SVM Algorithm for Vibration Fault Diagnosis in Centrifugal Pump

Vibration failure in the pumping system is a significant issue for indus-tries that rely on the pump as a critical device which requires regular maintenance.To save energy and money,a new automated system must be developed that can detect anomalies at an early stage.This paper presents a case study of a machine learning(ML)-based computational technique for automatic fault detection in a cascade pumping system based on variable frequency drive(VFD).Since the intensity of the vibrational effect depends on which axis has the most significant effect,a three-axis accelerometer is used to measure it in the pumping system.The emphasis is on determining the vibration effect on different axes.For experiment,various ML algorithms are investigated on collected vibratory data through Matlab software in x,y,z axes and performances of the algorithms are compared based on accuracy rate,prediction speed and training time.Based on the proposed research results,the multiclass support vector machine(MSVM)is found to be the best suitable algorithm compared to other algorithms.It has been demonstrated that ML algorithms can detect faults automatically rather than conventional meth-ods.MSVM is used for the proposed work because it is less complex and pro-duces better results with a limited data set.

Review:Measurement-Based Monitoring and Fault Identification in Centrifugal Pumps

Condition based maintenance(CBM) is one of the solutions to machinery maintenance requirements. Latest approaches to CBM aim at reducing human engagement in the real-time fault detection and decision making. Machine learning techniques like fuzzy-logic-based systems, neural networks, and support vector machines help to reduce human involvement. Most of these techniques provide fault information with 100% confidence. It is undeniably apparent that this area has a vast application scope. To facilitate future exploration, this review is presented describing the centrifugal pump faults, the signals they generate, their CBM based diagnostic schemes, and case studies for blockage and cavitation fault detection in centrifugal pump(CP) by performing the experiment on test rig. The classification accuracy is above 98% for fault detection. This review gives a head-start to new researchers in this field and identifies the un-touched areas pertaining to CP fault diagnosis.

Fault diagnosis and analysis of main sea water pump based on vibration monitoring in offshore oil field

The main sea water pump is the key equipment for the floating production storage and offloading （FPSO）. Affected by some factors such as hull deformation, sea water corrosion, rigid base and pipeline stress, the vibration value of main sea water pump in the horizontal direction is abnormally high and malfunctions usually happen. Therefore, it is essential to make fault diagnosis of main sea water pump, By conventional off-line monitoring and vibration amplitude spectrum analysis, the fault cycle is found and the alarm value and stop value of equipment are set, which is helpful to equipment maintenance and accident prevention.

Multiple fault diagnosis of down-hole conditions of sucker-rod pumping wells based on Freeman chain code and DCA

It is important to achieve continuous, stable and efficient pumping well operation in actual oilfield operation. Down-hole pumping well working conditions can be monitored in real-time and a reasonable production scheme can be designed when computer diagnosis is used. However, it is difficult to make a comprehensive analysis to supply efficient technical guidance for operation of the pumping well with multiple faults of down-hole conditions, which cannot be effectively dealt with by the common methods. To solve this problem, a method based on designated component analysis （DCA） is used in this paper. Freeman chain code is used to represent the down-hole dynamometer card whose important characteristics are extracted to construct a designated mode set. A control chart is used as a basis for fault detection. The upper and lower control lines on the control chart are determined from standard samples in normal working conditions. In an incompletely orthogonal mode, the designated mode set could be divided into some subsets in which the modes are completely orthogonal. The observed data is projected into each designated mode to realize fault detection according to the upper and lower control lines. The examples show that the proposed method can effectively diagnose multiple faults of down-hole conditions.

Fault diagnosis for down-hole conditions of sucker rod pumping systems based on the FBH-SC method

Dynamometer cards are commonly used to analyze down-hole working conditions of pumping systems in actual oil production. Nowadays, the traditional supervised learning methods heavily rely on the classification accuracy of the training samples. In order to reduce the errors of manual classification, an automatic clustering algorithm is proposed and applied to diagnose down-hole conditions of pumping systems. The spectral clustering （SC） is a new clustering algorithm, which is suitable for any data distribution. However, it is sensitive to initial cluster centers and scale parameters, and needs to predefine the cluster number. In order to overcome these shortcom- ings, we propose an automatic clustering algorithm, fast black hole-spectral clustering （FBH-SC）. The FBH algo- rithm is used to replace the K-mean method in SC, and a CritC index function is used as the target function to automatically choose the best scale parameter and clus- tering number in the clustering process. Different simulation experiments were designed to define the relationship among scale parameter, clustering number, CritC index value, and clustering accuracy. Finally, an example is given to validate the effectiveness of the proposed algorithm.

