June 15, 2019
June 15, 2019
June 19, 2019
NSF Grantees Poster Session
The rapid growth in the field of artificial intelligence (AI) and pattern recognition can be attributed in part to the success of neural network architectures and machine learning (ML) algorithms as well as the availability of cloud computing resources. The applications of neural networks range from simple classification or prediction tasks to complex tasks such as feature learning for complex data. Recently, neural networks combined with signal processing analytics have found applications in renewable energy systems. With artificial intelligence tools for solar array systems and other applications becoming popular in both academia and industry, there is a need to train undergraduate students on AI tools and their applications. In our case, we will focus training activities on applying neural nets (NN) to renewable energy systems. We planned an initial immersive learning experience with hands-on experiments to help students understand basic ML and NN concepts and applications. We will follow then with a learning activity that will cover neural-network online modules and software for solar energy applications in the HTML5 J-DSP environment. We choose J-DSP HTML 5 as the software environment to allow students to learn concepts using a user-friendly block diagram programming approach. We have previously used this approach in our signals and systems and DSP classes for online labs in filter design and spectral analysis. We note that the utility of J-DSP has also been extended through the years in several multi-disciplinary areas. This paper describes online teaching modules with HTML5 J-DSP software exercises with the focus on the application of neural networks for optimizing the efficiency of utility scale solar arrays. The module exposes the students to solar array analytics, monitoring and control by simulating a system in J-DSP. The module contains software functions that apply neural networks to optimize the performance of a solar array. The module also introduces students to the basics of ML and neural networks. The simulation enables students to experiment with a neural net and attempt to reconfigure the connection topology of the solar panels in order to optimize the overall output power. The solar array simulation considered is based on an existing 104 panels / 18kW testbed at our university which is equipped with programmable relays that enable dynamic and real-time series to parallel or parallel to series connection changes. The modules and exercises are assigned as a project in our undergraduate DSP class. The project outcomes are assessed using pre and post quizzes and interviews. The final paper will provide a comprehensive description of the modules, software, exercises and assessments. This work is supported by an NSF IUSE award.
Narayanaswamy, V. S., & Shanthamallu, U. S., & Dixit, A., & Rao, S., & Ayyanar, R., & Tepedelenlioglu, C., & Spanias, A. S., & Banavar, M. K., & Katoch, S., & Pedersen, E., & Spanias, P., & Turaga, P., & Khondoker, F. (2019, June), Board 162: Online Modules to Introduce Students to Solar Array Control using Neural Nets Paper presented at 2019 ASEE Annual Conference & Exposition , Tampa, Florida. https://peer.asee.org/31906
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2019 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015