Predictive maintenance

Predictive maintenance techniques are designed to help determine the condition of in-service equipment in order to estimate when maintenance should be performed. This approach promises cost savings over routine or time-based preventive maintenance, because tasks are performed only when warranted. Thus, it is regarded as condition-based maintenance carried out as suggested by estimations of the degradation state of an item.[1][2]

The main promise of predictive maintenance is to allow convenient scheduling of corrective maintenance, and to prevent unexpected equipment failures. The key is "the right infor quipment lifetime, increased plant safety, fewer accidents with negative impact on environment, and optimized spare parts handling.

Predictive maintenance differs from preventive maintenance because it relies on the actual condition of equipment, rather than average or expected life statistics, to predict when maintenance will be required.

Some of the main components that are necessary for implementing predictive maintenance are data collection and preprocessing, early fault detection, fault detection, time to failure prediction, maintenance scheduling and resource optimization.[3] Predictive maintenance has also been considered to be one of the driving forces for improving productivity and one of the ways to achieve "just-in-time" in manufacturing.[4]

Since 2001, the Center for Intelligent Maintenance Systems[5] Industry/University Collaborative Research Center) has been working in the development of advance methods and technologies for predictive maintenance. The developed approaches have been successfully validated in over 70 projects conducted with research and industry partners for enabling products and systems to achieve and sustain near-zero breakdown. The vision has been to estimate the current health of a plant equipment and predict the next fault event for improved productivity and asset utilization.[6][7] and later adopted by Fanuc in 2013 [8]


Predictive maintenance evaluates the condition of equipment by performing periodic (offline) or continuous (online) equipment condition monitoring. The ultimate goal of the approach is to perform maintenance at a scheduled point in time when the maintenance activity is most cost-effective and before the equipment loses performance within a threshold. This results in a reduction in unplanned downtime costs because of failure where for instance costs can be in the hundreds of thousands per day depending on industry.[9] In energy production in addition to loss of revenue and component costs, fines can be levied for non delivery increasing costs even further. This is in contrast to time- and/or operation count-based maintenance, where a piece of equipment gets maintained whether it needs it or not. Time-based maintenance is labor intensive, ineffective in identifying problems that develop between scheduled inspections, and so is not cost-effective. The fundamental idea is to transform the traditional ‘fail and fix’ maintenance practice to a ‘predict and prevent’ approach.[10]

The "predictive" component of predictive maintenance stems from the goal of predicting the future trend of the equipment's condition. This approach uses principles of statistical process control to determine at what point in the future maintenance activities will be appropriate.

Most predictive inspections are performed while equipment is in service, thereby minimizing disruption of normal system operations. Adoption of predictive maintenance can result in substantial cost savings and higher system reliability.

Reliability-centered maintenance emphasizes the use of predictive maintenance techniques in addition to traditional preventive measures. When properly implemented, it provides companies with a tool for achieving lowest asset net present costs for a given level of performance and risk.[11]

One goal is to transfer the predictive maintenance data to a computerized maintenance management system so that the equipment condition data is sent to the right equipment object to trigger maintenance planning, work order execution, and reporting.[12] Unless this is achieved, the predictive maintenance solution is of limited value, at least if the solution is implemented on a medium to large size plant with tens of thousands pieces of equipment. In 2010, the mining company Boliden, implemented a combined Distributed Control System and predictive maintenance solution integrated with the plant computerized maintenance management system on an object to object level, transferring equipment data using protocols like Highway Addressable Remote Transducer Protocol, IEC61850 and OLE for process control.


To evaluate equipment condition, predictive maintenance utilizes nondestructive testing technologies such as infrared, acoustic (partial discharge and airborne ultrasonic), corona detection, vibration analysis, sound level measurements, oil analysis, and other specific online tests. A new approach in this area is to utilize measurements on the actual equipment in combination with measurement of process performance, measured by other devices, to trigger equipment maintenance. This is primarily available in collaborative process automation systems (CPAS). Site measurements are often supported by wireless sensor networks to reduce the wiring cost.

