Improving autoencoder-based unsupervised damage detection in uncontrolled structural health monitoring under noisy conditions

Structural health monitoring is widely utilized in outdoor environments,especially under harsh conditions,which can introduce noise into the monitoring system.Therefore,designing an effective denoising strategy to enhance the performance of guided wave damage detection in noisy environments is crucial.This paper introduces a local temporal principal component analysis(PCA)reconstruction approach for denoising guided waves prior to implementing unsupervised damage detection,achieved through novel autoencoder-based reconstruction.Experimental results demonstrate that the proposed denoising method significantly enhances damage detection performance when guided waves are contaminated by noise,with SNR values ranging from 10 to-5 dB.Following the implementation of the proposed denoising approach,the AUC score can elevate from 0.65 to 0.96 when dealing with guided waves corrputed by noise at a level of-5 dB.Additionally,the paper provides guidance on selecting the appropriate number of components used in the denoising PCA reconstruction,aiding in the optimization of the damage detection in noisy conditions.

Massive Power Device Condition Monitoring Data Feature Extraction and Clustering Analysis using MapReduce and Graph Model

Effective storage,processing and analyzing of power device condition monitoring data faces enormous challenges.A framework is proposed that can support both MapReduce and Graph for massive monitoring data analysis at the same time based on Aliyun DTplus platform.First,power device condition monitoring data storage based on MaxCompute table and parallel permutation entropy feature extraction based on MaxCompute MapReduce are designed and implemented on DTplus platform.Then,Graph based k-means algorithm is implemented and used for massive condition monitoring data clustering analysis.Finally,performance tests are performed to compare the execution time between serial program and parallel program.Performance is analyzed from CPU cores consumption,memory utilization and parallel granularity.Experimental results show that the designed framework and parallel algorithms can efficiently process massive power device condition monitoring data.

System Reliability Analysis of Redundant Condition Monitoring Systems

The development and application of new reliability models and methods are presented to analyze the system relia- bility of complex condition monitoring systems.The methods include a method analyzing failure modes of a type of redundant con- dition monitoring systems (RCMS) by invoking failure tree model,Markov modeling techniques for analyzing system reliability of RCMS,and methods for estimating Markov model parameters.Furthermore,a computing case is investigated and many conclu- sions upon this case are summarized.Results show that the method proposed here is practical and valuable for designing condition monitoring systems and their maintenance.

Rolling Bearing Condition Monitoring Technique Based on Cage Rotation Analysis and Acoustic Emission

In this paper,we present an alternative technique for detecting changes in the operating conditions of rolling element bearings(REBs)that can lead to premature failure.The developed technique is based on measuring the kinematics of the bearing cage.The rotational motion of the cage is driven by traction forces generated in the contacts of the rolling elements with the races.It is known that the cage angular frequency relative to shaft angular frequency depends on the bearing load,the bearing speed,and the lubrication condition since these factors determine the lubricant film thickness and the associated traction forces.Since a large percentage of REB failures are due to misalignment or lubrication problems,any evidence of these conditions should be interpreted as an incipient fault.In this paper,a novel method for the measurement of the instantaneous angular speed(IAS)of the cage is developed.The method is evaluated in a deep groove ball bearing test rig equipped with a cage IAS sensor,as well as a custom acoustic emission(AE)transducer and a piezoelectric accelerometer.The IAS of the cage is analyzed under different bearing loads and shaft speeds,showing the dependence of the cage angular speed with the calculated lubricant film thickness.Typical bearing faulty operating conditions(mixed lubrication regime,lubricant depletion,and misalignment)are recreated.It is shown that the cage IAS is dependent on the lubrication regime and is sensitive to misalignment.The AE signal is also used to evaluate the lubrication regime.Experimental results suggest that the proposed technique can be used as a condition monitoring tool in industrial environments to detect abnormal REB conditions that may lead to premature failure.

Structural health monitoring of long-span suspension bridges using wavelet packet analysis

During the service life of civil engineering structures such as long-span bridges, local damage at key positions may continually accumulate, and may finally result in their sudden failure. One core issue of global vibration-based health monitoring methods is to seek some damage indices that are sensitive to structural damage, This paper proposes an online structural health monitoring method for long-span suspension bridges using wavelet packet transform （WPT）. The WPT- based method is based on the energy variations of structural ambient vibration responses decomposed using wavelet packet analysis. The main feature of this method is that the proposed wavelet packet energy spectrum （WPES） has the ability to detect structural damage from ambient vibration tests of a long-span suspension bridge. As an example application, the WPES-based health monitoring system is used on the Runyang Suspension Bridge under daily environmental conditions. The analysis reveals that changes in environmental temperature have a long-term influence on the WPES, while the effect of traffic loadings on the measured WPES of the bridge presents instantaneous changes because of the nonstationary properties of the loadings. The condition indication indices VD reflect the influences of environmental temperature on the dynamic properties of the Runyang Suspension Bridge. The field tests demonstrate that the proposed WPES-based condition indication index VD is a good candidate index for health monitoring of long-span suspension bridges under ambient excitations.

