The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fau...The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.展开更多
The behavior of matrix converter(MC) drive systems under the condition of MC short-circuit faults is comprehensively investigated. Two isolation strategies using semiconductors and high speed fuses(HSFs) for MC short-...The behavior of matrix converter(MC) drive systems under the condition of MC short-circuit faults is comprehensively investigated. Two isolation strategies using semiconductors and high speed fuses(HSFs) for MC short-circuit faults are examined and their performances are compared. The behavior of MC drive systems during the fuse action time under different operating conditions is explored. The feasibility of fault-tolerant operation during the fuse action time is also studied. The basic selection laws for the HSFs and the requirements for the passive components of the MC drive system from the point view of short-circuit faults are also discussed. Simulation results are used to demonstrate the feasibility of the proposed isolation strategies.展开更多
针对实际采煤机轴承故障诊断中存在变工况特征提取困难,故障训练样本不足等问题,结合当今流行的迁移学习的方法,提出了一种多尺度迁移学习的轴承诊断方法。首先通过经验模式分解(Empirical Mode Decomposition,EMD)从振动信号中分解成...针对实际采煤机轴承故障诊断中存在变工况特征提取困难,故障训练样本不足等问题,结合当今流行的迁移学习的方法,提出了一种多尺度迁移学习的轴承诊断方法。首先通过经验模式分解(Empirical Mode Decomposition,EMD)从振动信号中分解成不同频率的本征模态函数(Intrinsic Mode Function,IMF);其次将得到的不同频率的IMF与卷积神经网络中不同尺寸卷积核提取到的丰富特征互补构建多尺度特征融合;采用联合最大平均差异(Joint Maximum Mean Discrep⁃ancy,JMMD)特征迁移的方法使源域与目标域联合分布差异最小化,然后通过多尺度融合模型进行分类识别;最后在凯斯西储大学轴承数据集和江南大学数据集对该方法进行了验证。实验结果证明该模型在两种不同工况和型号的轴承数据集中均取得较高的准确率,表现出模型良好的泛化能力。展开更多
为了解决滚动轴承故障诊断中样本分布差异大、有效故障样本少以及不同故障样本数量不均衡所导致的诊断精度较低的问题;提出基于个性化联邦迁移学习(personalized federated transfer learning,PFTL)的滚动轴承故障诊断方法。在所提出的P...为了解决滚动轴承故障诊断中样本分布差异大、有效故障样本少以及不同故障样本数量不均衡所导致的诊断精度较低的问题;提出基于个性化联邦迁移学习(personalized federated transfer learning,PFTL)的滚动轴承故障诊断方法。在所提出的PFTL中,首先在预训练阶段,将不同分布的各类型故障样本作为联邦学习的各个客户端的输入,并引入贝叶斯层级模型对联邦学习的本地训练和聚合规则进行个性化调整,从而使得预训练模型在避免过拟合问题的同时具有较强的泛化能力;其次引入模型补丁,对预训练模型结构进行调整,并利用目标任务样本对模型进一步微调;最后在CWRU轴承数据集上进行故障诊断实验。实验结果证明所提方法的有效性。展开更多
运行数据显示全光纤电流互感器(fiber optic current transformer,FOCT)在极端环境下(温度为–45~85℃、振动加速度>15 m/s^(2))故障概率明显偏高,因此研究极端环境对FOCT性能的影响十分必要。在分析FOCT工作原理基础上,重点讨论了F...运行数据显示全光纤电流互感器(fiber optic current transformer,FOCT)在极端环境下(温度为–45~85℃、振动加速度>15 m/s^(2))故障概率明显偏高,因此研究极端环境对FOCT性能的影响十分必要。在分析FOCT工作原理基础上,重点讨论了FOCT核心模块的结构特征及极端环境的影响,并建立FOCT传变模型。根据FOCT真实工作环境,分析了极端环境对其测量准确性的影响。结果表明:温度的升高、光纤长度的增加、振动加速度的变大,都会使FOCT比差增大,测量精度下降。特别是在极端环境下,测量误差较大,无法满足0.2S级测量准确度的要求。为验证模型的可靠性,开展了温度和振动影响试验。针对现有试验缺乏对极端环境的考核,提出增加测点的温度试验方法和增加振动响应试验及振动耐久试验的振动试验方法。试验结果与仿真结果对比表明:两者结果具有一致性,偏差电流波形变化趋势比较一致。该研究为FOCT可靠性问题提供有益参考。展开更多
文摘The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.