Adaptive fault diagnosis of sucker rod pump systems based on optimal perceptron and simulation data

A highly precise and timely diagnosis technology can help effectively monitor and adjust the sucker rod production system(SRPS)used in oil wells to ensure a safe and efficient production.The current diagnosis method is pattern recognition of a dynamometer card(DC)based on feature extraction and perceptron.The premise of this method is that the training and target data have the same distribution.However,the training data are collected from a field SRPS with different system parameters designed to adapt to production conditions,which may significantly affect the diagnostic accuracy.To address this issue,in this study,an improved model of the sucker rod string(SRS)is derived by adding faultparameter dimensions,with which DCs under 16 working conditions could be generated.Subsequently an adaptive diagnosis method is proposed by taking simulated DCs generated near the working point of the target SRPS as training data.Meanwhile,to further improve the accuracy of the proposed method,the DC features are improved by relative normalization and using additional features of the DC position to increase the distance between different types of samples.The parameters of the perceptron are optimized to promote its discriminability.Finally,the accuracy and real-time performance of the proposed adaptive diagnosis method are validated using field data.

Design on Vibration Monitoring and Fault Diagnosis System of Large Water Pump Motor

Large water pump motor,whose operation decides the reliability of the whole production line,plays a very important role.Therefore,its online condition monitoring can help companies better know its operation,process fault analysis and protection.The essay mainly studies and designs large water pump motor′s real time vibration monitoring and fault diagnosis system.The essay completes the systems project design,the establishment of the system and performance test.Eddy-currentsensor,XM-120 vibration module,XM-320 axial translation module,XM-362 temperature module,XM-360 process amount module and XM-500 gateway module are used to measure the axial vibration and displacement of main motors.Laboratory tests prove that the system can meet the requirements of motor vibration monitoring.

Study of Fault Diagnosis in Bend Axis Piston Pump

On the basis of theoretical analysis and experimental rerearck, the vibration characteristics of the ZB1-107 bend axis piston pump that is wldely ed in mining machinery is studied in the paper, and the study provides the basis for pump fault diagnesis. The vibration signals of the rault-rree pump and tbe faulty pump have been compared in frequency domaln and it is round that tbere is obvious differeuce in their vibration frequency spectra. The experimentol results demonstrate that the raults, such as port plate wear and tear and the looseness or ball joint or the conuecting rod, can be effectively detected through vibration analysis.

Application of Technology in the Fault Diagnosis of Large Centrifugal Pump Units

In order to promote the stability of centrifugal pump units and maximize the role of centrifugal pumps, this paper analyzes the composition and basic working principle of centrifugal pumps, presents the main concerns of centrifugal pump maintenance, and finally investigates the common faults and maintenance methods of centrifugal pumps for reference.

Employing adaptive fuzzy computing for RCP intelligent control and fault diagnosis

Loss of coolant accident(LOCA),loss of fluid accident(LOFA),and loss of vacuum accident(LOVA)are the most severe accidents that can occur in nuclear power reactors(NPRs).These accidents occur when the reactor loses its cooling media,leading to uncontrolled chain reactions akin to a nuclear bomb.This article is focused on exploring methods to prevent such accidents and ensure that the reactor cooling system remains fully controlled.The reactor coolant pump(RCP)has a pivotal role in facilitating heat exchange between the primary cycle,which is connected to the reactor core,and the secondary cycle associated with the steam generator.Furthermore,the RCP is integral to preventing catastrophic events such as LOCA,LOFA,and LOVA accidents.In this study,we discuss the most critical aspects related to the RCP,specifically focusing on RCP control and RCP fault diagnosis.The AI-based adaptive fuzzy method is used to regulate the RCP’s speed and torque,whereas the neural fault diagnosis system(NFDS)is implemented for alarm signaling and fault diagnosis in nuclear reactors.To address the limitations of linguistic and statistical intelligence approaches,an integration of the statistical approach with fuzzy logic has been proposed.This integrated system leverages the strengths of both methods.Adaptive fuzzy control was applied to the VVER 1200 NPR-RCP induction motor,and the NFDS was implemented on the Kori-2 NPR-RCP.

