Optimization and Deployment of MOVES-Matrix 3.0.3 for High-Performance On-Road Energy and Emission Rate Modeling Applications of MOVES3

This project is an extension of the previous project to develop MOVES-Matrix 3.0 for EPA’s latest version of MOVES (also known as MOVES3). The MOVES-Matrix system developed by the Georgia Tech team can be used to conduct the emissions modeling 200-times faster than using MOVES, using a multidimensional array of MOVES outputs. The research team ran MOVES 2014 146,853 times for Atlanta, GA, and 30,429 times for other regions (across all combinations of input variables) to populate matrices that can be queried for energy and emissions analyses. The lookup system provides the exact same emission results as using MOVES 2014 for each analysis, ensuring the validity of modeling for regulatory analysis. The resulting emission rate matrix allow users to link emission rates to activity data for large projects and to evaluate changes in emissions for dynamic transportation systems in near-real-time, without having to launch MOVES, because all potential MOVES runs for the region have been pre-processed. In developing MOVES-matrix 3.0, the research team updated the programs to account for changes in data structures and source sub-types and applied the same conceptual design used in MOVES-Matrix 2.0. The development required significant code modifications to integrate new source type physics parameters (vehicle specific power/scaled-tractive power configurations for various source types based on model year), to address changes in input parameters, and to perform additional mandatory iterations within model runs. The previous project that focused on developing MOVES-Matrix 3.0 (based on MOVES version 3.0.2, published by the USEPA in September 2021) and running the updated model required significant additional time to perform each iterative run.

This project proposes to extend the development efforts to further optimize model runs on the supercomputing cluster to utilize all available computational resources and to accommodate additional changes in the latest version of MOVES3 (version 3.0.3) which is designed to support Java 17. The previous effort verified the conceptual design of MOVES-Matrix to apply to MOVES3. In this proposed project, the team will finish the development of MOVES-Matrix 3.0 (and complete the model runs on the supercomputing cluster) for the Atlanta Metropolitan Area within a new optimized modeling framework (requiring less manual surveillance and better utilization of the computational resources to reduce computer run time). This proposed project will also expand matrix generation to other regions in the nation based upon a thorough review of the updated region specifications (i.e., combined configurations of fuel specifications and I/M programs). The team will also perform regional and corridor level case studies of complex transportation systems and verify that the resulting outputs obtain the exact same results as using MOVES3 directly, as we did with the existing MOVES (2014)-Matrix 2.0. All model results will be uploaded to the Department of Energy’s Open Energy Data Initiative (OEDI) servers for public access. MOVES-Matrix 3.0 will allow users to rapidly query emissions and energy rates to produce the exact same results with MOVES3 (version 3.0.3).

Research Area