BCSSA-VMD and ICOA-ELM based fault diagnosis method for analogue circuits

• Published in: Analog integrated circuits and signal processing Vol. 123; no. 2; p. 20
• Main Authors: You, Dazhang, Liu, Shan, Yuan, Ye, Zhang, Yepeng
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.05.2025; Springer Nature B.V
• Subjects: Accuracy; Algorithms; Analog circuits; Bandpass filters; Bandwidths; Circuits; Circuits and Systems; Classification; Data processing; Decision trees; Decomposition; Electrical Engineering; Electronic devices; Engineering; Entropy; Fault diagnosis; Feature extraction; Fourier transforms; Integrated circuits; Machine learning; Neural networks; Optimization; Pattern recognition; Search algorithms; Signal,Image and Speech Processing; Support vector machines; Wavelet transforms; Fault diagnosis; Crayfish algorithm; Sparrow algorithm; Analogue circuits; Extreme learning machine; VMD algorithm
• ISSN: 0925-1030; 1573-1979
• DOI: 10.1007/s10470-025-02360-w

Analog circuits are an important component of integrated circuit systems, and circuit systems are the foundation for ensuring the normal operation of electronic devices. Therefore, it is necessary to efficiently diagnose and maintain faults in analog circuits. However, due to the tolerance, high nonlinearity, and susceptibility to environmental interference of analog circuit components, the development of related research on fault diagnosis has been hindered, and it cannot meet the current practical requirements for high safety and reliability of electronic devices. With the continuous increase in circuit scale and integration level, how to effectively and as much as possible extract more discriminative fault features is the key research direction of analog circuit fault diagnosis. Therefore, this article proposes a variational model decomposition (VMD) feature extraction method that combines Butterfly and Cauchy Sparrow search algorithms (BCSSA) and relies on an improved crayfish optimization algorithm (COA) to optimize the Extreme Learning Machine (ELM). Decomposition (VMD) feature extraction method, and rely on Improved Crayfish Optimization Algorithm (COA) Optimized Extreme Learning Machine (ELM) to complete the classification of faults. Firstly, the BCSSA algorithm is used to optimize the number of VMD decomposition modes K and the penalty factor α to achieve the optimal VMD decomposition of the original fault signal, obtain a series of Intrinsic Mode Function (IMF) and calculate its envelope entropy, determine the optimal IMF component by selecting the IMF component with the lowest envelope entropy., and calculate its time-domain parameter, then normalize and reduce the dimensionality to construct the vector that contains the characteristics of the fault. The normalized dimensionality reduction process constitutes the fault feature vector; secondly, the ICOA algorithm is introduced to optimize the ELM; Ultimately, the fault feature vector is fed into the ELM to acquire the fault diagnosis results. The simulation test examples of the Sallen-Key bandpass filter circuit and the Four-op-amp circuit show that the accuracy of the proposed improved VMD and ELM fault diagnosis method is as high as 99.68%.