Dec. 10, 2020 — Artificial intelligence (AI) refers to algorithms that can — for a given set of human-defined objectives — learn, predict and make decisions, significantly increasing the speed and efficacy of decision-making. Most AI applications use machine learning (ML) to find patterns in massive amounts of data. The patterns are then used for making predictions. AI and ML have factored prominently in the National Energy Technology Laboratory’s (NETL) computational science and engineering (CSE) work in 2020 through the development of science-based simulation models, mathematical methods and algorithms and software tools required to address the technical barriers to the development of next-generation technologies. This research helps to generate information and understanding beyond the reach of experiments alone, saving time, money and materials.
NETL accomplished many CSE successes in 2020. For instance, A groundbreaking NETL study demonstrated that ML and data analytics can be used to design next-generation alloys needed to operate fossil fuel-based power plants with greater efficiency and produce affordable electricity while lowering emissions of greenhouse gas. Completed by a team at NETL’s facility in Albany, Oregon, an internationally recognized center of excellence for alloy fabrication, the study validated the application of ML analysis to enable more rapid and exceptionally accurate design of high entropy alloys (HEAs) — critical materials for ultra-efficient power generation — and eliminate the trial-and-error method and other models to develop these advanced materials.
AI and ML techniques also factored into a collaborative effort to help predict well productivity in the oil and gas industries. In these industries, obtaining accurate predictions presents difficulties because the resources are underground. Even the best estimates come with significant unknowns especially considering that no two wells are the same depending on their location, production time, design and the resource extraction technology employed. AI and ML algorithms can help overcome these challenges by making sense of multiple variables and associated data sets. Taking all these variables into account, they can calculate a well’s estimated ultimate recovery and predict its performance prior to drilling. These predictions can give producers more certainty when deciding where and where not to drill.
In 2020, the Lab also looked to incorporate powerful tools like Google’s TensorFlow, which is now revolutionizing the way NETL researchers write CFD code to accelerate the design of more energy-efficient systems. TensorFlow is a machine learning framework that takes care of memory management, communication, data operations, optimization and parallelization, so researchers can focus on the algorithm. It eliminates rewriting code for specific hardware and acts as a practical and easy on-ramp for NETL code to run on the best hardware available.
NETL CSE work extended beyond in-house research alone in 2020. For example, NETL’s expertise was leveraged in external partnerships such as a collaboration with the Colorado School of Mines (CSM) to develop AI-enabled robots capable of evaluating and repairing power plant boilers, ensuring safer and more affordable energy production.
These accomplishments are a glimpse into the work the lab is doing through its Computational Science and Engineering competency to help solve the nation’s toughest energy challenges.