Status : Verified
| Personal Name | Cabaoig, Ronald R. |
|---|---|
| Resource Title | Hierarchical AC optimal power flow of multi-microgrid systems using a combination of learning-based and Lagrangian methods |
| Date Issued | February 2024 |
| Abstract | The concept of a multi-microgrid system (MMGS), an interconnected network of microgrids (MGs) sharing a common distribution system (DS), is gaining traction as a solution to improve grid resilience and self-sufficiency of consumers. A hierarchical control of an MMGS facilitates the effective coordination of MGs within a DS. In terms of power sharing, this control uses a hierarchical optimal power flow (OPF) at both the DS and MG levels to determine the optimal system setpoints. Now, with the increasing uncertainty brought about mainly by renewable energy (RE) resources, generator controls require more frequent adjustments to maintain supply-demand balance. However, these are dependent on the OPF runtimes. A large number of OPF simulations can also be conducted in some planning tasks, but the computational times could compound. Various linearizations or approximations of the full AC OPF are currently being used to attain faster solve times, but these sacrifice some information, resulting to suboptimal and AC-infeasible solutions. As such, it is still desirable to use the full AC OPF to ensure results of high fidelity. Learning-based AC OPF models are being explored as a potential solution by moving the computational burden offline or before operations and by providing fast online solutions while still maintaining a good accuracy. Applying a learning-based AC OPF approach to alleviate the computational speed concern in the context of a hierarchical control of an MMGS has not been explored yet. This study implements a bi-level AC OPF using a learning-based model, specifically an artificial neural network-multilayer perceptron (ANN-MLP), at the DS level and a Lagrangian/conventional method at the MG level. To supplement the baseline ANN for OPF in the literature where only the load values as inputs and generator control variables as outputs are included, the learning-based model at the DS level also incorporates the RE generation, MG power exchange information and |
| Degree Course | Master of Science in Electrical Engineering |
| Language | English |
| Keyword | multi-microgrid system; AC optimal power flow; hierarchical control; economic dispatch; machine learning |
| Material Type | Thesis/Dissertation |
Preliminary Pages
744.79 Kb
Category : F - Regular work, i.e., it has no patentable invention or creation, the author does not wish for personal publication, there is no confidential information.
Access Permission : Open Access