基于XGBoost方法的调距桨液压系统台架故障诊断  

Fault Diagnosis of Controlled Pitch Propeller Hydraulic System Test Bench Based on XGBoost Method

作　　者：孙林凯 娄康 SUN Linkai;LOU Kang(The 3rd Military Representative Office of the Naval Equipment Department in Shanghai Area,Shanghai 200031,China;Shanghai Marine Equipment Research Institute,Shanghai 200031,China)

机构地区：[1]海军装备部驻上海地区第三军事代表室,上海200031 [2]上海船舶设备研究所,上海200031

出　　处：《船舶工程》2024年第S02期190-196,共7页Ship Engineering

摘　　要：调距桨液压系统是调距桨装置中的关键核心部件,其相对故障率高、故障随机性大、结构组件复杂、故障影响恶劣,因此,有必要对液压系统的故障诊断展开研究。基于某调距桨台架模拟4种常见的故障类型,采集相关测点数据获得正常工况以及故障数据集。使用XGBoost对台架监测数据进行故障分类,为了提高模型的泛化能力并防止过拟合,使用添加高斯噪声、放缩变换和特征旋转对提取的特征进行数据增强,使用该方法后的故障分类精度达到98.1%,并对比不同特征对故障诊断性能的影响。研究结果可为实船加装测点提供理论支持。The hydraulic system of a controllable pitch propeller(CPP)is a critical and core component within the CPP assembly,characterized by a relatively high failure rate,significant randomness in faults,complex structural components,and severe fault impact.Therefore,it is essential to conduct fault diagnosis research on the hydraulic system.This study simulates four common fault types using a CPP test bench,collecting data from relevant measurement points to obtain both normal and fault datasets.XGBoost was employed to classify faults based on the monitoring data from the test bench.To enhance the model's generalization capability and prevent overfitting,data augmentation techniques such as Gaussian noise addition,scaling transformation,and feature rotation were applied to the extracted features.The proposed method achieved a fault classification accuracy of 98.1%,and by comparing the fault diagnosis performance across different features,theoretical support is provided for the installation of measurement points on actual vessels.

关 键 词：调距桨装置 液压系统 故障诊断 XGBoost 

分 类 号：U665.26[交通运输工程—船舶及航道工程]