Full Transcripts - On the Potential of ChatGPT to Generate Distribution Systems for Load Flow Studies using OpenDSS

Citation Author(s):
Rodrigo
Bonadia
ERA Energy Research and Analytics
Fernanda
Trindade
University of Campinas
Walmir
Freitas
University of Campinas
Bala
Venkatesh
Toronto Metropolitan University
Submitted by:
Rodrigo Bonadia
Last updated:
Fri, 06/30/2023 - 08:24
DOI:
10.21227/91fj-gr09
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

This is the full ChatGPT transcript for the IEEE Power Engineering Letter "On the Potential of ChatGPT to Generate Distribution Systems for Load Flow Studies using OpenDSS". The abstract for the letter is as follows:

In recent years, the Large Language Models have developed at an unprecedented pace with the potential to revolutionize various fields of knowledge, including power systems. This letter illustrates the current status and potential use of GPT-3.5 and GPT-4 to create test distribution systems modeled as DSS files for load flow studies using OpenDSS, focusing on educational and research purposes. A performance comparison of GPT-3.5 and GPT-4 large language models (with the ChatGPT frontend) has been conducted. More specifically, the ability of ChatGPT to generate simple test circuits to run in OpenDSS is verified, including elements such as lines, loads, transformers, and photovoltaic generators. The ability of ChatGPT to identify and solve simple engineering problems applied to the generated circuits is also briefly discussed. The results demonstrate that GPT-4 has the potential to create functional circuits and propose solutions for engineering problems if adequate guidance and examples are provided.

Funding Agency: 
National Council for Scientific and Technological Development (CNPq) and São Paulo Research Foundation (FAPESP)
Grant Number: 
CNPq: #304373/2020-6, FAPESP: #2021/11380-5, #2022/11517-3, and # 2023/01719-0