The increasing demand for advanced grid support functionality from a large number of DERs has sparked significant interest in that focuses on optimization methods for large-scale unbalanced power distribution systems, aimed at enhancing operational efficiency and resilience. Both the traditional mathematical optimization methods and machine-learning (ML) -based approaches are gaining traction to attain optimal solutions for the scaled power distribution systems. However, when developing any optimal power flow (OPF) algorithm for distribution system, it is crucial to integrate unbalanced distribution power flow models as constraints. This poses the most challenging aspects of formulating any OPF problem. Besides, for ML-based approaches, creating extensive training data is time-consuming task that significantly delays the process. This tutorial aims at presenting a python-based distribution OPF tool (DistOPF) that develops power flow models as a constraint from the standard network inputs, allowing users to focus on other aspects. Specifically, this python-based package can quickly create the underlying mathematical equations of the power flow model for any distribution system and allows researchers to avoid the need to repeatedly model power flow models and instead leverage existing models that can be readily selected. It also formulates standard OPF related constraints. Thus, this package alleviates challenges to formulate OPF problems to develop and provide grid-support functionality and enhance grid resilience for the power distribution system. In short, the DistOPF package/tool will provide researchers/users a unique platform (i) to create new OPF algorithms by leveraging formulated unbalanced power flow models as the optimization constraints, (ii) to benchmark developed OPF algorithms (for both planning stage solutions and operation stage controls) for the distribution networks using provided models/solutions, and (iii) to create extensive training data to train the advanced machine learning based OPF algorithms.”

Presenter(s)

1. Dr. Rabayet Sadnan,

Research Scientist, Pacific Northwest National Laboratory (PNNL)

2. Dr. Anamika Dubey

Associated Professor, School of Electrical Engineering and Computer Science, Washington State University

3. Nathan Gray

Research Assistant, School of Electrical Engineering and Computer Science, Washington State University