首页|Studies from Los Alamos National Laboratory Yield New Data on Machine Learning ( First-principles Performance Prediction of High Explosives Enabled By Machine Le arning)
Studies from Los Alamos National Laboratory Yield New Data on Machine Learning ( First-principles Performance Prediction of High Explosives Enabled By Machine Le arning)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators discuss new findings in Machine Learning. According to news reporting out of Los Alamos, New Mexico, by NewsRx editors, research stated, "Accurate modeling of the behavior of high-expl osive (HE) materials requires knowledge of the equation of state (EOS) for both the reactant and the product states of the material. Historically, EOS models ha ve been calibrated to reproduce experimental data, but there is growing interest in first-principles predictions of HE behavior." Funders for this research include Office of Defense Programs, United States Depa rtment of Energy (DOE), National Nuclear Security Administration of U.S. Departm ent of Energy. Our news journalists obtained a quote from the research from Los Alamos National Laboratory, "The product state is particularly challenging to model because of the wide range of density and temperature conditions that are relevant as well a s the requirement to include chemical reactivity in any kind of atomistic simula tion. Density functional theory (DFT) simulations are a natural choice for such simulations, but computational cost remains a challenge to the direct applicatio n of DFT simulations to HE product EOS development. We recently introduced a mac hine-learning-driven methodology to address these challenges that was successful ly applied to a single type of HE (penta-erythritol-tetranitrate, or PETN), but there were several open questions about the generality of the approach. In parti cular, we had to develop an approximate scheme to correct the DFT energies of th e product state using the energy differences between coupled cluster theory and DFT calculations for relevant molecular species in the gas phase to achieve good agreement with experiment. In this work, we apply the method to two additional HEs (octogen and 3,3(‘)-diamino-4,4(‘)-azoxyfurazan) to address these outstandin g questions. In this work, we again find deficiencies in the DFT energetic descr iption of the product state."
Los AlamosNew MexicoUnited StatesN orth and Central AmericaCyborgsEmerging TechnologiesMachine LearningLos Alamos National Laboratory
NETL
NSTL
