Failure time prediction using adaptive logical analysis of survival curves and multiple machining signals

172 basic Cessna plane in the process of operation, the production of equipment failure is random, so the evaluation of equipment performance and to predict its failure time to improve the safe operation of the 172 basic plane has important application value. On the plane this complex system, the grey theory combined with 172 basic Cessna plane, the collection of 172 basic aircraft fault information centralized data processing, analysis, prediction model GM (1, 1), through the calculation of the GM model data, and the error precision fitting test, better realize the basic 172 aircraft equipment failure time prediction.

Download Full-text

The Next Failure Time Prediction of Escalators via Deep Neural Network with Dynamic Time Warping Preprocessing

Applied Sciences ◽

10.3390/app10165622 ◽

2020 ◽

Vol 10 (16) ◽

pp. 5622

Author(s):

Zitong Zhou ◽

Yanyang Zi ◽

Jingsong Xie ◽

Jinglong Chen ◽

Tong An

Keyword(s):

Neural Network ◽

Dynamic Time Warping ◽

Deep Neural Network ◽

Short Term Memory ◽

Failure Time ◽

Low Complexity ◽

Time Warping ◽

Average Prediction Error ◽

Time Prediction ◽

Dynamic Time

The escalator is one of the most popular travel methods in public places, and the safe working of the escalator is significant. Accurately predicting the escalator failure time can provide scientific guidance for maintenance to avoid accidents. However, failure data have features of short length, non-uniform sampling, and random interference, which makes the data modeling difficult. Therefore, a strategy that combines data quality enhancement with deep neural networks is proposed for escalator failure time prediction in this paper. First, a comprehensive selection indicator (CSI) that can describe the stationarity and complexity of time series is established to select inherently excellent failure sequences. According to the CSI, failure sequences with high stationarity and low complexity are selected as the referenced sequences to enhance the quality of other failure sequences by using dynamic time warping preprocessing. Secondly, a deep neural network combining the advantages of a convolutional neural network and long short-term memory is built to train and predict quality-enhanced failure sequences. Finally, the failure-recall record of six escalators used for 6 years is analyzed by using the proposed method as a case study, and the results show that the proposed method can reduce the average prediction error of failure time to less than one month.

Download Full-text

Percolative approach for failure time prediction of thin film interconnects under high current stress

Microelectronics Reliability ◽

10.1016/j.microrel.2004.09.009 ◽

2005 ◽

Vol 45 (2) ◽

pp. 391-395 ◽

Cited By ~ 2

Author(s):

E. Misra ◽

Md M. Islam ◽

Mahbub Hasan ◽

H.C. Kim ◽

T.L. Alford

Keyword(s):

Thin Film ◽

Failure Time ◽

High Current ◽

Current Stress ◽

Time Prediction

Download Full-text

Reply to “Comments on ‘Evolutionary neural network modeling for software cumulative failure time prediction’ by Liang Tian and Afzel Noore,” by Bev Littlewood[Reliability Engineering and System Safety 87 (2005) 45–51]

Reliability Engineering & System Safety ◽

10.1016/j.ress.2005.02.002 ◽

2006 ◽

Vol 91 (4) ◽

pp. 487-491 ◽

Cited By ~ 5

Author(s):

Liang Tian ◽

Afzel Noore

Keyword(s):

Neural Network ◽

Failure Time ◽

Network Modeling ◽

Neural Network Modeling ◽

Reliability Engineering ◽

System Safety ◽

Evolutionary Neural Network ◽

Time Prediction

Download Full-text

Failure Prediction Model and ESR Modeling of Electrolytic Capacitor With Application to Predictive Maintenance

Volume 9: 2015 ASME/IEEE International Conference on Mechatronic and Embedded Systems and Applications ◽

10.1115/detc2015-46175 ◽

2015 ◽

Author(s):

Hadi Malek ◽

Sara Dadras ◽

YangQuan Chen

Keyword(s):

Life Span ◽

Fractional Order ◽

Prediction Models ◽

Failure Time ◽

Life Time ◽

Predictive Maintenance ◽

Electrolytic Capacitor ◽

Equivalent Series Resistance ◽

Time Prediction ◽

Actual Life

Being one of the most used passive components in power electronics, electrolytic capacitors have the shortest life span due to their wear-out failure which is mainly caused by vaporization and deterioration of capacitor electrolyte. Knowing these two phenomena increase Equivalent Series Resistance (ESR) of the capacitor, tracking ESR value over the system operating time can be a good indicator for state of health of an electrolytic capacitor. In order to set the maintenance schedule, various ESR monitoring algorithms computing remaining time before failure have been investigated in literature. These real-time algorithms use classical models for ESR and life-time estimation which are not precise enough and leads the maintenance program to be either risky if the prediction is more than the actual life-time or more expensive if it is much less than the actual life span. This paper presents a generalized equivalent model using fractional order element for electrolytic capacitor to estimate the ESR and impedance of faultless running capacitor. Unlike other existing fractional order models, proposed model considers a fractional order dynamic only in the dielectric losses and the terminal capacitor remains integer order as observed in actual capacitor’s behavior. Furthermore, a novel failure predictive model using Mittage-Leffler function is proposed to track the ESR increment due to aging of the capacitor and estimate the failure time based on the information which are provided through ESR monitoring system. Using this model increase the life-time prediction accuracy. Hence the predictive maintenance of the system with capacitors nearing their failure time can be set more precisely. These two fractional order models are compared against classical ESR and life-time prediction models to illustrate the enhanced performances of the proposed models.

Download Full-text

Health Monitoring in Landing Gears

Journal of KONES ◽

10.2478/kones-2019-0020 ◽

2019 ◽

Vol 26 (1) ◽

pp. 167-174

Author(s):

Zbigniew Skorupka ◽

Andrzej Tywoniuk

Keyword(s):

Data Analysis ◽

Health Monitoring ◽

Laboratory Tests ◽

Failure Time ◽

Computer Software ◽

Actual State ◽

Monitoring Methods ◽

Maintenance Costs ◽

Time Prediction ◽

Landing Gears

Abstract Every mechanical construction loses its properties in time due to the usage wear that leads to malfunctions and, in the end, to failure. Widely used method of failure time prediction base on extended laboratory tests where a device is tested against fatigue and wear. This method is well established but is expensive, time-consuming, and costly. Another way of failure prediction is to calculate it using advanced algorithms what is faster and cheaper but less accurate than actual tests. Furthermore, both methods are not optimal due to the principle of operation based on simplified assumptions. In such cases, it is common to make the lifespan of the safety wise devices for example landing gears much less than real in case of fatal failure not covered by the predictions. This can lead to much higher price and maintenance costs of the landing gear. Nowadays the worldwide trend is to monitor the behaviour of the devices in real time and predict failure using actual state. There are several methods of health monitoring, most of them including sensors, acquisition systems and computer software for analysis. In this article, authors describe possible landing gears health monitoring methods based on authors ’ laboratory experience in sensor appliance and test data analysis. The authors also present their idea of adding health monitoring to existing landing gears where no dedicated infrastructure was initially designed.

Download Full-text