TBM tunneling footage prediction method based on multi-modal deep fusion model

Abstract

The application discloses a TBM tunneling footage prediction method based on a multi-modal deep fusion model, and belongs to the technical field of tunnels and underground engineering. The TBM tunneling footage prediction method based on the multi-modal deep fusion model comprises the following steps: step 1, collecting first multi-modal data of a TBM historical construction section, and performing data alignment and format unification processing on the first multi-modal data; step 2, constructing an initial TBM tunneling footage prediction model; step 3, performing parameter training on the initial TBM tunneling footage prediction model by using a sample set to obtain a final TBM tunneling footage prediction model; and step 4, inputting second multi-modal data into the final TBM tunneling footage prediction model to output a tunneling footage prediction value of a to-be-tunneled section. The application realizes joint modeling and deep fusion of geological environment semantics in construction documents and spatial geometric constraints in engineering drawings, converts long-term accumulated static data into calculable prediction features, and improves the accuracy and stability of TBM tunneling footage prediction.

Images