Aeroengine fault detection method and system based on grouped convolutional autoencoder

Abstract

The invention relates to an aeroengine fault detection method and system based on a grouped convolutional autoencoder, wherein the method includes: a variable grouping step, and based on the correlation between the variables, the variables of the aircraft communication addressing and reporting system data are divided into multiple Variable group; feature extraction step, using convolutional denoising autoencoder model to independently extract the features of each variable group; fault identification step, merging the features of all variable groups to form a feature vector, based on the feature vector using support vector machine to identify faulty samples. The invention does not require a large amount of expert knowledge and experience, avoids cumbersome data preprocessing work, and still has good comprehensive fault detection performance without a large number of good labeled samples, and has good robustness, and is suitable for engineering practice , with low computational and time costs.

Description

technical field [0001] The invention relates to the technical field of aero-engines, in particular to an aero-engine fault detection method and system based on group convolution autoencoders. Background technique [0002] With the development of civil aviation, the safety, reliability and economy of aircraft engines have received more and more attention. Engine fault detection is an important way to improve the above performance. It can help managers allocate monitoring resources more reasonably, improve the safety and reliability of each flight, formulate scientific maintenance plans, and minimize operation and maintenance costs. [0003] Many current engine fault detection methods rely on performance deviation data. However, the performance deviation data comes from the engine manufacturer (OEM), and airlines need to pay high fees to obtain the performance deviation data. If the cooperation between airlines and OEMs is interrupted due to some unpredictable reasons, it wi...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is that, aiming at the problem that the existing fault detection method is not suitable for dealing with ACARS data with complex relationships, high dimensions and containing a lot of noise, the grouping operation, convolutional neural network, autoencoder and support vector machine Combined, a method and system for aero-engine fault detection based on grouped convolutional denoising autoencoder is proposed for processing ACARS data

Images