynamic time warping distance (DDTW)
  16. ...and 18 more sections

Figures (23)

  • Figure 1: Proposed Methodology.
  • Figure 1: Boiler and return water temperature difference for each house throughout the day
  • Figure 2: Data from a typical residential heating system during a day
  • Figure 2: Boiler modulation level for each house throughout the day
  • Figure 3: IL score for every number of clusters and algorithm for the three distance metrics in the boiler dimension
  • ...and 18 more figures