Condition Monitoring of Turbines Using Nonlinear Mapping Method

Aiming at the non-linear nature of the signals generated from turbines, curvilinear component analysis (CCA), a novel nonlinear projection method that favors local topology conservation is presented for turbines conditions monitoring. This is accomplished in two steps. Time domain features are extracted from raw vibration signals, and then they are projected into a two-dimensional output space by using CCA method and form regions indicative of specific conditions, which helps classify and identify turbine states visually. Therefore, the variation of turbine conditions can be observed clearly with the trajectory of image points for the feature data in the two-dimensional space, and the occurrence and development of failures can be monitored in time. Key words condition monitoring - turbines - nonlinear mapping - curvilinear component analysis CLC number TP 17 - TH 17 Foundation item: Supported by the National Key Basic Research Special Found of China (2003CB716207) and the National Natural Science Foundation of China (50375047)Biography: Liao Guang-lan (1974-), male, Ph. D. candidate, research direction: fault diagnosis, pattern recognition and neural networks.

Preliminary Study on the Chinese Version of Soil Sample Spectrum Analysis System

An imported energy spectrum analyzer is powerful, but English operation interface is not easy to use. According to actual work needs, preliminary design of the Chinese energy spectrum analysis system is introduced in the paper.

Integrated Condition Monitoring of Large Captive Power Plants and Aluminum Smelters

Condition monitoring is implementation of the advanced diagnostic techniques to reduce downtime and to increase the efficiency and reliability.The research is for determining the usage of advanced techniques like Vibration analysis,Oil analysis and Thermography to diagnose ensuing problems of the Plant and Machinery at an early stage and plan to take corrective and preventive actions to eliminate the forthcoming breakdown and enhancing the reliability of the system.Nowadays,the most of the industries have adopted the condition monitoring techniques as a part of support system to the basic maintenance strategies.Major condition monitoring technique they follow is Vibration Spectrum Analysis,which can detect faults at a very early stage.However implementation of other techniques like Oil analysis or Ferrography,Thermography etc.can further enhance the data interpretation as they would detect the source of abnormality at much early stage thus providing us with a longer lead time to plan and take the corrective actions.In Large Captive Power Plants and Aluminium Smelters,Integrated Condition Monitoring techniques play an important role as stoppage of primary system and its auxiliaries(boiler,steam turbine,generator,coal and ash handling plants etc.)results into the stoppage of the entire plant,which in turn leads to loss of productivity.From economical and operational point of view,it is desirable to ensure optimum level of system availability.

Comparison of Economic Life of Tractors in Iran Using Condition Monitoring （CM） and Breakdown Maintenance （BM） Methods

Repair and maintenance costs are the most important factors affecting decision making about substituting agricultural machineries. This decision is made based on the economic life （time） of machineries. In this research, condition monitoring of MF285 and MF399 tractors was performed using engine oil analysis to find the optimum life time of tractor substitution in comparison with the breakdown maintenance method in Iran. All recorded information about fixed and variable costs were selected as data base and analyzed. Data were divided （classified） based on period of annual working time. Using power regression analysis led to find mathematical models for the optimum time life definition. Cumulative working time （X） was selected as independent and cumulative costs based on definite percent of initial price （Y） was considered as dependent variable and a power law equation was found to express the costs of both MF399 and MF285 tractors as a function of working time. Results showed that in CM method, average of economic life was 13 and 11 years for MF399 and MF285, respectively. It was also found that in BM method, economic life wasl0 and 8.5 years for MF399 and MF285, respectively.

Using Oil Analysis to Study the Wear Condition of Bearing in Trunnion of Convertor During/After Run-in Period

The bearings in the trunnion of convertor are characterized by low-speed, heavy-load and huge-dimension. In case they experience failure in operation, the output of the convertor and even that of the whole product line would be affected and the huge loss would be resulted in. Thus it is very important to master the working conditions of the bearings. Vibration and oil analysis are two main techniques to monitor the conditions of the rotary machine at present. But normal vibration analysis cannot be used here because of the limitation of their sensors in signal collecting for the rotary frequencies of the bearings are too low. In this paper, the wear condition of the bearing on the driving side of the No.5 convertor during/after the run-in period was monitored through oil analysis including atomic emissive spectrum and ferrography. It has been observed that its run-in period was as long as 19 months. This is mainly attributed to the relative short accumulated working time of the bearing.

