Skip to content

albertodiazdurana/residential-heating-data-science-guide

Repository files navigation

Data Science for Residential Energy Systems

A domain knowledge framework bridging energy engineering fundamentals with data science and machine learning for building energy optimization. Developed as part of ongoing research toward a methodological contribution: A Framework for ML Deployment in Building Energy Management Systems.

Focus: Systematic documentation of German heating standards (DIN, VDI, GEG) with applied ML methodologies for residential heating system optimization. Structured using Take AI Bite, a framework for human-AI collaboration powered by the Deliberate Systematic Methodology (DSM).


Motivation

The energy sector faces a critical skills gap. According to the IEA World Energy Employment 2025 report, 60% of energy companies report labor shortages, with particular demand for data analytics and ML expertise. Meanwhile, academic reviews note that resources bridging ML research and building energy applications remain fragmented across disciplines.

This repository addresses that gap through:

  • Comprehensive documentation of German heating standards for data scientists, ML engineers, and data engineers
  • ML methodologies specifically adapted for energy time series analysis
  • Production-grade MLOps and data engineering patterns for IoT and sensor data systems
  • Applied case studies demonstrating real-world optimization approaches

Repository Structure

Part Topic Scope
Part I Domain Fundamentals Thermodynamics, heating system types, German standards
Part II Data Science & ML Time series analysis, forecasting, anomaly detection, optimization
Part III Production & MLOps Data pipelines, deployment patterns, monitoring
Part IV Technical Stack Python, SQL, GraphQL, API design
Part V Applied Scenarios Case studies, system design, cross-functional collaboration
References References Academic sources, regulations, technical glossary

Target Audience

  • Data Scientists, ML-, and Data-Engineers transitioning into energy and building optimization domains
  • Energy Engineers integrating ML and data-driven approaches into practice
  • Researchers in building science, smart buildings, and energy efficiency
  • Graduate Students in energy systems, building physics, or applied ML

Content Overview

Part I: Domain Fundamentals

Systematic review of residential heating system engineering:

  • Chapter 1: Thermodynamic principles (heat transfer mechanisms, thermal mass, U-values)
  • Chapter 2: Heating system typology (gas boilers, heat pumps, district heating, CHP)
  • Chapter 3: Control parameters (Heizkennlinie, Vorlauf/Rücklauf temperature management)
  • Chapter 4: Hydraulic balancing methodology (Hydraulischer Abgleich)
  • Chapter 5: Sector coupling (PV + heat pump + storage integration)

Part II: Data Science & ML

Applied machine learning methodologies for energy optimization:

  • Chapter 6: Time series fundamentals for energy data
  • Chapter 7: Heat demand forecasting and energy production prediction
  • Chapter 8: Anomaly detection frameworks for heating systems
  • Chapter 9: Control and optimization algorithms (MPC, reinforcement learning)
  • Chapter 10: Supervised learning applications
  • Chapter 11: Unsupervised learning (building clustering, load profile segmentation)

Part III: Production & MLOps

Deployment and operational considerations for ML systems:

  • Chapter 12: Data pipeline architecture for IoT/energy systems (MQTT, TimescaleDB)
  • Chapter 13: Production-grade algorithm development
  • Chapter 14: MLOps frameworks (MLflow, experiment tracking)
  • Chapter 15: Deployment patterns (batch inference, real-time scoring, edge deployment)

Part IV: Technical Stack

Implementation and data engineering reference for energy systems:

  • Chapter 16: Python for energy data science (Pandas, NumPy, scikit-learn)
  • Chapter 17: Data access patterns (SQL window functions, GraphQL, REST APIs)

Part V: Applied Scenarios

Case study analysis and system design:

  • Chapter 18: Case study walkthroughs (district heating, heat pump systems, building portfolios)
  • Chapter 19: System design exercises
  • Chapter 20: Cross-functional collaboration patterns

Technical Glossary (German-English)

German English Technical Context
Heizkennlinie Heating curve Flow temperature as function of outdoor temperature
Vorlauftemperatur Flow/supply temperature Water temperature leaving the heat source
Rücklauftemperatur Return temperature Water temperature returning to heat source
Spreizung Temperature spread Difference between flow and return (Vorlauf - Rücklauf)
Brennwertnutzung Condensing operation Recovering latent heat from flue gas
Hydraulischer Abgleich Hydraulic balancing Optimizing flow rates across heating circuits
Nachtabsenkung Night setback Reducing heating during unoccupied hours
Wärmepumpe Heat pump Device using refrigeration cycle for heating
Fernwärme District heating Centralized heat distribution network
BHKW CHP (Combined Heat & Power) Simultaneous electricity and heat generation
Anschlussleistung Connection capacity Contracted power for district heating
Wärmegestehungskosten Heat generation costs Total cost per kWh of useful heat

See 06_References.md for complete glossary.


Key Performance Indicators

Metric Typical Range Significance
Flow temperature at 4°C outdoor 50-70°C Lower values indicate higher efficiency
Return temperature for condensing < 55°C Required threshold for gas boiler condensing operation
Heat pump COP 3-5 Coefficient of performance; higher = more efficient
Optimization savings potential 10-20% Achievable through control optimization
Systems without night setback ~70% Represents immediate optimization opportunity
Heat-pump-ready buildings ~17% Without major renovation requirements

Regulatory Framework (Germany)

Regulation Scope
GEG (Gebäudeenergiegesetz) Building energy efficiency standards
GEG §60b Mandatory heating system inspection requirements
GEG §60c Hydraulic balancing requirements
DIN EN 12831 Heat load calculation methodology
VDI 2067 Economic calculation for energy systems
VDI 6030 Radiator sizing and selection

Repository Ecosystem

This project is part of a three-repository system:

Repository Role Content
Residential Heating DS Guide (this repo) Framework & Theory Domain knowledge, methodology, German standards documentation
Residential Energy Apps (Live App) Implementation Heating curve simulator, interactive Streamlit applications
Take AI Bite (DSM) Meta-Methodology Human-AI collaboration framework powered by the Deliberate Systematic Methodology

Integration: Theory (this repo) is validated by Practice (code repo), executed using the Meta-methodology (DSM).


Contributing

Contributions are welcome. This repository aims to serve as a professional reference for the energy transition.

Contribution areas:

  • Technical corrections and clarifications
  • Code examples and implementation notebooks
  • Translation to other languages
  • Anonymized case studies

See CONTRIBUTING.md for guidelines.


Author

Alberto Diaz-Durana Freelance Data Scientist & ML Engineer | 10+ Years Experience

  • MSc Process, Energy & Environmental Systems Engineering (TU Berlin)
  • PhD Candidate, Energy Planning & Machine Learning (TU Berlin) -- CPOTE 2020 publication
  • Specialization in building energy systems, thermo-economic modeling, and applied ML for energy optimization
  • Production ML experience across energy, cement manufacturing, and process mining domains
  • GitHub | LinkedIn
  • Website | Blog

License

This work is licensed under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).

You are free to:

  • Share — copy and redistribute the material in any medium or format
  • Adapt — remix, transform, and build upon the material for any purpose

Under the following terms:

  • Attribution — Give appropriate credit and indicate if changes were made
  • ShareAlike — Distribute contributions under the same license

Acknowledgments

  • German heating standards documentation based on DIN, VDI, and GEG regulations
  • Weather data patterns informed by DWD (Deutscher Wetterdienst) open data
  • Load profiles referenced from BDEW standard load profiles

Status: Active development Last Updated: March 2026

About

Domain knowledge bridging energy engineering with ML/data science for residential heating optimization, developed with Take AI Bite (German standards, DIN/VDI/GEG)

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors