摘要
传统的动力装置故障诊断方法需要大量的故障数据样本,导致诊断效率和实时性差,无法满足现代船舶航行的需求。针对上述问题,提出人工智能技术在船舶动力装置故障诊断中的应用。使用小波包分析技术对传感器采集的信号进行去噪、分解重构以及能量谱特征提取处理后,构建船舶动力装置故障集。使用D-S理论对BP神经网络输出的诊断结果进行数据融合和置信度判断,得到可靠的诊断结果完成故障诊断。对比实验数据显示,利用人工智能的方法诊断精度较高,并且诊断响应效率高,具有良好的泛化能力。
Traditional power plant fault diagnosis methods need a large number of fault data samples, resulting in poor diagnosis efficiency and real-time, which can not meet the needs of modern ship navigation. In view of the above problems,the application of artificial intelligence technology in fault diagnosis of marine power plant is proposed. Wavelet packet analysis technology is used to denoise, decompose and reconstruct the signal collected by the sensor and extract the feature of energy spectrum, then the fault set of marine power plant is constructed. D-S theory is used for data fusion and confidence judgment of BP neural network output diagnosis results to obtain reliable diagnosis results and complete fault diagnosis.Compared with the experimental data, the method of artificial intelligence has high diagnosis accuracy, high diagnosis response efficiency and good generalization ability.
作者
马海洲
丁爱萍
MA Hai-zhou;DING Ai-ping(Yellow River Conservancy Technical Institute of College of Information Engineering,Kaifeng 475001,China)
出处
《舰船科学技术》
北大核心
2021年第12期109-111,共3页
Ship Science and Technology
基金
河南省高等教育教学改革研究与实践项目(2017SJGLX140)
关键词
人工智能技术
船舶动力装置
故障诊断
神经网络
artificial intelligence techniques
ship power units
fault diagnosis
neural networks
