针对船舶压载水系统故障样本不均衡且故障诊断精度较低，传统鲸鱼优化算法（WOA）易陷入局部最优、全局搜索能力不足等问题，提出一种基于自适应合成过采样方法（ADASYN）的故障数据均衡化方法，结合改进鲸鱼优化算法（IWOA）优化极限学习机（ELM）的船舶压载水系统故障诊断模型。首先，对不平衡故障数据集采用随机森林算法（RF）进行特征重要性排序并对特征参数进行降维；其次，使用ADASYN对故障样本进行自适应过采样以平衡数据集；最后，通过引入最优邻域扰动、自适应权重、变螺旋位置更新等策略对WOA进行有效改进，并对ELM的权重和阈值进行优化，并利用优化后的ELM模型对故障数据进行诊断识别。实验结果表明，对少数类故障样本进行ADASYN扩充后，IWOA-ELM模型的故障诊断准确率为96.22%，与GWO-ELM、PSO-ELM、WOA-ELM模型相比，诊断精度分别提高了2.89%、3.44%和1.22%。
Aiming at the problems of imbalanced fault samples and low fault diagnosis accuracy in ship ballast water systems, traditional Whale Optimization Algorithm (WOA) is prone to getting stuck in local optima and insufficient global search ability. A fault data balancing method based on Adaptive Synthesis oversampling (ADASYN) is proposed, combined with an improved Whale Optimization Algorithm (IWOA) to optimize the Extreme Learning Machine (ELM) for ship ballast water system fault diagnosis model. Firstly, the random forest algorithm (RF) is used to rank the feature importance of the imbalanced fault dataset and reduce the dimensionality of the feature parameters; Secondly, ADASYN is used to adaptively oversample fault samples to balance the dataset; Finally, effective improvements were made to WOA by introducing optimal neighborhood perturbations, adaptive weights, and variable spiral position updates. The weights and thresholds of ELM were optimized, and the optimized ELM model was used to diagnose and identify fault data. The experimental results show that after ADASYN expansion on minority fault samples, the fault diagnosis accuracy of the IWOA-ELM model is 96.22%. Compared with the GWO-ELM, PSO-ELM, and WOA-ELM models, the diagnostic accuracy is improved by 2.89%, 3.44%, and 1.22%, respectively.
2025,47(14): 74-81 收稿日期：2024-10-15
DOI：10.3404/j.issn.1672-7649.2025.14.012
分类号：U664.8
基金项目：国家重点研发计划项目（2022YFB4301400）；LNG船液货与机电模拟演练系统研制项目（CBG3N21-3-3）
作者简介：郭骞(1999-),男,硕士研究生,研究方向为轮机自动化与智能化
参考文献：
[1] 黄英双. 基于优化支持向量机的压载水系统故障诊断研究[D]. 大连: 大连海事大学, 2020.
[2] 李伟真, 商蕾, 汪敏, 等. 基于不平衡数据与集成学习的柴油机故障诊断研究[J]. 武汉理工大学学报(交通科学与工程版), 2024, 48(4): 661-667.
LI W Z, SHANG L, WANG M, et al. Research on diesel engine fault diagnosis based on unbalanced data and ensemble learning [J]. Journal of Wuhan University of Technology (Transportation Science and Engineering Edition), 2024, 48 (4): 661-667.
[3] 王泷德, 曹辉, 魏来. 不平衡数据下船舶主机在线故障诊断研究[J]. 中国舰船研究, 2023, 18(5): 269-275.
WANG L D, CAO H, WEI L. Research on online fault diagnosis of ship main engine under unbalanced data [J]. China Shipbuilding Research, 2023, 18 (5): 269-275.
[4] 谢桦, 陈俊星, 赵宇明, 等. 基于SMOTE和决策树算法的电力变压器状态评估知识获取方法[J]. 电力自动化设备, 2020, 40(2): 137-142+1.
XIE H, CHEN J X, ZHAO Y M, et al. Knowledge acquisition method for power transformer state assessment based on SMOTE and decision tree algorithm [J]. Power Automation Equipment, 2020, 40 (2): 137-142+1.
[5] 张涛, 王朝阳, 吴鑫辉, 等. 基于多维特征与IGWO-SVM的电机轴承故障诊断[J]. 兵器装备工程学报, 2023, 44(9): 149-154+210.
ZHANG T, WANG C Y, WU X H, et al. Fault diagnosis of motor bearings based on multidimensional features and IGWS-SVM [J]. Journal of Weapon Equipment Engineering, 2023, 44(9): 149-154+210.
[6] 李琨, 张久亭. 基于TSMAAPE与WOA-KELM的液压泵故障诊断[J]. 机床与液压, 2022, 50(9): 201-209.
LI K, ZHANG J T. Hydraulic pump fault diagnosis based on TSMAAPE and WOA-KELM [J]. Machine Tool and Hydraulic, 2022, 50 (9): 201-209.
[7] 刘迪迪, 王洋, 刘辉乾, 等. 基于ADASYN平衡化数据集的POA-SVM变压器故障诊断[J]. 电网与清洁能源, 2023, 39(8): 36-44.
LIU D D, WANG Y, LIU H Q, et al. POA-SVM transformer fault diagnosis based on ADASYN balanced dataset [J]. Power Grid and Clean Energy, 2023, 39 (8): 36-44.
[8] ADNAN R M, MOSTAFA R R, KISI O, et al. Improving streamflow prediction using a new hybrid ELM model combined with hybrid particle swarm optimization and grey wolf optimization [J]. Knowledge-Based Systems, 2021, 230.
[9] 李梦瑶, 周强, 于忠清. 基于KPCA和优化ELM的齿轮箱故障诊断[J]. 组合机床与自动化加工技术, 2021, (4): 87-90+95.
LI M Y, ZHOU Q, YU Z Q. Fault diagnosis of gearbox based on KPCA and optimized ELM [J]. Combination Machine Tool and Automation Processing Technology, 2021, (4): 87-90+95.
[10] 许德刚, 王再庆, 郭奕欣, 等. 鲸鱼优化算法研究综述[J]. 计算机应用研究, 2023, 40(2): 328-336.
XU D G, WANG Z Q, GUO Y X, et al. A Review of Whale Optimization Algorithm Research [J]. Computer Application Research, 2023, 40 (2): 328-336.
[11] Yao F, Liu Q, Ji B, et al. Open circuit fault diagnosis of three‐phase inverter based on SR‐WOA‐ELM[J]. International Journal of Circuit Theory and Applications, 2023, 52(6): 2786-2802.
[12] 刘磊, 白克强, 但志宏, 等. 一种全局搜索策略的鲸鱼优化算法[J]. 小型微型计算机系统, 2020, 41(9): 1820-1825.
LIU L, BAI K Q, DAN Z H, et al. A whale optimization algorithm with global search strategy [J]. Small Microcomputer System, 2020, 41 (9): 1820-1825.
[13] 史俊冰, 赵如意, 王迎敏, 等. 基于变量优化和IWOA-LSTM的锅炉系统水冷壁温度预测[J]. 热能动力工程, 2023, 38(10): 103-112.
SHI J B, ZHAO R Y, WANG Y M, et al. Prediction of water-cooled wall temperature in boiler system based on variable optimization and IWOA-LSTM [J]. Thermal Power Engineering, 2023, 38 (10): 103-112.
[14] 杨雨亭. 基于RF特征优选的ISSA-SVM变压器故障诊断方法[D]. 南京: 南京邮电大学, 2023.