基岩油气成藏特征与中国陆上深层基岩油气勘探方向

基于全球基岩油气藏数据库和中国基岩油气藏解剖,深入分析基岩油气成藏特征,探讨深层基岩油气成藏的有利条件和勘探方向。研究表明:全球已发现的基岩油气田主要分布在埋深小于4500 m的中浅层,层位以太古宇和前寒武系为主,储集层岩性以花岗岩和变质岩为主;规模较大的基岩油气田主要分布在中新生代构造运动活跃的裂谷盆地、弧后盆地和前陆盆地。基岩油气成藏特征主要表现为:(1)以孔隙-裂缝型低孔特低渗储集层为主,非均质性强,强抗压实作用导致储集层物性不受埋深控制,规模成储期为盆地基底风化剥蚀期及后期构造改造期;(2)他源供烃,成藏组合可划分为烃源岩-基岩接触型和烃源岩-基岩分离型两大类;(3)烃源岩异常高压和基岩储集层常压-低压,导致烃源岩与储集层之间存在较大的压力差,有利于深层基岩抽吸成藏。基底构造活动性、成藏组合关系、深大断裂(尤其走滑断裂)发育程度及区域性盖层等是深层基岩选区评价的主要参数;古老克拉通盆地陆内裂谷边缘的前寒武系结晶基底、紧邻生烃凹陷的古生代褶皱基底和中新生代块断基底,均具有较好的成藏条件,是未来深层基岩油气勘探的主要方向。

基于FFT-LSTM的抽水蓄能发电机定子匝间短路故障诊断方法

定子短路故障是抽水蓄能发电机常见的故障之一,其会对发电机的性能和安全性产生严重影响,为确保抽水蓄能发电机安全稳定运行,提出基于FFT-LSTM的抽水蓄能发电机定子短路故障诊断方法。建立抽水蓄能发电机定子绕组匝间短路故障模型,分析定子绕组匝间短路故障时,发电机定子电、磁相关状态。以抽水蓄能发电机定子绕组匝间短路故障时的三相电流信号为依据,基于磁势相等原理将三相电流变换成两相电流后,利用FFT转换定子两相电流的时域信号为频域信号,获取故障电流频谱图输入LSTM网络中进行处理,输出抽水蓄能发电机定子绕组匝间短路故障诊断结果。实验结果表明,该方法可以更好地区分抽水蓄能发电机正常与故障状态,实现抽水蓄能发电机定子绕组匝间短路故障诊断,且故障诊断的交叉熵损失低。

基于信号处理的液压泵故障检测方法研究综述

液压泵的故障诊断是其正常工作和健康管理的关键,基于信号处理的液压泵诊断方法已经成为主流。近年来,学者们对液压泵故障诊断的研究非常活跃,但对液压泵故障分析和诊断方法缺少系统的总结和分析。通过对液压泵相关文献进行统计分析,系统地总结了液压泵故障产生的原因、故障诊断的基本方法及研究进展,指出了液压泵故障分析和诊断领域的发展前景,为研究人员和相关维修人员提供了参考价值并指明了研究方向。

基于多点电压电流信息融合的泵房设备故障诊断方法

为了实现自来水厂泵房及其二次加压泵房中配电设备与水泵故障的快速在线诊断与报警,保障泵房设备的安全运行,提出一种基于多点电压与电流信息融合技术的泵房设备故障在线诊断方法.基于泵房设备配电系统与设备构成,从理论上系统地分析了泵房配电设备与水泵常见故障所具有的电压与电流特征及其相互影响机理,使用配电线路中多个关键节点的实时电压、电流测量值进行融合处理,提取泵房配电主回路供电电压超限、配电线路中空气开关或接触器触头开路、水泵电机绕组开路、水泵电机绕组不平衡与匝间短路、水泵堵转、电机与水泵连接机构断开6类故障的故障特征,实现这些故障的在线诊断.阐述了故障诊断方法的执行条件、实现原理及具体实现步骤.使用MATLAB/Simulink搭建仿真模型进行实验验证,仿真结果表明,本文方法能有效地诊断泵房设备故障,克服了现有人工排查耗时耗力与离线分析不及时的不足,为自来水厂及其二次加压泵房配电设备与水泵的安全运行提供有效保障.

基于GWO-FCM的输油泵故障诊断模型自学习框架

随着输油泵场站无人化建设的发展,企业对输油泵故障诊断技术的要求也越来越高。目前,被广泛使用的利用机器学习算法进行输油泵故障诊断的方法都只能针对模型训练集中已包含的几类故障进行诊断,在企业的实际使用中,仍会出现其他不包含在训练集中的故障而不能被正确自动识别、诊断。针对上述问题,设计了一种输油泵故障诊断模型自学习框架,通过信号处理技术结合深度学习提取深层故障特征,提高工业现场数据的可分性;通过模糊C均值聚类结合相似度度量判别已知故障和未知故障,对出现的未知故障模式进行识别和记录;利用频繁出现的未知故障数据重训练模型,在原有诊断功能的基础上提高对未知故障的识别、诊断及学习能力。为验证方法的有效性,使用工业现场采集的输油泵数据进行实验,结果表明,现有诊断方法所提出的输油泵故障诊断模型自学习框架能够实现对未知故障的准确识别。

