Nov. 5, 2021 — Three technologies developed by researchers at the U.S. Department of Energy’s (DOE) Argonne National Laboratory and partner organizations have been named as 2021 R&D 100 Award winners, building on a decades-long history of wins.

Recognizing the 100 most innovative technologies of the past year, the R&D 100 Awards are considered the ​“Oscars” of innovation. Sponsored by R&D World magazine, the renowned worldwide competition received entries from 17 countries/regions.

Started in 1963, the R&D 100 Awards serve as the nation’s most prestigious innovation awards program, honoring R&D pioneers and their revolutionary ideas in science and technology. Technologies are chosen from three categories: Mechanical/Materials, Process/Prototyping and Software/Services.

Argonne scientists have received more than 130 R&D 100 Awards since the competition began. Past winners include Fortune 500 companies, DOE national laboratories, academic institutions and smaller companies.

The Argonne technologies described below were selected as winners by an independent panel of more than 40 industry leaders. Winners were recognized at the virtual 2021 R&D 100 Conference on Oct. 21–22.

ML-GA: a Machine Learning-Genetic Algorithm for Rapid Product Design Optimization

Pinaki Pal, Opeoluwa Owoyele, Ahmed Abdul Moiz, Janardhan Kodavasal, Sibendu Som

Machine Learning — Genetic Algorithm (ML-GA) is a unique software technology that harnesses the power of advanced machine learning to speed up the virtual design of products and manufacturing processes across a wide range of industries.

Industrial product design involves a large number of control parameters that are often time-consuming and costly to optimize, even with computer modeling. By embedding ML into the design process, ML-GA dramatically speeds up computer-aided engineering simulation-driven virtual prototyping, dramatically shrinking the product development phase from a few months to a few days compared with traditional approaches. In doing so, the software reduces computational costs as well.

ML-GA operates on a novel combination of advanced ensemble ML-based surrogate modeling, adaptive sampling of design space (via active learning) for on-the-fly refinements of the ML surrogate model and a GA optimizer — all within an automated, modular, end-to-end workflow.

Owing to its highly parallelizable and portable framework, ML-GA can be readily coupled with any simulation tool and run efficiently on high performance computing (HPC) clusters/supercomputers and cloud-based platforms. These unique features allow for easy adoption by industries ranging from automotive, aerospace and defense to energy and oil and gas.

ML-GA is funded with support from DOE’s Vehicle Technologies Office (VTO) through a Technology Commercialization Fund (TCF) project. (VTO is part of the Office of Energy Efficiency and Renewable Energy [EERE].) ML-GA was recently licensed on a nonexclusive basis by industry partner Parallel Works Inc., a Chicago-based HPC software platform company, as part of the TCF project.

Mochi: a Customizable Data Navigation Tool

Robert Ross, Philip Carns, Matthieu Dorier, Robert Latham, George Amvrosiadis, Charles Cranor, Tyler Reddy, Robert Robey, Dana Robinson, Galen Shipman, Shane Snyder, Jerome Soumagne, Qing Zheng

While most scientists rely on data storage systems to gather and analyze data, many struggle to manage the data generated by their research. Mochi is a novel navigation tool that offers a solution. Rather than using a one-size-fits-all approach to data, Mochi allows scientists to rapidly customize a suite of data services to suit the needs of a specific domain and problem. By shaving weeks or months off the time needed to produce actionable information from collected data, scientists are able to realize their discoveries faster.

Composition is key to Mochi’s success. The open-source, state-of-the-art tool offers communication, data storage, concurrency management and group membership capabilities, along with a collection of building blocks scientists can use to craft a data storage system designed to address their own specific needs. Each scientist benefits from using a specialized storage service without having to create one from scratch. These specialized services offer greater efficiency and flexibility than a traditional monolithic file system and are applicable to new technologies as they emerge.

Regardless of which components are used, they all share the same underlying communication framework, known as Mercury, to efficiently move large volumes of data between storage and compute resources. Mochi is scalable, enables high performance and can help reduce the coding and maintenance burden for teams building data services.

While research continues on Mochi, the core components are widely used inside and outside of Argonne.

SZ: a Lossy Compression Framework for Scientific Data

Franck Cappello, Sheng Di, Jon Calhoun, Griffin Dube, Ali Murat Gok, Sian Jin, Xin Liang, Cody Rivera, Dingwen Tao, Jiannan Tian, Robert Underwood, Chengming Zhang, Kai Zhao

Exascale simulations and next-generation scientific instruments are important to address issues such as climate change, cosmology, materials science, advanced manufacturing and the development of new disease treatments and drugs. Many of these simulations and instruments need methods to reduce significantly the data they produce.

Developed by Argonne scientists, SZ is a lossy compression framework for scientific floating-point data featuring an innovative, highly customizable and configurable design with strict compression error control. The technology was initiated at Argonne along with university and industry collaborators. Argonne is the lead and main developer of the SZ compressor.

With its unique combination of capabilities, SZ offers an exceptionally wide scope of application use cases for scientific simulations and instrument facilities and demonstrates excellent performance in compression ratios, speed and accuracy. SZ can be used to visualize data, accelerate simulation, reduce the simulation data footprint for storage, compute larger simulation problems (compression in memory), accelerate execution by avoiding recomputation and reducing memory bandwidth bottlenecks, and reduce instrument data stream intensity.

SZ has applications in simulation (cosmology, quantum chemistry, molecular dynamics, climate), seismic imaging and X-ray crystallography. It is also used by researchers to advance the development of compression methods for scientific data.

The SZ project is supported primarily by DOE’s Exascale Computing Project, a collaborative effort of DOE’s Office of Science and the National Nuclear Security Administration.

The Office of Energy Efficiency and Renewable Energy’s (EERE) mission is to accelerate the research, development, demonstration, and deployment of technologies and solutions to equitably transition America to net-zero greenhouse gas emissions economy-wide by no later than 2050, and ensure the clean energy economy benefits all Americans, creating good paying jobs for the American people — especially workers and communities impacted by the energy transition and those historically underserved by the energy system and overburdened by pollution.

Argonne National Laboratory seeks solutions to pressing national problems in science and technology. The nation’s first national laboratory, Argonne conducts leading-edge basic and applied scientific research in virtually every scientific discipline. Argonne researchers work closely with researchers from hundreds of companies, universities, and federal, state and municipal agencies to help them solve their specific problems, advance America’s scientific leadership and prepare the nation for a better future. With employees from more than 60 nations, Argonne is managed by UChicago Argonne, LLC for the U.S. Department of Energy’s Office of Science.

The U.S. Department of Energy’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://​ener​gy​.gov/​s​c​ience.

---

Source: Beth Burmahl, Argonne National Laboratory