Aircraft abnormal dynamic load rapid positioning method based on deep recurrent neural network

The invention provides an aircraft abnormal dynamic load rapid positioning method based on a deep recurrent neural network, which divides a dynamic load positioning problem into a multivariable time sequence classification problem, and comprises the following steps: firstly, obtaining a vibration response of an aircraft subjected to an abnormal load through simulation or calculation, and dividing to obtain training set data, verification set data and test set data; and a corresponding dynamic load position label. Then constructing an LSTM neural network model for performing feature extraction and judgment on a time sequence signal in the input vibration response time sequence data, and outputting a corresponding dynamic load position label; and the LSTM neural network model is trained and predicted through the obtained data, so that the model can accurately position the dynamic load position. And finally, during actual application, the abnormal dynamic load actually borne by the aircraft can be quickly positi