Electric Shock Feature Extraction Method Based on Adaptive Variational Mode Decomposition and Singular Value Decomposition

This paper proposes a feature extraction method combining adaptive variational mode decomposition (AVMD) and singular value decomposition (SVD) for electric shock fault-type identification. The AVMD algorithm is utilized to adaptively decompose the electric shock signal into intrinsic mode components, each containing distinct frequency information. Subsequently, the correlation coefficient is employed to extract the intrinsic mode component with amplitudes greater than or equal to 0.1 (gamma k${\gamma }_k$ >= 0.1). Feature extraction is then performed using SVD on the gamma k${\gamma }_k$ >= 0.1 intrinsic mode component, based on its maximum singular value and singular entropy. This approach effectively overcomes the limitation of the traditional VMD that necessitates manual K value setting. Moreover, it achieves dimensionality reduction and feature extraction of the intrinsic mode components through SVD, resulting in enhanced computational efficiency and fault identification accuracy. Extensive simulations demonstrate the remarkable recognition rates of electric shock fault types in animals and plants using the proposed AVMD-SVD method, achieving a recognition rate as high as 99.25%. Comparative performance analysis further verifies the superiority of AVMD-SVD over similar empirical mode decomposition-SVD feature extraction techniques. This paper proposes a feature extraction method combining adaptive variational mode decomposition (AVMD) and singular value decomposition for electric shock fault-type identification. The number of mode components (K) in VMD is adaptively determined through singular entropy relative increment. The optimal AVMD modal components are selected through the correlation coefficient, constructing the Hankel matrix and extracting the maximum singular value and singular entropy from the Hankel matrix as characteristic phasors for animal and plant electric shock.image