Blocked and Accelerated Wavelet De-noising Algorithm Based on Data Splitting and Wavelet Analysis in Large Data Environment for Aero-Engine Health Monitoring

Abstract

Data de-noising is a necessary part of health management, and it is the premise and foundation of effective feature extraction, condition monitoring and fault diagnosis for aero-engine. Random noise can cause serious interference to effective signals, and even lead to signal distortion and misdiagnosis of health condition. In view of the contradiction between the limited computing power of aircraft airborne system and the large amount of data processing, an blocked wavelet de-noising algorithm for large data is proposed based on the principle of data splitting theory and the wavelet theory under the multiple constraints of large data, high de-noising precision and fast processing speed. The algorithm used data splitting principle to split large data into small data sets, reduced the computational requirements of large data, and accelerated the speed of wavelet de-noising. The processing results of the theoretical data and the actual airborne aero-engine monitoring data showed that, compared with the traditional algorithms, the algorithm can protect the effective information and maintain the same de-noising accuracy, and the data de-noising time in the aero engine health monitoring data environment was accelerated by 4 times at least. 