ChatGPT & Mechanical Engineering: Examining performance on the FE Mechanical Engineering and Undergraduate Exams
Project Overview
The document examines the role of generative AI, particularly ChatGPT, in enhancing education, with a focus on its applications in mechanical engineering and mathematical principles related to heat transfer. It highlights ChatGPT's testing in answering engineering exam questions and its performance across different educational contexts, revealing both benefits, such as aiding learning and problem-solving, and challenges, including inconsistencies and inaccuracies that necessitate careful use by students and educators. Additionally, it delves into the mathematical foundations of convective and radiative heat transfer, providing essential formulas and a structured approach to calculating the convective heat transfer coefficient and heat loss from a heated plate based on established laws. This comprehensive analysis underscores the potential of generative AI to support educational outcomes while also emphasizing the need for critical engagement with its limitations.
Key Applications
ChatGPT for Educational Support
Context: Various contexts within undergraduate mechanical engineering education, including exam preparation for the Fundamentals of Engineering Exam (FE), development of learning materials such as syllabi and lesson plans, and aiding students in understanding heat transfer principles through theoretical calculations.
Implementation: ChatGPT is utilized to assist students and educators by providing answers to engineering exam questions, generating personalized learning materials, and facilitating the application of theoretical concepts in practical scenarios. It has been tested for accuracy in exam settings and is used to create tailored educational resources.
Outcomes: ChatGPT has demonstrated a 76% accuracy rate on exam questions, showing its capability to enhance learning and exam preparation. It can also improve educational resources and provide tailored problem sets, enhancing students' understanding of complex engineering principles.
Challenges: Despite its utility, ChatGPT presents challenges such as the inability to process images, a tendency to provide incorrect answers with confidence, and an overall inconsistency in responses. Educators must verify the accuracy of content generated for learning materials, and students may struggle with applying theoretical equations to practical scenarios.
Implementation Barriers
Technical Barrier
ChatGPT currently cannot process or generate images, limiting its application in engineering subjects that require visual understanding.
Proposed Solutions: Instructors can design assessments that do not rely on visual aids, focusing on text-based questions and ensuring students engage with the material directly.
Educational Barrier
There is a concern that over-reliance on ChatGPT may weaken students' critical thinking and problem-solving skills. Additionally, students may find it difficult to relate abstract equations to practical applications in heat transfer.
Proposed Solutions: Encouraging a focus on the problem-solving process rather than merely obtaining correct answers can help maintain students' analytical skills. Incorporate practical labs and simulations to demonstrate heat transfer principles in real-time.
Resource Availability
Access to up-to-date tables of air property values may be limited.
Proposed Solutions: Provide digital resources or online databases where students can access current air properties.
Project Team
Matthew Frenkel
Researcher
Hebah Emara
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Matthew Frenkel, Hebah Emara
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai