Intelligent spaceflight liquid engine health state identification method under fault-free sample

The invention discloses an intelligent spaceflight liquid engine health state identification method under a fault-free sample, and the method comprises the steps of taking one-dimensional signals collected by a plurality of sensor channels at a steady-state operation stage of a spaceflight liquid engine as multi-source data, carrying out the unified preprocessing of all channel signals, and dividing the channel signals into a training set and a test set; the fault data generation method based on time sequence tracking comprises generating a fault sample through a training set fault-free sample, and merging the fault sample into a training set; then constructing a cyclic convolutional neural network as an intelligent state recognition model, and performing adaptive fusion and feature extraction on multi-source data of a training set by the model to realize intelligent recognition of the health state of the aerospace liquid engine; and finally, applying the model to intelligent identification of the health stat