Pressure chamber fault diagnosis model design based on segmented control and adaptive fuzzy neural network

Abstract

Abstract With the advent of the information age, the evolution of aerospace technology has rendered high-altitude flights increasingly common and vital. Nonetheless, the fault diagnosis of the pressure chamber, a crucial aspect of ensuring flight safety, remains an urgent challenge. The integration of segmented control technology in this domain further augments system stability and safety. This paper introduces a fault diagnosis model using EWTLM-FNN framework for monitoring and analyzing the state of the pressure chamber. The EWTLM-FNN framework commences with denoising and filtering of barometric pressure monitoring data to eliminate noise interference, followed by the extraction of frequency-domain modal information using the empirical wavelet transform (EWT). Subsequently, a three-layer Long Short-Term Memory Network conducts a profound analysis of the time and frequency domain features. The extracted features are then input into a fuzzy neural network (FNN) for fault identification and diagnosis, thus achieving high-precision monitoring of pressure chamber faults. Experimental results demonstrate that the proposed EWTLM-FNN framework exhibits superior fault diagnosis performance across multiple barometric pressure monitoring datasets, achieving over 90% diagnostic accuracy on the self-constructed pressure chamber fault dataset, and surpassing all indices compared to traditional machine learning and single deep learning models, thereby providing a theoretical and methodological foundation for future aircraft pressure fault diagnosis.

Keywords

• FNN
• Segmented control
• Fault detection
• EWT