Method for predicting tunneling performance of cantilever tunneling machine based on deep belief network

The invention provides a method for predicting the tunneling performance of a cantilever tunneling machine based on a deep belief network. The method comprises the following steps: collecting tunneling parameters of the cantilever tunneling machine in a tunneling state and rock mass state parameters of a tunneling face; carrying out preprocessing on the obtained data; constructing a hard rock cantilever heading machine performance prediction database, and dividing the database into a training subset, a verification subset and a test subset; establishing a deep belief network model; predicting the tunneling performance of the cantilever tunneling machine and evaluating the result; new parameters of the rock mass in front of the tunnel face are collected and input into the trained network, and the instantaneous tunneling speed of the cantilever tunneling machine with the tunneling section in front of the tunnel face is obtained. The continuous and accurate prediction of the tunneling performance of the cantilev