Vibration analysis is most productive on high-speed rotating equipment and can be the most expensive component of a PdM program to get up and running.[13] Vibration analysis, when properly done, allows the user to evaluate the condition of equipment and avoid failures. The latest generation of vibration analyzers comprises more capabilities and automated functions than its predecessors. Many units display the full vibration spectrum of three axes simultaneously, providing a snapshot of what is going on with a particular machine. But despite such capabilities, not even the most sophisticated equipment successfully predicts developing problems unless the operator understands and applies the basics of vibration analysis.[14]

In certain situations, strong background noise interferences from several competing sources may mask the signal of interest and hinder the industrial applicability of vibration sensors. Consequently, motor current signature analysis (MCSA) is a non-intrusive alternative to vibration measurement which has the potential to monitor faults from both electrical and mechanical systems.[15]

Remote visual inspection is the first non destructive testing. It provides a cost-efficient primary assessment. Essential information and defaults can be deduced from the external appearance of the piece, such as folds, breaks, cracks and corrosion. The remote visual inspection has to be carried out in good conditions with a sufficient lighting (350 LUX at least). When the part of the piece to be controlled is not directly accessible, an instrument made of mirrors and lenses called endoscope is used. Hidden defects with external irregularities may indicate a more serious defect inside.

Acoustical analysis can be done on a sonic or ultrasonic level. New ultrasonic techniques for condition monitoring make it possible to "hear" friction and stress in rotating machinery, which can predict deterioration earlier than conventional techniques.[16] Ultrasonic technology is sensitive to high-frequency sounds that are inaudible to the human ear and distinguishes them from lower-frequency sounds and mechanical vibration. Machine friction and stress waves produce distinctive sounds in the upper ultrasonic range. Changes in these friction and stress waves can suggest deteriorating conditions much earlier than technologies such as vibration or oil analysis. With proper ultrasonic measurement and analysis, it’s possible to differentiate normal wear from abnormal wear, physical damage, imbalance conditions, and lubrication problems based on a direct relationship between asset and operating conditions.

Sonic monitoring equipment is less expensive, but it also has fewer uses than ultrasonic technologies. Sonic technology is useful only on mechanical equipment, while ultrasonic equipment can detect electrical problems and is more flexible and reliable in detecting mechanical problems.

Infrared monitoring and analysis has the widest range of application (from high- to low-speed equipment), and it can be effective for spotting both mechanical and electrical failures; some consider it to currently be the most cost-effective technology. Oil analysis is a long-term program that, where relevant, can eventually be more predictive than any of the other technologies. It can take years for a plant's oil program to reach this level of sophistication and effectiveness. Analytical techniques performed on oil samples can be classified in two categories: used oil analysis and wear particle analysis. Used oil analysis determines the condition of the lubricant itself, determines the quality of the lubricant, and checks its suitability for continued use. Wear particle analysis determines the mechanical condition of machine components that are lubricated. Through wear particle analysis, you can identify the composition of the solid material present and evaluate particle type, size, concentration, distribution, and morphology.[17]

The use of Model Based Condition Monitoring for predictive maintenance programs is becoming increasingly popular over time. This method involves spectral analysis on the motor’s current and voltage signals and then compares the measured parameters to a known and learned model of the motor to diagnose various electrical and mechanical anomalies. This process of "model based" condition monitoring was originally designed and used on NASA’s space shuttle to monitor and detect developing faults in the space shuttle’s main engine.[18] It allows for the automation of data collection and analysis tasks, providing round the clock condition monitoring and warnings about faults as they develop.