Fractional Envelope Analysis for Rolling Element Bearing Weak Fault Feature Extraction

The bearing weak fault feature extraction is crucial to mechanical fault diagnosis and machine condition monitoring. Envelope analysis based on Hilbert transform has been widely used in bearing fault feature extraction. A generalization of the Hilbert transform, the fractional Hilbert transform is defined in the frequency domain, it is based upon the modification of spatial filter with a fractional parameter, and it can be used to construct a new kind of fractional analytic signal. By performing spectrum analysis on the fractional envelope signal, the fractional envelope spectrum can be obtained. When weak faults occur in a bearing, some of the characteristic frequencies will clearly appear in the fractional envelope spectrum. These characteristic frequencies can be used for bearing weak fault feature extraction. The effectiveness of the proposed method is verified through simulation signal and experiment data. © 2017 Chinese Association of Automation.

TIME-SERIES MODELI NG AND FAULT FORECAST STUDY ON SPECTRAL ANALYSIS OF LUBRICATING OIL

The application of ti me-series modeling and forecasting method to the spectral analysis for lubricat ing oil of mechanical equipment is discussed. The AR model is used to perform a time-series modeling and forecasting analysis for the spectral analysis data co llected from aero-engines. In the oil condition monitoring field of mechanical equipment, the use of the method of time-series analysis has rarely been report ed. As indicated in the satisfactory example, a practical method for condition m onitoring and fault forecasting of mechanical equipment has been achieved.

Structural Health Monitoring of Offshore Buoyant Leg Storage and Regasification Platform:Experimental Investigations

Offshore platforms are of high strategic importance,whose preventive maintenance is on top priority.Buoyant Leg Storage and Regasification Platforms(BLSRP)are special of its kind as they handle LNG storage and processing,which are highly hazardous.Implementation of structural health monitoring(SHM)to offshore platforms ensures safe operability and structural integrity.Prospective damages on the offshore platforms under rare events can be readily identified by deploying dense array of sensors.A novel scheme of deploying wireless sensor network is experimentally investigated on an offshore BLSRP,including postulated failure modes that arise from tether failure.Response of the scaled model under wave loads is acquired by both wired and wireless sensors to validate the proposed scheme.Proposed wireless sensor network is used to trigger alert monitoring to communicate the unwarranted response of the deck and buoyant legs under the postulated failure modes.SHM triggers the alert mechanisms on exceedance of the measured data with that of the preset threshold values;alert mechanisms used in the present study include email alert and message pop-up to the validated user accounts.Presented study is a prima facie of SHM application to offshore platforms,successfully demonstrated in lab scale.

新能源汽车电池的在线监测与原位分析技术

为控制新能源汽车电池耗电量,并对其电能消耗行为进行原位分析,针对新能源汽车电池的在线监测与原位分析技术展开研究。完善动力锂电池组结构模型,根据汽车电池的热耦合特性,推导监测核函数,实现新能源汽车电池的在线监测。在此基础上,确定汽车电池的电极材料,并分别从表征特性、拉曼光谱两个角度,对其电池元件的原位特性进行研究。

无人机载CeBr3航空监测系统无限大面源探测能力研究

在核设施发生重大事故后,快速获取事故应急区内的辐射水平,对后续应急决策及救援、保护公众免受辐射危害具有至关重要的意义,核应急航空辐射监测是核应急早期快速获取辐射水平的重要技术手段和方法之一,因此核工业航测遥感中心自主研发了无人机载CeBr3航空监测系统。为准确评估该监测系统的探测能力,文章使用蒙特卡罗方法对该监测系统进行了建模,并通过60Co点源对模型的准确性进行了验证。在此基础上,采用全能谱分析、蒙特卡罗及实验相结合的方法获得监测系统对天然辐射单位能谱,并以核事故后地面受到一个新的137Cs放射性沉降场景为例,计算监测系统在不同测量条件下(天然辐射本底、晶体体积、测量高度)对无限大137Cs面源的探测效率及最小可探测活度浓度。通过模拟及计算监测系统对天然辐射本底及核应急人工放射性核素的单位能谱,准确评估了自研的无人机载CeBr3核应急航空监测系统的探测能力,为后续利用该监测系统开展核与辐射航空监测可提供重要理论依据及数据支撑。

大型铺管船托管架疲劳损伤实时监测应用与研究

托管架长期承受着交变载荷的作用,结构疲劳破坏问题不容忽视。为此,本文分别基于波浪谱和监测数据对托管架的疲劳问题进行研究。通过计算结果的对比分析,发现基于波浪谱的疲劳损伤计算,由于波浪荷载的随机性很强,导致计算结果偏于保守,适用于设计阶段结构的疲劳寿命预报评估;而基于实时监测数据的疲劳损伤计算,数据相对更加准确,适用于实时评估结构当前的疲劳损伤情况,并依此提出适当的维修和保养建议。

