Prediction of H₂S Dispersion Distance in Complex Terrain Using DEM-Based Deep Neural Networks

This dataset contains the simulation results used for training and validating the deep neural network model proposed in the associated manuscript. The data were generated through Computational Fluid Dynamics (CFD) simulations under real complex terrain conditions in two pipeline scenarios (Sichuan gas field and Xinjiang oil field). Each row represents one simulation case corresponding to a specific leakage location and dispersion direction. The output variable corresponds to the predicted hazardous concentration boundary distance for hydrogen sulfide (H₂S) at 200 ppm. The dataset supports the development and validation of a DEM-based deep learning framework for rapid prediction of H₂S dispersion distance.