Formerly. ME 292E "MEC ENG 292E Advanced Special Topics in Energy Science and Technology

Machine Learning Tools for Modeling Energy Transport and Conversion Processes

Instructor: Prof. Van P. Carey

Course Description

This course will provide an introduction to basic probability, linear algebra concepts, and the foundation mathematics principles used in machine learning tools. Python programming will be used to implement machine learning methods in class projects. Projects will be scoped so Python programming can be learned as students work on them. In subsequent sections of the course, students will construct a genetic algorithm and a neural network model from scratch to explore basic features of these tools, and will then progress to using Python neural network programming tools to develop models for energy conversion and energy transport process applications.

The course will emphasize physics-inspired strategies that frame the problem, preprocess data, and execute training of the model so that the usefulness of the results is maximized. The model of the physics for projects will be described in materials made available to students, so specialized knowledge of the physics for project systems will not be required. Students will work in teams of two on projects that will involve construction of one or more computer programs to implement a machine learning methodology to analyze data from models or experiments for energy conversion or transport processes. Students can select the topic for their final project from options provided by the instructor, or can propose a topic related to their research or capstone project. Students completing this course will develop an understanding of the fundamental aspects of machine learning tools and will gain expertise in using them for different types of applications.

Prerequisites: Undergraduate classes in computer programming and thermodynamics. Undergraduate courses in fluid mechanics and heat transfer are recommended but not required." show less