Dear Reader,

For the 52end issue of our heatbeat Research Newsletter, we take a look into a recently published research paper in the field of data driven models in district heating.

The paper „Enhanced methodology for disaggregating space heating and domestic hot water heat loads of buildings in district heating networks“ from Borgato et al. investigates different linear regression models to disaggregate space heat and domestic hot water demands based on district heating monitoring data.

The research paper presents an advanced methodology for disaggregating space heating (SH) and domestic hot water (DHW) heat loads in district heating networks (DHNs), addressing the limitations of existing models that struggle with daily fluctuations and seasonal variations. The ability to accurately separate SH and DHW consumption is essential for optimizing DHN operations, improving energy efficiency, and supporting retrofitting interventions in buildings.

The paper evaluates four different regression-based models to enhance heat load disaggregation, each improving upon the traditional Energy Signature Curve (ESC) method:

1. CPT-1: Uses a single ESC for the entire year based on Change Point Temperature (CPT) to separate SH and DHW.
2. CPT-24: Extends CPT-1 by introducing 24 independent ESC models, one for each hour of the day, allowing better representation of daily fluctuations.
3. Eref-1: Replaces CPT with a heat consumption threshold (Eref) to more accurately identify seasonal transitions, avoiding errors caused by temperature-based segmentation.
4. 4. Eref-24: Combines the improvements of Eref-1 with 24 hourly ESC models, providing the most detailed and accurate disaggregation of SH and DHW loads.

The models were tested on data from 27 buildings in Tartu, Estonia, using smart meter readings and climatic data. Various key performance indicators (KPIs) were used to assess model accuracy, including R² values, Yearly Energy Consumption Deviation (YEC), Winter and Summer Energy Consumption Deviations (WEC, SEC), and new qualitative metrics such as Dynamic Time Warping Distance (DTWD) and Percentage of Matching Labels (PML, PMLHR).

The study demonstrates that the Eref-24 model significantly outperforms the state-of-the-art CPT-1 method in multiple ways:

1. Improved Accuracy in Seasonal Transitions
2. Better Representation of Daily Heat Consumption Trends
3. More Realistic DHW Load Estimation
4. Reduction of Errors in Heat Load Forecasting
5. Enhanced Model Robustness for Different Building Types

The improved disaggregation methodology has significant implications for DHN management, enabling better demand forecasting, integration of renewable energy sources, and optimization of heating supply strategies. Operators can use these insights to fine-tune heat production, reducing inefficiencies and lowering operational costs.

The study presents a major advancement in heat load disaggregation, with the Eref-24 model proving to be the most accurate and reliable approach. By improving seasonal detection, capturing hourly variations, and better estimating DHW consumption, this methodology provides valuable insights for DHN optimization, energy-efficient retrofitting, and climate-friendly urban heating solutions.

Further Information

As always, we recommend reading the article in full. We would also like to share a little note on our own behalf: In addition to this research newsletter and various blog posts, we have added a monthly feature update to our blog, summarizing important developments and new features in our heatbeat Digital Twin. You can find the first entry at "https://heatbeat.de/en/blog/64/ "

The next issue of our newsletter will be published on March 5, 2025.