Self-organizing voltage regulation in the distribution networks: Insights into planning, operation and validation

Global energy trends are experiencing a significant shift characterized by a growing movement toward the integration of Distributed Renewable Energy Sources (DRES), such as wind and solar energy, into the power grid. Accelerated by technological advancements and supportive policy initiatives, this transition aims to reduce our reliance on fossil fuels, promote local energy generation, and improve energy security. However, the extensive penetration of DRES into the power grid presents its unique set of challenges.

A big challenge associated with the integration of DRES is their innate intermittency and unpredictability, which induce fluctuations in power availability and demand. Such fluctuations could lead to voltage instability, frequency deviations, and general power quality problems within the power grid. Moreover, the traditional power grid, which is largely unidirectional in design, cannot manage the bidirectional power flow resulting from DRES integration. As a result, ensuring the stability and reliability of the power grid becomes essential with the widespread integration of DRES. Furthermore, incorporating DRES requires innovative grid planning and operation methodologies to optimize resources and prevent potential congestion.

Motivated by these challenges, this thesis develops and implements the method on identifying the main barriers to increasing the integration of DRES in distribution networks (DN) and developing the solution that can enable high DRES penetration levels in power grids, thereby supporting the transition to a future 100\% renewable energy system. This thesis provides a solution for three critical phases for future smart power grids: planning, operation, and validation.

Planning phase: The thesis introduces a stochastic simulation-based approach to assess DRES penetration levels and the capacity requirements for the central Battery Energy Storage System (BESS) in DNs while ensuring technical constraints. The stochastic method creates a wide range of scenarios under various conditions. For each scenario, my proposed approach calculates the maximum allowable DRES penetration level and the required BESS capacity with different DRES control logic. The maximum allowable DRES penetration level and the BESS capacity requirements are then determined by analyzing various simulation results. The unique contribution lies in equipping distribution system operators (DSO) with the ability to compare results and select the most appropriate voltage control and power smoothing methods. This helps address the challenges associated with voltage violation and intermittency issues arising from DRES-generated power, thus improving the overall resilience and reliability of the power grid. Moreover, data analysis techniques are utilized to compare the efficacy of various local voltage and BESS control methodologies, offering valuable insights for network planners.
Operation phase (DSO): Building on the foundational knowledge acquired in the planning phase about DRES high penetration level network, a novel algorithm is proposed to achieve optimal voltage regulation through the self-organizing actions of agents. This algorithm empowers distributed agents to coordinate and collaborate in real time to regulate voltage in DNs with high DRES penetration. The proposed method can minimize the number of agents involved in the voltage regulation and the change of required power for voltage regulation, which together minimizes the need for re-dispatching, i.e., the impact of voltage regulation on the exchange of energy. Moreover, the proposed method performs online optimization, i.e., the value of the decision variable is physically implemented as a controller set-point at each iteration, which reduces the response time. The presented algorithm is benchmarked against the alternating direction method of multipliers (ADMM) algorithm and centralized optimization to validate its efficiency.

Operation phase (Energy community): While coordinating DRES strategies from the grid's viewpoint is vital, it's equally important to consider the energy management of energy communities. I propose a comprehensive four-stage energy management approach that employs receding-horizon optimization to stabilize power fluctuations in a residential energy community system. This system comprises a photovoltaic (PV) installation, a BESS, and a hydrogen system with an electrolyzer, a fuel cell, and a hydrogen storage. This innovative approach uniquely integrates four optimization stages, i.e., yearly, monthly, day-ahead, and intra-day. It blends long-term and short-term optimization techniques in EMS development to utilize hydrogen generated via electrolysis as seasonal storage. The introduced algorithm incorporates three modes with distinct objective functions for enhanced user adaptability. The approach is tested through simulations and operational analysis of an on-site PV–BESS–electrolyzer–fuel cell energy system field lab, including an in-depth analysis of system failure rates, system efficiency evaluation, and performance comparison across different modes of operation.


Validation Phase: To make our research more hands-on and highlight the challenges of DRES, I developed a practical simulation tool named The Illuminator. The Illuminator helps illustrate the challenges of DRES integration, acts as a sandbox for testing new research concepts in real and nonreal time, and allows real-world equipment simulations to check an algorithm before it is fully used. The Illuminator technology is primarily a modular software platform developed on a Raspberry Pi cluster. It is open-source, available on GitHub and developed in Python. The Illuminator comprises models of common energy technologies, such as PV panels, wind turbines, BESS, and hydrogen systems. The uniqueness of The Illuminator is in its modularity and flexibility to reconfigure scenarios and cases on the fly, even by non-experts in a plug-and-play fashion. I introduce The Illuminator and show its performance in two simple case studies.


This research improves the collective understanding of DRES integration by developing practical tools and methodologies that can significantly influence the design and operation of future power grids. Consequently, it paves the way for a cleaner, more efficient, and reliable energy system.