基于WPES与MEEMD的船用主机振动研究

为揭示船用长冲程低速柴油机健康状态下的振动特征,采用小波包能量谱(Wavelet Packet Energy Spectrum, WPES)和改进的总体平均经验模态分解(Modified Ensemble Empirical Mode Decomposition, MEEMD)结合的特征提取方法,对典型推进工况下低速机的表面振动信号进行3层小波包分解和重构。通过对能量占比较大的节点采用MEEMD方法进行分解,获得IMF1分量频谱。研究结果表明,在40%以下的较低发动机负荷时,各单次燃烧循环的振动波动较小,振动幅值基本一致。提升至50%以上发动机负荷时,燃烧引起振动波动明显增强。50%工况下,中高频能量占总能量的41.51%,为主要振动源。

实验室检测环境条件开展监测情况分析

目的了解检验检测机构检测环境条件监测情况,分析存在问题并提供相应对策。方法通过自行设计问卷,调查实验室所在领域、类型、现有能力、信息化现状与监测现状等信息。通过第三方服务平台问卷星开展问卷调查,回收问卷183份,并通过问卷星SPSSAU进行数据分析。结果调查了183家检验检测机构,实验室涉及的环境条件主要集中在温度、湿度、微生物污染、酸碱度和毒害腐蚀、有毒有害气体等,96.17%的实验室涉及温湿度,83.06%的实验室已经开展环境条件监测,监测的主要环境条件为温湿度,占86.34%,对电流、电压、光照、压力、噪声、振动、有害有毒气体等监测不足25%。仅有18.03%的检验检测机构已开展信息化监测,其他机构仍然通过传统的手工记录的方式开展监测登记。9.84%的检验机构建立了检验检测环境条件数据模型和趋势预警。结论本次调研数据分析发现国内包括政府实验室、企业实验室、第三方检验检测机构在内的实验室平台,其平台能力构成复杂、要求监测的环境条件种类繁多。但目前在监测技术融合、动态监测以及监测覆盖面方面还存在一定的不足,建议应通过建立环境条件监测信息化平台,利用数据模型建立起危险趋势预警机制、实现环境条件监测优化机制与干预管理。

考虑场址危险一致性的中国完备条件目标谱与记录选取

即使所处地区的宏观抗震设防烈度相同,不同城市的地震地质构造与活动性背景仍然存在差异。为使地震动输入与所在场址的地震危险性水平保持一致,该研究衔接我国第五代地震区划图的概率地震危险性分析结果,给出了完备条件目标谱的构建方法。采用与第五代地震区划图相同的三级潜源分布和地震动预测方程,进行目标场址的危险性分析与设定地震解耦,衔接三维(震级-经纬度)设定地震解耦结果,给出了考虑所有潜源走向与位置相对贡献的“完备”条件谱。以北京近郊某城市作为算例,结果表明,完备条件谱的均值介于使用长轴和短轴衰减关系的粗略条件谱之间,条件标准差高于粗略条件谱。研究给出的完备条件谱适用于我国区划图的长短轴类型衰减关系,可以很好体现潜源的空间分布特征与背景地震危险性差异。基于我国强震动记录数据库拟合的加速度反应谱相关系数得到的条件谱要略高于基于国外数据库的计算结果。为了实现加速度反应谱均值和标准差的高效匹配,该研究采取多元高斯分布的协方差矩阵近似替换完备条件谱高斯混合分布的协方差矩阵,结合贪心优化算法,选取得到满足目标完备条件谱分布的强震动记录数据集。研究结果系统给出了我国条件目标谱进行记录选取的构建思路与选波方案,可为后续全概率结构性态分析提供依据。

基于改进信息熵的直接刀具状态监测设备部署

在不拆刀情况下,基于机器视觉的在线刀具状态监测系统可完成刀具磨损测量和状态评估,但与在线捕获刀具图像质量息息相关的系统部署参数选择却鲜有研究.为解决上述问题,本文构建基于改进信息熵的多项式回归模型以实现刀具状态监测系统的最优部署.首先,使用自适应阈值方法去除捕获刀具图像中背景要素干扰,并通过信息熵指标评估图像中刀具磨损区域的成像质量;然后,构建相机工作距离、曝光时间与所提出评价指标之间的多项式回归模型以描述部署参数与提出评价指标的映射关系;最后,应用最小二乘法求取多项式模型系数获得最优部署参数.在确保自变量的因子水平涵盖最优部署参数情况下设计正交实验,实验结果表明:提出的评价指标与工作距离、曝光时间等部署参数之间均存在主效应关系,符合光学成像系统的变化规律;与支持向量机、决策树和K近邻(K-nearest neighbor,KNN)算法等非线性回归预测模型相比,三次多项式回归模型预测误差最小,其平均绝对误差、均方误差、均方根误差分别为0.022631,0.00068,0.026069;在多项式回归模型求解的最优部署参数下,所捕获的刀具图像的测量精度达到96.76%,提高0.74%,满足刀具状态监测的精度要求.