Using AI to Analyze Water Flow Beneath a Dam

In this assignment, students performed seepage analysis beneath a dam using flow nets and compared their work with AI-generated flow nets. They tried different prompts and checked the AI's results. 

Authors: 

Civil and Environmental Engineering Department

  • Mohamad Mahdi Hallal, Assistant Professor
  • Adda Athanasopoulos-Zekkos, Associate Professor
  • Arwen Gosch, Graduate Student
  • James Nguyen, Project Engineer

Course Number & Title: CIVENG 175, Geotechnical and Geoenvironmental Engineering

This work is licensed under CC BY-NC-SA 4.0

Details
Assignment Title
 		Using AI to Analyze Water Flow Beneath a Dam
Delivery FormatIn-person
Learning Objectives
  • Apply flownet principles to estimate seepage, uplift pressures, and exit gradients beneath hydraulic structures.
  • Develop effective prompts that communicate engineering procedures and constraints to AI tools.
  • Evaluate AI-generated engineering solutions by verifying results against fundamental principles and published research findings.
  • Develop AI literacy skills, including prompt development, iterative refinement, critical evaluation of AI outputs, and engineering judgment in assessing AI-assisted analyses.
Brief Summary of Assignment

Students perform seepage analysis beneath a dam using flow nets and compare traditional engineering analysis with AI-assisted flow net generation. The exercise introduces prompt development, where students translate engineering procedures and constraints into instructions for a GenAI tool, evaluate AI-generated results, and compare their findings with published research on AI-assisted seepage analysis. Students assess the capabilities and limitations of AI for engineering problems requiring spatial reasoning, which continues to be a challenge, and compare their experiences with findings from a published research article to determine whether the observed strengths and limitations of AI tools are consistent with current research. 

GenAI tools used: Any comparable large language model.

Guardrails Shared with StudentsThe assignment includes AI-use indicators that communicate expectations at the question level. These indicators can distinguish between questions where AI is not permitted (NAI), partially permitted (PAI), or fully permitted (FAI). In the NAI category, AI tools are not permitted at any stage of the task. In the PAI category, AI tools may be used for limited purposes, such as brainstorming, outlining, debugging, or providing feedback, but not for generating final solutions. In the FAI category, AI tools may be used throughout the task, including generating content or final solutions. These AI-use indicators would then be included next to each question to clearly and explicitly communicate AI use expectations.
Assignment Details
Content tags