基于在线半监督学习的故障诊断方法研究

针对机械故障诊断中准确、完备的故障训练样本获取困难,而现有分类方法难以有效地发掘大量未标记故障样本中蕴含的有用信息,提出了一种基于在线半监督学习的故障诊断方法.该方法基于Tri-training算法将在线贯序极限学习机从监督学习模式扩展到半监督学习模式,利用少量不精确的标记样本构建初始分类器,并从大量未标记样本中在线扩充标记样本,对分类器进行增量式更新以提高其泛化性能.半监督基准数据试验结果表明,训练样本总数相同但标记样本数与未标记样本数比例不同时,所提算法得到的分类准确率相当且训练时间相差小于1.2倍.以柴油机8种工况的故障模式为对象进行试验验证,结果表明标记故障样本较少时,未标记故障样本的加入可使故障分类准确率提高5％～8％.     

Feature selection and fault‐severity classification–based machine health assessment methodology for point machine sliding‐chair degradation

In this paper, we propose an offline and online machine health assessment (MHA) methodology composed of feature extraction and selection, segmentation‐based fault severity evaluation, and classification steps. In the offline phase, the best representative feature of degradation is selected by a new filter‐based feature selection approach. The selected feature is further segmented by utilizing the bottom‐up time series segmentation to discriminate machine health states, ie, degradation levels. Then, the health state fault severity is extracted by a proposed segment evaluation approach based on within segment rate‐of‐change (RoC) and coefficient of variation (CV) statistics. To train supervised classifiers, a priori knowledge about the availability of the labeled data set is needed. To overcome this limitation, the health state fault‐severity information is used to label (eg, healthy, minor, medium, and severe) unlabeled raw condition monitoring (CM) data. In the online phase, the fault‐severity classification is carried out by kernel‐based support vector machine (SVM) classifier. Next to SVM, the k‐nearest neighbor (KNN) is also used in comparative analysis on the fault severity classification problem. Supervised classifiers are trained in the offline phase and tested in the online phase. Unlike to traditional supervised approaches, this proposed method does not require any a priori knowledge about the availability of the labeled data set. The proposed methodology is validated on infield point machine sliding‐chair degradation data to illustrate its effectiveness and applicability. The results show that the time series segmentation‐based failure severity detection and SVM‐based classification are promising.     

多波束声呐海底底质半监督学习分类方法

多波束测深声呐的反向散射数据中包含海底表层的声学信息,可以用来进行海底表层底质分类。但实际中通过物理采样获得大范围的底质类型的标签信息所需成本过高,制约了传统监督分类算法的性能。针对实际应用中只拥有大量无标签数据和少量有标签数据的情况,文章提出了基于自动编码器预训练以及伪标签自训练的半监督学习底质分类算法。利用2018年和2019年两次同一海域实验采集的多波束测深声呐反向散射数据,对所提算法进行了验证。数据处理结果表明,相比仅利用有标签数据的监督分类算法,提出的半监督学习分类算法保证分类准确率的同时所需的有标签数据更少。自动编码器预训练的半监督学习分类方法在有标签样本数量极少时的准确率仍高于75%。     

A hybrid SVM and supervised learning approach to fuzzy min‐max hyperbox classifiers

Abstract

In this paper, a fuzzy min‐max hyperbox classifier is designed to solve M‐class classification problems using a hybrid SVM and supervised learning approach. In order to solve a classification problem, a set of training patterns is gathered from a considered classification problem. However, the training set may include several noisy patterns. In order to delete the noisy patterns from the training set, the support vector machine is applied to find the noisy patterns so that the remaining training patterns can describe the behavior of the considered classification system well. Subsequently, a supervised learning method is proposed to generate fuzzy min‐max hyperboxes for the remaining training patterns so that the generated fuzzy min‐max hyperbox classifier has good generalization performance. Finally, the Iris data set is considered to demonstrate the good performance of the proposed approach for solving this classification problem.     

Unsupervised Anomaly Detection via DBSCAN for KPIs Jitters in Network Managements

For many Internet companies, a huge amount of KPIs (e.g., server CPU usage, network usage, business monitoring data) will be generated every day. How to closely monitor various KPIs, and then quickly and accurately detect anomalies in such huge data for troubleshooting and recovering business is a great challenge, especially for unlabeled data. The generated KPIs can be detected by supervised learning with labeled data, but the current problem is that most KPIs are unlabeled. That is a time-consuming and laborious work to label anomaly for company engineers. Build an unsupervised model to detect unlabeled data is an urgent need at present. In this paper, unsupervised learning DBSCAN combined with feature extraction of data has been used, and for some KPIs, its best F-Score can reach about 0.9, which is quite good for solving the current problem.     

An unsupervised approach for fault diagnosis of power transformers

Electrical utilities apply condition monitoring on power transformers (PTs) to prevent unplanned outages and detect incipient faults. This monitoring is often done using dissolved gas analysis (DGA) coupled with engineering methods to interpret the data, however the obtained results lack accuracy and reproducibility. In order to improve accuracy, various advanced analytical methods have been proposed in the literature. Nonetheless, these methods are often hard to interpret by the decision-maker and require a substantial amount of failure records to be trained. In the context of the PTs, failure data quality is recurrently questionable, and failure records are scarce when compared to nonfailure records. This work tackles these challenges by proposing a novel unsupervised methodology for diagnosing PT condition. Differently from the supervised approaches in the literature, our method does not require the labeling of DGA records and incorporates a visual representation of the results in a 2D scatter plot to assist in interpretation. A modified clustering technique is used to classify the condition of different PTs using historical DGA data. Finally, well-known engineering methods are applied to interpret each of the obtained clusters. The approach was validated using data from two different real-world data sets provided by a generation company and a distribution system operator. The results highlight the advantages of the proposed approach and outperformed engineering methods (from IEC and IEEE standards) and companies legacy method. The approach was also validated on the public IEC TC10 database, showing the capability to achieve comparable accuracy with supervised learning methods from the literature. As a result of the methodology performance, both companies are currently using it in their daily DGA diagnosis.     

