Abdul Azzam

Scientific Publications & Preprints:

• Google Scholar
• Preprints

Student Theses:

• Tim Dubies: "Development and Analysis of a Model-Predictive Control for an optimized Operation of Heating Grids in Energy Quarters," Master's Thesis, 2024
  • A district heating simulation model was built using PyDHN, integrating CHP, heat pump, solar thermal, and thermal storage. An MPC was implemented and compared against rule-based control using campus data from the University of Stuttgart. The MPC showed clear cost advantages when multiple flexible generators and storage are available, particularly with intraday electricity prices embedded in the objective function. Model inaccuracies causing suboptimal operation were also identified, pointing to further optimization potential.
• Lakshay Panjwani: "Structuring energy data using a database to evaluate ML models for load forecasts," Master's Thesis, 2024
  • This thesis developed and compared ML models for building energy load forecasting — linear regression, ridge regression, SARIMA, and LightGBM — incorporating weather, calendar, and building-specific features. SARIMA achieved the highest accuracy. A modular DBMS architecture was designed to manage time series energy data and support multi-stakeholder access, providing a scalable foundation for future integration of real-time data and advanced analytics.
• Jonas Schleh: "Distributed Optimized Control of the Heating Network at Campus Vaihingen", Research Thesis, 2025
  • This thesis developed a distributed MPC framework for the University of Stuttgart's district heating network using an ADMM algorithm to decompose the system into communicating subsystems (CHP, heat pump, solar thermal, thermal storage, heat consumer). The approach was validated against centralized MPC. The centralized controller outperformed on cost and demand coverage due to global optimization, while the distributed approach achieved a higher share of renewable energy but showed inaccuracies in exact demand satisfaction. The work demonstrates ADMM's capability to coordinate local optimization across subsystems in a real-world district heating context.
• Kareem Hassan: "Development and Analysis of a Dynamic Model for State Estimation of Electrical Networks for Application in Model Predictive Control", Student Research Project, 2025
  • This thesis developed a dynamic state estimation model for medium-voltage grids with PV and battery storage within an MPC framework. Both DC power flow and linearized AC optimal power flow models were implemented and compared. A degradation-aware battery model reduced operational costs and extended battery lifetime. The DCPF model proved computationally efficient for large-scale optimization, while the linearized ACOPF captured reactive power and voltage behavior more accurately. Results were validated against established power flow tools.
• Leon Scheurer: "Development and Analysis of a Dynamic Grid and Heat Pump Model for a Model Predictive Control of an Energy District", Thesis, 2025
  • Building on the existing DiTEnS simulation model, this thesis developed a dynamic large-scale heat pump model based on empirical efficiency and start-up/shutdown data, and analyzed two dynamic district heating network prediction approaches: an instationary energy balance model and a graph-theory-based model. The large-scale heat pump produced realistic seasonal behavior. The energy balance network model yielded comparable results to the reference but introduced temperature fluctuations that increased costs. The graph-theory model caused the storage to deactivate and produced unrealistic temperature peaks, indicating that further model detail is needed for practical application.
• Hassan Ghazle, Michael Neubrander, Charlotta Wallentin: "Forecast Model for Revenue Estimation of Charging Stations Considering Variable Influencing Factors", Student Project (Renewable Energies), 2026
  • Using real charging data from four MVV Energie AG stations in Mannheim, this project implemented and compared SARIMA, LightGBM, and LSTM models for EV charging load forecasting. The LSTM achieved the lowest errors and was the only model recommended for practical deployment. SARIMA and LightGBM did not reach sufficient accuracy on the available dataset. The authors note that the limited training data constrains the conclusions, and that larger datasets would likely improve all models and potentially change the relative rankings.
• Caroline Fegert: "MPC-Based Control for a Residential Energy District with Bidirectional EV Charging", Thesis, 2025
  • This thesis developed an MPC framework for a residential district with PV generation and uni-/bidirectional EVs, investigating three scenarios: self-consumption optimization, cost optimization, and grid-oriented peak shaving. Each scenario was benchmarked against a unidirectional reference. Bidirectional charging improved objective achievement in all three cases, but the magnitude of benefit depends strongly on the objective: improved PV self-consumption, economically motivated temporal storage flexibility, and effective peak reduction at the grid connection point, respectively. The work also identifies trade-offs around battery utilization and grid interaction.
• Chenming Zhan: "Robustness Improvement of Model Predictive Control (MPC) for Building Energy Systems", Research Thesis
  • This thesis developed a robustness-focused MPC framework for small-scale building energy systems, tackling the computational instability that conventional MPC can experience under extreme conditions. The approach uses soft constraints and slack variables as a mathematical robustness layer, complemented by physical redundancy for resilience. A novel "Component Health" concept monitors individual components in real time and automatically adjusts constraints and parameters when anomalies are detected, triggering degradation or exclusion strategies. Multiple user-selectable multi-objective reconfiguration strategies add operational flexibility. The model is implemented in Python.

Experience:

• Current Position:
  • PhD Student, Institute for Energy Economics and Rational Energy Use, University of Stuttgart (since 2023) Project: Discursive Transformation of Energy Systems (DiTEnS)
  • Lecturer, DHBW Stuttgart (September 2023 - present) Lectures: Fundamentals of Automation, Control Engineering
• Previous Positions:
  • Research Thesis, Centre for Solar Energy and Hydrogen Research Baden-Württemberg (ZSW) (August 2021 - October 2022): Development and Analysis of a Distributed Predictive Control Algorithm for Smart Grids with Decentralized Virtual Power Plants
  • Research Assistant, Centre for Solar Energy and Hydrogen Research Baden-Württemberg (ZSW) (July 2019 - September 2021)
  • Technical Assistant, Nagel Machines and Tool Factory GmbH (October 2018 - May 2019)
  • Cooperative Studies Mechanical Engineering (DHBW), bielomatik GmbH (October 2015 - September 2018)

Education:

• University of Stuttgart: Master of Science - MS, Engineering Cybernetics (April 2019 - March 2023) Fields: Autonomous Systems & Control Theory, Automation in Energy Engineering, Artificial Intelligence, Computer Science
• Universitat de Barcelona: Master of Science - MS (February 2022 - July 2022) Fields: Neurobiology and Neuroscience, Parallel Programming (CUDA and OpenMP)
• Hochschule Esslingen - University of Applied Science: Master of Engineering, Systems Engineering/Mechatronics (October 2018 - April 2019) Fields: Control Engineering & Numerical Mathematics
• Baden-Wuerttemberg Cooperative State University (DHBW): Bachelor of Engineering - BE, Mechanical Engineering (September 2015 - September 2018) Field: Design & Development