This study proposes an intelligent scheduling factor-incorporated optimisation method for energy flow dispatch in integrated energy systems. The methodology innovatively integrates CNN-based high-precision load forecasting with D–S evidence theory. This approach fuses the predicted load values (Evidence E1) with production-side operational constraints (e.g.

oil production plan

load rate

E2 and E3)

establishing a robust

dynamic scheduling decision framework. Initially

by determining equipment parameters and control factor states under the surplus electricity-to-hydrogen mode

a heterogeneous flow coupling model incorporating scheduling factors is developed for the energy supply system. Subsequently

an intelligent scheduling factor prediction framework is proposed: A CNN-based load forecasting model is constructed to obtain preliminary predictions

followed by correlation analysis to select multisource evidence sets and establish an information fusion model for determining optimal scheduling factors with confidence intervals. Finally

the proposed method provides the optimal scheduling factor expectation as the decision command and the D–S confidence interval as a safety constraint

which is validated in case studies on a deepwater oil and gas field (DWOGF) in the South China Sea.