Sponsor: Foothill Section Chapter, PE31
Speaker: Dr. Koji Yamashita

Meeting Date: 30 Jun 2023
Time: 02:00 PM to 03:00 PM
Cost:
Location: This is a Virtual Presentation. You must register through the link provided under “REGISTRATION”.
Reservations: IEEE
Summary:
Recently, the integration of inverter-based resources (IBRs) in renewable energy systems has introduced new dynamic challenges to the power grid. These challenges include converter-driven oscillations and the expansion of power outage areas, which pose a significant risk to the reliability of future grids. Grid sensor technology has been developed, enabling the collection of a vast amount of measurement data. This data presents an opportunity for machine-learning experts to leverage big data analytics and tackle various grid reliability analysis challenges specific to the future power system. However, additional technical hurdles are related to data integrity and consistency, particularly regarding labeling and other factors. In power system analysis, time-domain simulation models have traditionally been used for contingency analysis, post-mortem analysis, grid controller design, and grid-wide protection. Synthetic data, which refers to simulated responses, has emerged as a promising approach for imputing missing events and measurements in the data. However, the validation of simulation models, especially for grid-wide events, has been insufficient. This means that synthetic data from numerical simulation models are accurate for limited power system dynamic studies but may not be adequately validated for broader applications. In addition, there is a lack of collaboration between modeling experts specializing in electromagnetic transient modeling, electromechanical modeling, IBR modeling, and grid modeling. Furthermore, no generally accepted model validation procedure or requirements is in place. As a result, well-validated model parameters for simulation studies are scarce, even though the models themselves are readily available. Given these challenges, machine-learning experts must perceive the potential missing dynamics or model errors when utilizing synthetic data for data-driven research studies. This presentation aims to address the current quality of synthetic data and provide insights on navigating the complexities associated with big data in grid reliability analysis. Join us for an engaging discussion on this topic.

Bio: Dr. Koji Yamashita received his B.S. and M.S. degrees in Electrical Engineering from Waseda University, Tokyo, Japan, in 1993 and 1995, respectively. He went on to complete his Ph.D. in Electrical and Computer Engineering from Michigan Technological University, Houghton, MI, USA, in 2020.