基金Supported by National Natural Science Foundation of P. R. China （60574083）, Key Laboratory of Process Industry Automation, State Education Ministry of China （PAL200514）
基金Project(50807002) supported by the National Natural Science Foundation of ChinaProject(SKLD10KM05) supported by Opening Fund of State Key Laboratory of Power System and Generation EquipmentsProject(201206025007) supported by the National Scholarship Fund,China
文摘The behavior of matrix converter(MC) drive systems under the condition of MC short-circuit faults is comprehensively investigated. Two isolation strategies using semiconductors and high speed fuses(HSFs) for MC short-circuit faults are examined and their performances are compared. The behavior of MC drive systems during the fuse action time under different operating conditions is explored. The feasibility of fault-tolerant operation during the fuse action time is also studied. The basic selection laws for the HSFs and the requirements for the passive components of the MC drive system from the point view of short-circuit faults are also discussed. Simulation results are used to demonstrate the feasibility of the proposed isolation strategies.
文摘针对实际采煤机轴承故障诊断中存在变工况特征提取困难,故障训练样本不足等问题,结合当今流行的迁移学习的方法,提出了一种多尺度迁移学习的轴承诊断方法。首先通过经验模式分解(Empirical Mode Decomposition,EMD)从振动信号中分解成不同频率的本征模态函数(Intrinsic Mode Function,IMF);其次将得到的不同频率的IMF与卷积神经网络中不同尺寸卷积核提取到的丰富特征互补构建多尺度特征融合;采用联合最大平均差异(Joint Maximum Mean Discrep⁃ancy,JMMD)特征迁移的方法使源域与目标域联合分布差异最小化,然后通过多尺度融合模型进行分类识别;最后在凯斯西储大学轴承数据集和江南大学数据集对该方法进行了验证。实验结果证明该模型在两种不同工况和型号的轴承数据集中均取得较高的准确率,表现出模型良好的泛化能力。
文摘为了解决滚动轴承故障诊断中样本分布差异大、有效故障样本少以及不同故障样本数量不均衡所导致的诊断精度较低的问题;提出基于个性化联邦迁移学习(personalized federated transfer learning,PFTL)的滚动轴承故障诊断方法。在所提出的PFTL中,首先在预训练阶段,将不同分布的各类型故障样本作为联邦学习的各个客户端的输入,并引入贝叶斯层级模型对联邦学习的本地训练和聚合规则进行个性化调整,从而使得预训练模型在避免过拟合问题的同时具有较强的泛化能力;其次引入模型补丁,对预训练模型结构进行调整,并利用目标任务样本对模型进一步微调;最后在CWRU轴承数据集上进行故障诊断实验。实验结果证明所提方法的有效性。
文摘运行数据显示全光纤电流互感器(fiber optic current transformer,FOCT)在极端环境下(温度为–45~85℃、振动加速度>15 m/s^(2))故障概率明显偏高,因此研究极端环境对FOCT性能的影响十分必要。在分析FOCT工作原理基础上,重点讨论了FOCT核心模块的结构特征及极端环境的影响,并建立FOCT传变模型。根据FOCT真实工作环境,分析了极端环境对其测量准确性的影响。结果表明:温度的升高、光纤长度的增加、振动加速度的变大,都会使FOCT比差增大,测量精度下降。特别是在极端环境下,测量误差较大,无法满足0.2S级测量准确度的要求。为验证模型的可靠性,开展了温度和振动影响试验。针对现有试验缺乏对极端环境的考核,提出增加测点的温度试验方法和增加振动响应试验及振动耐久试验的振动试验方法。试验结果与仿真结果对比表明:两者结果具有一致性,偏差电流波形变化趋势比较一致。该研究为FOCT可靠性问题提供有益参考。
