Fault diagnosis of marine electric thruster gearbox based on MPDCNN under strong noisy environments

Table 1. Experimental data sets
View table
View in Article
Table 1. Experimental data sets
齿轮状态转速/（r·min⁻¹）样本量分类标签训练集验证集测试集
正常g0 300 500 0 300 100 100
缺齿g1 300 500 1 300 100 100
表面磨损g2 300 500 2 300 100 100
齿根裂纹g3 300 500 3 300 100 100

Table 2. Architecture parameters of MPDCNN network

View table

View in Article

Table 2. Architecture parameters of MPDCNN network

模型结构	网络层	参数
一维卷积	第1个卷积层	卷积核大小：15×15；个数：64；步幅：2
	第1个最大池化层	池化大小：2×2
	第2个卷积层	卷积核大小：5×5；个数：32；步幅：1
	第2个自适应最大池化层	池化大小：2×2
二维卷积	第1个卷积层	卷积核大小：3×3；个数：64；步幅：2
	第1个最大池化层	池化大小：2×2
	第2个卷积层	卷积核大小：3×3；个数：32；步幅：1
	第2个最大池化层	池化大小：2×2
其他参数		批大小：64；最大迭代次数：100；优化器：Adam；学习率：0.000 1；权重：0.001；参数省略比例：0.2

Table 3. Comparison results of other feature extraction methods with the method proposed in this paper
View table
View in Article
Table 3. Comparison results of other feature extraction methods with the method proposed in this paper
特征提取方法准确率/% 训练时间/s
EMD-PDCNN 78.50 41.290
VMD-PDCNN 65.25 41.286
CWT-PDCNN 90.75 33.825
EWT-PDCNN 83.50 37.682
本文方法 98.00 34.918

Table 4. Comparison results of other fault diagnosis methods wiht the method proposed in this paper
View table
View in Article
Table 4. Comparison results of other fault diagnosis methods wiht the method proposed in this paper
故障诊断方法准确率/% 训练时间/s
SVM 93.75 8.600
KNN 85.75 8.753
Naive Bayes 88.75 4.023
Random Forest 88.00 3.819
AlexNet 95.75 37.561
Resnet 97.25 58.364
VGG 97.50 51.227
本文方法 98.00 34.918

Tools

Get Citation

Copy Citation Text

Qianming SHANG, Wanying JIANG, Yi ZHOU, Zhengqiang WANG, Yubo SUN. Fault diagnosis of marine electric thruster gearbox based on MPDCNN under strong noisy environments[J]. Chinese Journal of Ship Research, 2025, 20(2): 30

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Ship Intelligent O&M, and Fault Diagnosis

Received: Mar. 8, 2024

Accepted: --

Published Online: May. 15, 2025

The Author Email:

DOI:10.19693/j.issn.1673-3185.03814

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Experimental data sets

Table 1. Experimental data sets

Table 2. Architecture parameters of MPDCNN network

Table 2. Architecture parameters of MPDCNN network

Table 3. Comparison results of other feature extraction methods with the method proposed in this paper

Table 3. Comparison results of other feature extraction methods with the method proposed in this paper

Table 4. Comparison results of other fault diagnosis methods wiht the method proposed in this paper

Table 4. Comparison results of other fault diagnosis methods wiht the method proposed in this paper