一种往复压缩机气阀故障诊断方法
By optimizing the Hankel matrix window length of DMD using CEO and combining it with two-dimensional multi-scale attention entropy (MATE2D) with PSO to optimize LSSVM parameters, the problem of inappropriate parameter selection in existing technologies is solved, achieving high-precision diagnosis of valve faults in reciprocating compressors and improving fault identification capability and model generalization performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENYANG LIGONG UNIV
- Filing Date
- 2026-05-07
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods for diagnosing valve faults in reciprocating compressors suffer from several problems, including the lack of universality in selecting the Hankel matrix window length, incomplete extraction of one-dimensional entropy features, and the influence of LSSVM model parameters on diagnostic accuracy, leading to inaccurate fault diagnosis.
The Chaotic Evolutionary Optimization (CEO) algorithm is used to optimize the Hankel matrix window length in Dynamic Mode Decomposition (DMD), and the kernel parameters and penalty parameters of Least Squares Support Vector Machine (LSSVM) are optimized by combining the two-dimensional multi-scale attention entropy (MATE2D) and particle swarm optimization (PSO) algorithm. A PSO-LSSVM model is constructed to achieve efficient decomposition and feature extraction of valve vibration signals.
It improves the stability and universality of DMD decomposition, enhances the ability to extract time-frequency domain features of fault signals, strengthens the generalization ability of LSSVM model, realizes high-precision valve fault diagnosis, and ensures industrial production safety.
Smart Images
Figure CN122149843B_ABST