Abstract:As a critical high-speed rotating machinery, the centrifugal compressor has been widely used in various modern industries. However, it is subject to a potential damaging phenomenon called surge that can be caused by several factors such as unmatched design, improper operation, inlet and outlet blockage, and so on, which thus may result in catastrophic accidents. This paper concerns the incipient surge detection and diagnosis (ISDD) of the centrifugal compressor based on its bearing vibration signals, leveraging adaptive feature fusion and sparse ensemble learning approach to develop a data-driven-based intelligent diagnostic model. Firstly, vibration signals are decomposed by the empirical mode decomposition (EMD) method, and various features can be extracted from the obtained intrinsic mode functions (IMFs) with rich information of the time series. Next, maximum likelihood estimation (MLE) is utilized to compute the intrinsic dimension of the extracted feature vectors, which are then reduced to the corresponding dimensionality adaptively by kernel principal component analysis (kernel PCA). Coping with the multi-dimensional input and nonlinearly separable classification problem, an ensemble learning approach with L1 -regularization term is adopted for incipient surge identification of the centrifugal compressor. A novel and hyperparameter-relaxed hybrid algorithm is employed for the sparse ensemble learning procedure, which guarantees transcendental accuracy and preeminent generalization capability while ensuring high robustness and sparsity of the intelligent model. Finally, in addition to diagnostic tasks of the centrifugal compressor datasets under different working conditions, the ISDD performance of the proposed method in imbalanced sample scenarios, varying working conditions and additional noise interference circumstances is comprehensively investigated, and the overall success indicates the robustness and superiority of the proposed method compared with other existing approaches.

Compressor Surge Detecting Based on Time Series Analysis

Compressor Surge Detection Based on Online Learning

Compressor Surge Detecting Based on Short-Time Energy

A Review on Theoretical and Numerical Research of Axial Compressor Surge

Surge Process of a High-Speed Axial–Centrifugal Compressor

Compressor Stability at Pulsating Condition Via Gas Dynamics Responses Analysis

Instability Early Warning Based on Frequency-Domain Characteristic of Surge

Methods of surge point judgment for compressor experiments

Analyzing the Acoustic Signal of Compressor Surge by Using Fast Fourier Transform and Wavelet Transform

Rapid diagnosis and active disturbance rejection control of compressor surge based on hybrid deep learning

Identification of Surge Inception in Aero-Engine Compressor Using S-transform

Surge Monitoring Method Based on Nonlinear Manifold Learning

Robust adaptive backstepping active control of compressor surge based on wavelet neural network

Identification of Stall and Surge in Centrifugal Compressor Experiments

Experimental Research on Surge Characteristics of Centrifugal Compressors

Numerical model of predicting surge boundaries in high-speed centrifugal compressors

Active Surge Control for Variable Speed Axial Compressors.

Experimental Analysis of Surge-Detection System Based on Pressure Derivatives at Part-Speed Operation

A novel fault measure and early warning system for air compressor

Vibration-based incipient surge detection and diagnosis of the centrifugal compressor using adaptive feature fusion and sparse ensemble learning approach

A Study of Calculational Model for Predicting Surge Line in a Centrifugal Compression System with Cooler and Condenser