Research Group
The Center for Combined Smart Energy Systems (CoSES) is a low-inertia, sector-coupled, low voltage active distribution grid laboratory to investigates smart operation strategies for sustainable energy systems at distribution grid prosumer levels. Our focus lies on the complex interaction of different subsystems, considering three integrated layers: thermal energy supply, electrical energy supply, communication & computation network. The research of our young group covers data driven model validation and calibration, real-time and optimal energy management, and dynamic control strategies of electrical and heat prosumers. The lab hosts the energy infrastructure of five realistic buildings, with real and emulated DERs, and a bidirectional interaction with the electrical and heating networks. The lab uses Power Hardware-in-the-loop (PHiL) philosophy to implement developed control and simulation models in real-time over a real long reconfigurable electrical grid and an emulated heating grid.
Topic Background
Optimally operated home energy management systems (HEMS) have been shown in literature to be effective in relieving the grid expansion costs, to accommodate the massive electrification of electrical, heating and mobility sectors through renewable energy resources. The uncertainties in generation, individual load profiles and the interplay between heat and electrical networks, provides many interesting challenges for EMS control algorithms.
Research Topic
EMS algorithms fall into rule-based controls, mathematical optimization, and data-driven methods like Reinforcement-Learning (RL). Rule-based methods are simple to develop but lack predictive control efficiency. Mathematical optimization minimizes system but gets intractable if it tries to mimic stochastic real-world behaviors. RL models can learn real world behavior, handle forecast errors well, but demand intricate training set. A gap remains in EMS algorithm development for real-world dynamics and uncertainties while being robust with manageable implementation effort. Physics based machine learning is an approach to combine the robustness of data driven modelling with physical constraints in the model to derive the best of both – mathematically precise and heuristics world. Could this be the key to develop a practically useful EMS algorithm?
Tasks:
• Literature review of active distribution grid control and EMS algorithms.
• Familiarization with the control system of CoSES laboratory.
• Familiarization with RL implementation frameworks
• Building up co-simulation environment to collect training set, ground truth for a data driven EMS algorithm.
• Developing a basis for a physical loss function.
• Benchmarking on a suitably chosen scenario set.
• Result processing, evaluation and presentation
• Preparation of a report describing the key aspects and findings of the conducted work
• Presenting the results to a scientific audience.
• Performing daily lab work duties of a student research assistant.
