为实现船舶调距桨螺距失控故障的精准诊断，提出基于径向基函数（Radial Basis Function，RBF）神经网络的智能诊断方法。通过AMESim仿真平台构建调距桨液压系统多工况模型，模拟液压泵吸入口堵塞、液压缸内泄漏、安全阀弹簧失效等5类典型故障，采集系统压力、流量及温度等9维特征参数构建数据集。采用Z-score标准化方法消除量纲差异，结合网格搜索算法优化RBF神经网络扩展参数，建立单隐层故障分类模型，并通过Matlab实现网络训练和验证。结果表明，该方法分类准确率达96%，与传统BP神经网络相比，诊断效率提升23%，误报率降低至3.8%，验证了该模型对非线性故障特征的强适应性和高可靠性。研究成果可为船舶机电设备智能诊断提供可推广技术方案。
This study aims to accurately diagnose pitch - runaway faults in ship controllable - pitch propellers. An intelligent diagnosis method using Radial Basis Function neural networks is proposed. First, an AMESim - based multi - condition model of the propeller's hydraulic system is built. Five typical faults are simulated, like blocked hydraulic - pump inlets and leaking cylinders. Nine - dimensional data on pressure, flow, and temperature are collected to form a dataset. Z - score standardization removes dimensional differences, and grid - search optimizes RBF network parameters. A single-hidden-layer fault classification model was developed and implemented through Matlab for network training and validation. Test results show 96% accuracy, a 23% higher efficiency than BP neural networks, and a 3.8% false - alarm rate. This proves the model's adaptability and reliability for non - linear faults, offering a practical solution for ship electromechanical equipment diagnosis.
2026,48(1): 114-119 收稿日期：2025-4-10
DOI：10.3404/j.issn.1672-7649.2026.01.016
分类号：U664.33
基金项目：广西高校中青年教师（科研）基础能力提升项目（2022KY0421）
作者简介：刘润泽(1987-),男,硕士,研究方向为船舶设备故障诊断、可靠性等
参考文献：
[1] 王磊, 李唐, 刘猛, 等. 某型船调距桨偶发螺距失控故障原因分析[J]. 广东造船, 2024, 43(1): 104-106.
WANG L, LI T, LIU M, et al. Analysis of the causes of occasional pitch runaway fault of a certain type of ship's controllable pitch propeller[J]. Guangdong Shipbuilding, 2024, 43(1): 104-106.
[2] ORHAN M, CELIK M. A literature review and future research agenda on fault detection and diagnosis studies in marine machinery systems[J]. Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment, 2024, 238(1): 3-21.
[3] CAI H, CHEN W, FU H, et al. A multimodal fault diagnosis model utilizing GAF transformation for controllable pitch propeller hydraulic system[C]//2024 10th International Conference on Control, Decision and Information Technologies (CoDIT). IEEE, 2024: 2881-2886.
[4] FENG Y, CHEN W, FU H, et al. Fault diagnosis of controllable pitch propeller as few-shot classification with mechanism simulation data augmentation[C]//2023 IEEE 2nd Industrial Electronics Society Annual On-Line Conference (ONCON). IEEE, 2023: 1-5.
[5] BAI X, LING H, LUO X, et al. Reliability and availability evaluation on hydraulic system of ship controllable pitch propeller based on evidence theory and dynamic Bayesian network[J]. Ocean Engineering, 2023, 276: 114125.
[6] 卢石松, 唐建中, 夏城城, 等. 新型建模语言在调距桨电液系统故障诊断中的应用[J]. 液压与气动, 2025, 49(1): 1-12
LU S, TANG J, XIA C, et al. Application of a novel modeling language in fault diagnosis of electro-hydraulic systems for controllable pitch propellers[J]. Hydraulics & Pneumatics, 2025, 49(1): 1-12
[7] 刘沁. 船舶舵机液压系统的智能故障诊断方法研究[D]. 北京: 北京交通大学, 2021.
[8] 孙建波. 船舶柴油主推进装置及其控制系统的建模与仿真研究[D]. 大连: 大连海事大学, 2007.
[9] 郜超见. 基于AMESim和置信推理的调距桨故障诊断研究[D]. 北京: 中国舰船研究院, 2023.
[10] 陈涛. 基于神经网络的船舶柴油机故障诊断研究[D]. 武汉: 武汉理工大学, 2014.
[11] 蓝润荣. 基于改进RBF神经网络的信用评级分析[D]. 北京: 中国科学技术大学, 2014.
[12] 董航程. 基于径向基函数的自编码器算法研究[D]. 哈尔滨: 哈尔滨工业大学, 2019.
[13] 刘学辉. 基于粒子群算法优化RBF神经网络的船用主机能效分析[D]. 天津: 天津理工大学, 2022.
