Method for identifying working state of aero-engine

The invention discloses an aero-engine working state identification method, which comprises the following steps: S1, data acquisition: acquiring one or more sortie data of an aero-engine, and extracting parameter record data such as rotating speed, exhaust temperature, throttle lever angle, exhaust nozzle area and guide vane angle from the sortie data and variable quantities of a plurality of time steps of the record data; and S2, model construction: constructing a multi-scale convolutional neural network-bidirectional long-short term memory neural network model, importing input data, and establishing a model. The aero-engine working states including the transition state can be recognized, the working states are divided meticulously, and the model recognition accuracy is improved; the transition state can be better identified; state division is more careful, transition working states such as starting, acceleration, deceleration, connection and disconnection of stress application, accelerator stopping and the