Fault Diagnosis Method of Analog Circuit Based on Radial Base Function Neural Network

Gebregiorgis, Meron; He, Yigang; Ning, Shuguang; Shi, Wang

doi:10.1088/1742-6596/1575/1/012148

Cited by 2 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To confirm the validity of the Att-GCN-based fault diagnosis method for analogue circuit handling of graphs, experiments have been chosen on the Sallen-Key bandpass filter circuit, the four-op-amp biquadratic filter circuit, and the amplifier board circuit, with the accuracy of the method compared to those based on DNN [7], one-dimensional convolutional neural network (1DCNN) [19], RBFNN [20], ResNet network [24], GMKL-SVM [5], MKELM [25] models for accuracy comparison.…”

Section: Experimental Results and Analysesmentioning

confidence: 99%

“…In the Sallen-Key bandpass filter circuit and four-op-amp biquadratic filter circuit, methods based on Att-GCN, DNN [7], 1DCNN [19], RBFNN [20], Resnet [24], GMKL-SVM [5], and MKELM [25] were applied for incipient fault diagnosis. Models, accuracies, training times and testing times were calculated as portrayed in tables 6-8.…”

Section: Comparison Of Methodsmentioning

confidence: 99%

“…A short-time Fourier transform (STFT) with the time window width of 2 was performed for obtaining the time-frequency image, which was then fed into resnet50 [24]. Set the spread parameter of RBFNN [20] as 1. To build the GMKL-SVM [5], db5 was chosen as the wavelet basis function for 6 layers wavelet packet transform.…”

Section: Comparison Of Methodsmentioning

confidence: 99%

“…The softplus function is an approximately smoothed version of the relu, with the advantage that it remains valid within the negative range. Finally, the softmax function is selected to normalise the results to obtain α ij for ease of comparison and computation, as shown in (20),…”

Section: Fault Extraction Modulementioning

confidence: 99%

“…Yang et al [19] proposed the 1-dimensional convolutional neural network structure. Gebregiorgis et al [20] applied a radial basis function-based neural network (RBFNN) for diagnosis and pattern classification. He et al [21] dealt with the response signal through cross wavelet transform to obtain a spectrum image, which was then fed into a generative adversarial network as a feature image for processing.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

An incipient fault diagnosis method based on Att-GCN for analogue circuits

Yang¹,

Li²,

Gao³

2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Incipient faults for analogue circuits in modern electronic systems are difficult to diagnose due to poor fault features. To address this issue, a method based on the attention weighted graph convolution network (Att-GCN) is proposed in this paper. The structural and data features of samples are jointly extracted to mine the effective characteristics from incipient faults. First, a wavelet packet energy transform and a probabilistic principal component analysis (ProbPCA) are employed to enhance the sample fault information. Then, the distance clustering method is deployed to construct the sample set into a non-European structure sample graph, where the structural features of fault samples are preserved. Second, an Att GCN, which combines the spatial-domain graph convolution network and improved self-attention mechanism, is constructed to extract the structural features and data features to obtain more effective fault information. Additionally, the multisample dropout method is introduced to reduce network overfitting in the training process. To assess the method's actual performance for fault diagnosis, experiments are carried out in the Sallen-Key bandpass filter circuit, the four-op-amp biquadratic filter circuit and the amplifier board circuit. The outcomes indicate that this method improves the incipient fault diagnosis accuracy for analogue circuits.

show abstract