Project Description

Due to ever-increasing demands for flexibility and new features in the bulk power system brought on by the rapid build of renewable generation and energy storage assets, the time performance of long-term planning models and short-term wholesale power market optimization suffers with the increased size and complexity of the optimization problems. Better-suited, but more complex problem formulations in power system planning and operation can be halted or replaced with simpler interim solutions due to computational limitations, which at times may delay the inevitable need for the ideal solution.

The student will use simplified unit commitment and economic dispatch power system modeling and the Clarkson HPC environment to explore methods to parallelize this common application to improve its performance. Additional analysis tools can be used as needed for the project. Guidance on the use of the model code, the underlying concepts in Power Engineering, and the HPC environment will be provided. The modeling tools are open source and students are encouraged to modify or add to them as needed for the project.

Upon completion of the research, students will provide insights on the methods available to improve the performance of the economic dispatch problem. These insights shall be presented to industry advisors from the New York Independent System Operator (NYISO) and Clarkson faculty and should include an explanation of the theory and testing of each method used and results in comparison to no parallelization/deconstruction and commonly available computer equipment. Students are also encouraged to document their work for potential publication.