多能互补发电系统故障识别与测距方法

随着大量新能源的接入,使得多端柔性直流系统(modular multilevel converter based multi-terminal direct current, MMC-MTDC)故障特征愈加复杂,快速准确的故障识别与测距是亟需解决的关键难题之一。为此,提出了一种风-光-储-蓄互补发电站经柔性直流输电外送系统故障识别与测距方法。首先,搭建风-光-储-蓄互补发电站经柔直外送系统,在此基础上,提出了一种Teager能量算子能量熵的新方法,利用测量点正负极Teager能量算子能量熵的比值构建故障选极及区段识别判据。接着,针对已识别的故障线路,提出变分模态分解(variational mode decomposition, VMD)与Teager能量算子(teager energy operator, TEO)相结合的故障测距方法。最后,利用PSCAD/EMTDC进行仿真,结果表明所提识别方法可以准确判断故障所在线路,所提测距方法能在故障发生2 ms时间窗内实现故障测距,误差率不超过2.55%,并具有较高的耐过渡电阻能力。

基于多领域耦合建模的轴向柱塞泵故障诊断方法

针对轴向柱塞泵传统单一领域建模方法存在的建模困难、仿真精度低以及故障诊断所需故障样本不足的问题,开展基于多领域耦合建模的轴向柱塞泵故障诊断方法。利用Simscape构建轴向柱塞泵多领域耦合模型,并对柱塞泄漏、主轴轴承磨损以及组合故障3种常见的故障进行模拟,再通过故障注入技术和MATLAB快速重启功能获取多种工况、不同故障程度下的压力和流量数据;随后从时域和频域对故障数据进行特征提取,同时利用单因素方差分析对故障特征进行选择;最后利用得到的特征对K邻近、朴素贝叶斯、决策树、神经网络、支持向量机等5种故障诊断算法进行训练,得到故障诊断准确率最高的算法,其平均诊断准确率为98.5%。该方法提高了轴向柱塞泵多领域耦合建模的精确性,实现了对轴向柱塞泵的有效故障诊断。

基于CNN-LSTM的钻井泵液力端故障诊断方法研究

钻井泵液力端工作环境复杂,容易发生故障,传统故障诊断方法难以满足钻井现场需求。针对五缸式钻井泵,开展了基于深度神经网络的钻井泵液力端故障诊断研究,设计了CNN-LSTM故障诊断模型结构,研究了LSTM对故障诊断模型性能影响。结果表明,提出的CNN-LSTM模型实现了钻井泵液力端多种工况下9类故障快速准确诊断,通过引入LSTM结构,将故障诊断准确率提升了7.85%,达到了97.67%。因此提出的CNN-LSTM故障诊断模型可为钻井现场提供一种高效准确的钻井泵液力端故障诊断方法。

基于多维振动特征图谱的特高压换流阀主循环泵轻量化故障诊断模型

针对特高压换流阀阀冷系统中主循环泵故障特征提取难和故障诊断模型规模大的问题,提出一种基于多维振动特征图谱的轻量化主循环泵故障诊断模型。首先,基于振动轨迹图像(vibration locus image,VLI)和伪颜色编码的时域特征提取方法,构建主循环泵的时域特征图谱。其次,融合马尔科夫变迁场(Markov transition fields,MTF)和小波包变换(wavelet packet transform,WPT),全尺度提取振动信号的低频和高频故障特征,构建主循环泵的频域、时频域特征图谱。最后,通过全维度动态卷积(omni-dimensional dynamic convolution,ODconv)优化轻量化卷积神经网络模型框架,构建了轻量化主循环泵故障诊断模型(OD-ShuffleNet)。并融合时域、频域和时频域故障特征,在减少硬件资源占用的基础上进一步提升模型的故障诊断精度。分析结果表明,模型的诊断准确率为95.0%,优于经典卷积神经网络架构。

基于无监督学习的抽油机井示功图自动聚类与批量标注方法

为充分利用大量未标注样本、节约人力与时间,提出了基于无监督学习的抽油机井示功图自动聚类与批量标注方法。首先,将抽油机驴头往复运动产生的位移、载荷数据转化为示功图图片样本,其中,示功图的横坐标为位移,纵坐标为载荷;其次,加载在ImageNet上训练过的带有一系列权重参数、具有强特征提取能力的卷积神经网络模型;然后,去除该网络模型的全连接层,利用该网络模型提取示功图图片样本的特征;最后,利用k-means聚类算法对提取到的特征进行聚类分析,将具有相似特征的示功图聚到同一文件夹中。批量的对示功图聚类结果进行快速标注,从而形成抽油机井故障诊断的示功图样本集。实验随机搜集了100口抽油机井的20 000条示功图数据,结果表明,基于无监督学习的抽油机井示功图自动聚类与批量标注方法耗时短、准确率高,为示功图样本集标注提供了一种高效方法,对于充分挖掘油田大数据的应用价值具有示范意义。