Figures & data
Figure 1. Overview of the framework of the proposed SSDL.
![Figure 1. Overview of the framework of the proposed SSDL.](/cms/asset/c2f49d29-327f-4559-bdbd-4164e696a499/tprs_a_2032860_f0001_oc.jpg)
Figure 2. Structure of SAE and SSAE-Softmax neural network: (a) symmetrical structure neural network of SAE, where is the dimension of each sample,
is the number of neurons in the hidden layer 1; and (b) structure of SSAE-Softmax classifier, where
and
are the number of neurons in the hidden layer 2 and hidden layer 3, respectively,
denotes the probability that
is classified as operational state
(
).
![Figure 2. Structure of SAE and SSAE-Softmax neural network: (a) symmetrical structure neural network of SAE, where x is the dimension of each sample, g is the number of neurons in the hidden layer 1; and (b) structure of SSAE-Softmax classifier, where d and q are the number of neurons in the hidden layer 2 and hidden layer 3, respectively, P(h=i|x) denotes the probability that s is classified as operational state i (i=1,2,…,y).](/cms/asset/03273f6a-19f2-41bd-b5f1-2c66385c1a2a/tprs_a_2032860_f0002_oc.jpg)
Figure 3. Flowchart of constructing SSDL model for machinery fault diagnosis.
![Figure 3. Flowchart of constructing SSDL model for machinery fault diagnosis.](/cms/asset/eb4fa237-79a9-4fae-9dc6-0d29cc6e08db/tprs_a_2032860_f0003_oc.jpg)
Table 1. Description of the four datasets for bearing fault diagnosis.
Figure 4. Experimental data preparation for delta 3-D printer: (a) test-rig; (b) installed attitude sensor; (c) normal and fault state of a synchronous belt; and (d) normal and fault state of a joint bearing screw.
![Figure 4. Experimental data preparation for delta 3-D printer: (a) test-rig; (b) installed attitude sensor; (c) normal and fault state of a synchronous belt; and (d) normal and fault state of a joint bearing screw.](/cms/asset/4f9ac43b-0779-442f-86b1-da03720285d0/tprs_a_2032860_f0004_oc.jpg)
Table 2. Description of the four datasets for 3-D printer fault diagnosis.
Table 3. Diagnosis accuracies of SSDL under different combinations of network architecture and enlarging factor.
Table 4. Hyper-parameter settings for all the approaches on bearing datasets.
Table 5. Hyper-parameter settings for all the approaches on 3-D printer datasets.
Figure 5. The prediction accuracies obtained by all the algorithms on each testing dataset.
![Figure 5. The prediction accuracies obtained by all the algorithms on each testing dataset.](/cms/asset/4445920d-e965-4961-b604-76ff86d0a37e/tprs_a_2032860_f0005_oc.jpg)
Figure 6. The iterative Accu and Accs values of the contrastive approaches on the BD-1: (a) SSDL-SS; (b) SSDL-SC; (c) SSDL-ES; and (d) SSDL.
![Figure 6. The iterative Accu and Accs values of the contrastive approaches on the BD-1: (a) SSDL-SS; (b) SSDL-SC; (c) SSDL-ES; and (d) SSDL.](/cms/asset/82913873-ab47-4d15-bc90-001849d14813/tprs_a_2032860_f0006_oc.jpg)
Figure 7. The iterative Accu and Accs values of the contrastive approaches on the BD-4: (a) SSDL-SS; (b) SSDL-SC; (c) SSDL-ES; and (d) SSDL.
![Figure 7. The iterative Accu and Accs values of the contrastive approaches on the BD-4: (a) SSDL-SS; (b) SSDL-SC; (c) SSDL-ES; and (d) SSDL.](/cms/asset/c140663c-d929-4207-b6d0-8698d35a01b2/tprs_a_2032860_f0007_oc.jpg)
Figure 8. The prediction accuracies obtained by SSDL under different number of labelled samples on datasets PD-3 and PD-4.
![Figure 8. The prediction accuracies obtained by SSDL under different number of labelled samples on datasets PD-3 and PD-4.](/cms/asset/f55368b9-d03b-4ccb-8e59-b044f6673b3f/tprs_a_2032860_f0008_oc.jpg)
Table 6. Comprehensive comparison of diagnosis accuracy.