Reproduce “Deep Learning Surrogates for Real-Time Gas Emission Inversion” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Thomas Newman, Christopher Nemeth, Matthew Jones, Philip Jonathan
The authors use a neural network trained on computational fluid dynamics simulations to predict methane concentrations within a sequential particle filter estimating source locations and emission rates from sensor measurements. During a ten-minute controlled release at Chilbolton, final particles averaged 5.82 metres from the known source, versus 11.09 metres for a Gaussian-plume filter; total computation took 83.3 minutes, not real-time.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.