Estimation of Complicated Profiles in Phase I,Clustering and S‐estimation Approaches

Some quality characteristics are well defined when expressed as a function of an independent variable. This function is usually called a profile. If the functional form of the profile is known, parametric methods could be used to monitor the profile representing a process. However, some processes are complicated, and it is not suitable to use parametric models. In these cases, nonparametric methods may be used to monitor the profiles. One of the powerful nonparametric profile monitoring methods is to use wavelets. In this paper, the issue of estimating the complicated profiles in phase I is studied. In order to monitor the process using wavelets, it is required to estimate the vector of wavelet coefficients. Classical estimators are usually used to estimate the coefficients vector. These estimators should be used when the data do not contain outliers. However, it is possible that the data set is contaminated and includes some outliers. Thus, it is better to use robust estimators that are insensitive to the presence of outliers. In this paper, two robust estimators for estimating the complicated profiles using wavelets are proposed. In the first approach, the dimension of the coefficients vector is reduced by means of PCA incorporated into clustering. The second approach is based on the S‐estimation method. An extensive simulation study is performed using matlab ® software to evaluate the proposed methods and to compare the results with an existing classical method. The results show the well performance of the suggested methods in estimating the model parameters when the data set is not contaminated and in the presence of outliers. Copyright © 2016 John Wiley & Sons, Ltd.     

Modeling and analysis of disease and risk factors through learning Bayesian networks from observational data

This paper focuses on identification of the relationships between a disease and its potential risk factors using Bayesian networks in an epidemiologic study, with the emphasis on integrating medical domain knowledge and statistical data analysis. An integrated approach is developed to identify the risk factors associated with patients' occupational histories and is demonstrated using real‐world data. This approach includes several steps. First, raw data are preprocessed into a format that is acceptable to the learning algorithms of Bayesian networks. Some important considerations are discussed to address the uniqueness of the data and the challenges of the learning. Second, a Bayesian network is learned from the preprocessed data set by integrating medical domain knowledge and generic learning algorithms. Third, the relationships revealed by the Bayesian network are used for risk factor analysis, including identification of a group of people who share certain common characteristics and have a relatively high probability of developing the disease, and prediction of a person's risk of developing the disease given information on his/her occupational history. Copyright © 2007 John Wiley & Sons, Ltd.     

基于BDPCA聚类算法的航空发动机故障数据标记

航空发动机作为飞行器的动力核心对飞行器的安全飞行有着举足轻重的作用,保证航空发动机的平稳运行对飞行安全有着重大意义。在基于有监督学习的航空发动机故障诊断技术不断取得进展的同时,如何将平时获取的大量未标记数据转换为能够用来训练故障诊断模型的带标记数据,成为了制约行业发展的一大瓶颈。针对这一问题引入了基于无监督学习的DPCA算法,用以实现对未标记数据集的准确分类与标记,并针对DPCA算法在局部密度计算与簇类别数选择方面的缺陷进行了优化:针对原始DPCA算法应用标准高斯核计算局部密度易造成误识别的状况,引入共享邻域算法对局部密度的计算方法进行优化;针对原始DPCA算法需要人工研判确定簇类别数易造成的误识别状况,引入BIC选择准则对簇类别数的选择方法进行优化;提出了原始DPCA算法与共享邻域算法以及BIC选择准则相结合的BDPCA算法。最后应用航空发动机转子故障数据对BDPCA算法进行了性能验证并取得了良好的结果,证实了BDPCA算法在航空发动机气路故障诊断领域有较高的实用价值。     

Condition‐based Maintenance Optimization Using Neural Network‐based Health Condition Prediction

Artificial neural network (ANN)‐based methods have been extensively investigated for equipment health condition prediction. However, effective condition‐based maintenance (CBM) optimization methods utilizing ANN prediction information are currently not available due to two key challenges: (i) ANN prediction models typically only give a single remaining life prediction value, and it is hard to quantify the uncertainty associated with the predicted value; (ii) simulation methods are generally used for evaluating the cost of the CBM policies, while more accurate and efficient numerical methods are not available, which is critical for performing CBM optimization. In this paper, we propose a CBM optimization approach based on ANN remaining life prediction information, in which the above‐mentioned key challenges are addressed. The CBM policy is defined by a failure probability threshold value. The remaining life prediction uncertainty is estimated based on ANN lifetime prediction errors on the test set during the ANN training and testing processes. A numerical method is developed to evaluate the cost of the proposed CBM policy more accurately and efficiently. Optimization can be performed to find the optimal failure probability threshold value corresponding to the lowest maintenance cost. The effectiveness of the proposed CBM approach is demonstrated using two simulated degradation data sets and a real‐world condition monitoring data set collected from pump bearings. The proposed approach is also compared with benchmark maintenance policies and is found to outperform the benchmark policies. The proposed CBM approach can also be adapted to utilize information obtained using other prognostics methods. Copyright © 2012 John Wiley & Sons, Ltd.