POWERGPT

By Rithika Mathew, Siyuan Du
Slide 1: Title slide for PowerGPT presentation by Rithika Mathew and Siyuan Du

Slide-1

Faculty Mentor: Yufei Tang, Ph.D.

Student Mentor: Raul Mendy

Rithika Mathew, Siyuan Du

Slide 2: Smart Grid PowerGPT overview showing electrical network concepts and renewable energy integration

Slide-2

SMART GRID POWERGPT

Electrical Network

Match supply and demand

Integration of Renewable Energy

Incorporates Advanced Tech

[1] Schaller, J. (2020, June 1). Future strategies for data center smart grid integration. Mission Critical Magazine RSS.

Slide 3: Background section discussing smart grid challenges and existing maintenance strategies

Slide-3

BACKGROUND POWERGPT

Integration of new technologies into the electrical grid has a multitude of benefits, but also creates diverse strains on the system

Smart grids need more advanced maintenance and management techniques

Existing strategies:

Scheduled - Periodic deployment of teams to oversee machines and perform maintenance

Reactive - Relies on highly accurate sensing equipment, Prolonged service delays

Slide 4: Research gaps identifying need for proactive strategies and improved system-operator interaction

Slide-4

RESEARCH GAPS POWERGPT

Proactive Strategies: Lack of proactive prognostics strategies

Self-monitoring: Less expensive and time efficient

Ease of use: Need for more organic system-operator interaction

Slide 5: Proposed solution PowerGPT featuring large scope, dynamic visuals, predictive capabilities, and UI improvements

Slide-5

PROPOSED SOLUTION - POWERGPT

Large Scope: National grid map and wind power prediction

Dynamic Visuals: Transformer graphics, map, and diagrams

Predictive: Transformer based fault detection

UI: Tkinter to React based

History: PostgreSQL storage

Session Management: Login Page

Slide 6: Features overview including login, history, interactive features, wind power, and power grid capabilities

Slide-6

FEATURES POWERGPT

Login: User authentication

History: Chat retrieval

Interactive Features: Graph & diagram

Wind Power: Live generation predictions

Power Grid: Color-coded plants

Slide 7: Login and Chat History features showing user authentication and database storage capabilities

Slide-7

LOGIN / CHAT HISTORY POWERGPT

User Query history

Automatically logs to PostgreSQL database for retrieval

Login page for session management

Information stored on database

Password login planned for future implementation

Features such as chat history will be per-user based

Slide 8: Interactive fault detection system using transformer-based model with visual diagram highlighting

Slide-8

INTERACTIVE FAULT DETECTION POWERGPT

Uses trained transformer-based model to predict power transformer faults

When prompted for fault detection, chatbot takes user defined signal and feeds it into the transformer model

Bot outputs fault prediction

Highlights faulty parts on a diagram of a power transformer

Graphs the signal

Slide 9: GridWatch feature showing global power grid monitoring with world map visualization

Slide-9

GRIDWATCH POWERGPT

This slide displays a world map visualization showing global power grid monitoring capabilities, with various power generation facilities and grid connections represented across different continents.

Slide 10: Wind Power prediction system details including dataset, training, integration, and user interaction features

Slide-10

WIND POWER POWERGPT

Dataset: 2017 data of a 2.5 MW Clipper Liberty wind turbine in Minnesota

Training: LightGBM library, Final RMSE of around 150 kW

Integration: Accesses live wind speed data for predictions

Interaction: User can query and ask for predictions from any city.

Slide 11: Wind Power interface screenshot showing prediction results and data visualization

Slide-11

WIND POWER POWERGPT

This slide shows a screenshot of the wind power prediction interface, displaying real-time wind power generation predictions with numerical data and interactive elements for user queries and results visualization.

Slide 12: Chatbot functionality description emphasizing organic conversation and framework integration

Slide-12

CHATBOT POWERGPT

Makes interaction with the system organic and conversational

Prompt-based interaction with various features.

React and Flask Framework for connecting frontend and backend

Slide 13: Future works outlining security, enhanced chatbot, chat history, and wind prediction improvements

Slide-13

FUTURE WORKS POWERGPT

Security: Implement secure password protection for user accounts.

Enhanced Chatbot: Improve the chatbot's responsiveness to match the capabilities of ChatGPT.

Chat History: Include diagram and graph

Wind Prediction: Change to transformer model – better long-term dependencies

Slide 14: References section listing all cited sources and documentation

Slide-14

REFERENCES POWERGPT

  1. Schaller, J. (2020, June 1). Future strategies for data center smart grid integration. Mission Critical Magazine RSS. https://www.missioncriticalmagazine.com/articles/93008-future-strategies-for-data-center-smart-grid-integration
  2. Accurso, J., Mendy, R., Torres, A., & Tang, Y. (2023). A ChatGPT-like solution for power transformer condition monitoring. In 2023 International Conference on Machine Learning and Applications (ICMLA) (pp. 1716-1722). IEEE. https://doi.org/10.1109/ICMLA58977.2023.00260
  3. Asimislam. (2022, November 20). Global Power Generation - Eda & World Map. Kaggle. https://www.kaggle.com/code/asimislam/global-power-generation-eda-world-map
  4. Google. (n.d.). NWS weather - apps on Google Play. Google. https://play.google.com/store/apps/details?id=com.daugherty.nws_remastered&hl=en_US
  5. Davison, Brian. (2019). Rich Data for Wind Turbine Power Performance Analysis. Retrieved from the Data Repository for the University of Minnesota (DRUM), https://doi.org/10.13020/1etn-1q17.
Slide 15: Thank you slide with acknowledgments for NSF and FPL funding support

Slide-15

THANK YOU! POWERGPT

Acknowledgement: This work was supported in part by the U.S. National Science Foundation under Grant Nos. CNS-1950400 (I-SENSE REU SITE) & CMMI-2145571 (CAREER Award) and the FPL Intelligent Energy Technologies (InETech) Center Summer Internship Program at FAU.

Last slide: Contains plain text stating 'End of presentation. Click the right arrow to return to beginning of slide show.'

End of Presentation

Click the right arrow to return to the beginning of the slide show.

For a downloadable version of this presentation, email: I-SENSE@FAU.

Additional Information
The Institute for Sensing and Embedded Network Systems Engineering (I-SENSE) was established in early 2015 to coordinate university-wide activities in the Sensing and Smart Systems pillar of FAU’s Strategic Plan for the Race to Excellence.
Address
Florida Atlantic University
777 Glades Road
Boca Raton, FL 33431
i-sense@fau.edu