English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Efficient modeling of jet diffusion during accidental release is critical foroperation and maintenance management of hydrogen facilities. Deep learning hasproven effective for concentration prediction in gas jet diffusion scenarios.Nonetheless, its reliance on extensive simulations as training data and itspotential disregard for physical laws limit its applicability to unseenaccidental scenarios. Recently, physics-informed neural networks (PINNs) haveemerged to reconstruct spatial information by using data fromsparsely-distributed sensors which are easily collected in real-worldapplications. However, prevailing approaches use the fully-connected neuralnetwork as the backbone without considering the spatial dependency of sensordata, which reduces the accuracy of concentration prediction. This studyintroduces the physics-informed graph deep learning approach (Physic_GNN) forefficient and accurate hydrogen jet diffusion prediction by usingsparsely-distributed sensor data. Graph neural network (GNN) is used to modelthe spatial dependency of such sensor data by using graph nodes at whichgoverning equations describing the physical law of hydrogen jet diffusion areimmediately solved. The computed residuals are then applied to constrain thetraining process. Public experimental data of hydrogen jet is used to comparethe accuracy and efficiency between our proposed approach Physic_GNN andstate-of-the-art PINN. The results demonstrate our Physic_GNN exhibits higheraccuracy and physical consistency of centerline concentration prediction givensparse concentration compared to PINN and more efficient compared to OpenFOAM.The proposed approach enables accurate and robust real-time spatial consequencereconstruction and underlying physical mechanisms analysis by using sparsesensor data.

Hydrogen jet diffusion modeling by using physics-informed graph neural network and sparsely-distributed sensor data