Commercial software

Applications (by industry)


  • Detect problems before they cause downtime for linear, fixed and mobile assets.[20]
  • Improving safety and track void detection through a new vehicle cab-based monitoring system
  • Siemens Tracksure track monitoring system is able to identify voids underneath track from the acceleration measured in the vehicle cab.[21]
  • Can also identify the type of track asset that the void is located under and provide an indication of the severity of the void
  • Health Monitoring of point Machines (devices used to operate railway turnouts) can aid in detecting early symptoms of degradation prior to failure.[22][23][24]


  • Early fault detection and diagnosis in the manufacturing industry.[4]
  • Manufacturers increasingly collect big data from Internet of Things (IoT) sensors in their factories and products and using different algorithms for the collected data to detect warning signs of expensive failures before they occur.[25][26]
  • Manufacturing industry: predict equipment failures can be easily found out using big data.[27][28]

Oil and Gas

  • Oil and gas companies often lack visibility into the condition of their equipment, especially in remote offshore and deep-water locations.[29]
  • Big data can provide insight to oil and gas companies, this way equipment failures and the optimal lifetime of the system and components can be analyzed and predicted.[29]
  • Considerable work has been done in the area of health monitoring and fault diagnosis of rotating machinery equipment in manufacturing industry.[30][13]


  • Detecting underlying degradation and predicting how soon a battery will reach a level of unsatisfactory performance.[31]
  • Health assessment of batteries in electric vehicles for accurate quantification of the State of Health (SOH) and its subsequent impact on vehicle mobility.[32]

See also


  1. Goriveau, Rafael; Medjaher, Kamal; Zerhouni, Noureddine (2016-11-14). From prognostics and health systems management to predictive maintenance 1 : monitoring and prognostics. ISTE Ltd and John Wiley & Sons, Inc. ISBN 978-1-84821-937-3.
  2. Mobley, R. Keith. An introduction to predictive maintenance (2nd ed.). Butterworth-Heinemann. ISBN 978-0-7506-7531-4.
  3. Amruthnath, Nagdev; Gupta, Tarun (2018). "Fault Class Prediction in Unsupervised Learning using Model-Based Clustering Approach". doi:10.13140/rg.2.2.22085.14563. Cite journal requires |journal= (help)
  4. Amruthnath, Nagdev; Gupta, Tarun (2018). "A Research Study on Unsupervised Machine Learning Algorithms for Fault Detection in Predictive Maintenance". doi:10.13140/rg.2.2.28822.24648. Cite journal requires |journal= (help)
  5. "Center for Intelligent Maintenance Systems". IMS Center. Retrieved 23 January 2014.
  6. Huang, R.; Xi, L.; Lee, J.; Liu, C. R. (21 February 2007). "The framework, impact and commercial prospects of a new predictive maintenance system: intelligent maintenance system". Production Planning & Control. 16 (7): 652–664. doi:10.1080/09537280500205837.
  7. Qiu, Hai; Jay Lee. "Near-zero downtime: Overview and trends". Reliable plant. Noria.
  8. "FANUC ZDT Application Zero Down Time". Fanuc America.
  9. "How Much Does Predictive Maintenance Save You Money?". Retrieved 2017-12-03.
  10. Lee, Jay; Jun Ni; Dragan Djurdjanovic; Hai Qiu; Haitao Liao (August 2006). "Intelligent prognostics tools and e-maintenance". Computers in Industry. 57 (6): 478. doi:10.1016/j.compind.2006.02.014.
  11. Mather, D. (2008). "The value of RCM". Plant Services.
  12. Peng, K. (2012). Equipment Management in the Post-Maintenance Era: A New Alternative to Total Productive Maintenance (TPM). CRC Press. pp. 132–136. ISBN 9781466501942. Retrieved 18 May 2018.
  13. Lee, Jay (2013). "Prognostics and health management design for rotary machinery systems—Reviews, methodology and applications". Mechanical Systems and Signal Processing. 42 (1–2): 314–334. doi:10.1016/j.ymssp.2013.06.004.
  14. Yung, Chuck (June 9, 2006). "Vibration analysis: what does it mean?". Plant Services.
  15. Imza, Inaki Bravo; Hossein Davari Ardakani; Zongchang Liu; Alfredo García-Arribas; Aitor Arnaiz; Jay Lee (2017). "Motor current signature analysis for gearbox condition monitoring under transient speeds using wavelet analysis and dual-level time synchronous averaging". Mechanical Systems and Signal Processing. 94: 73–84. Bibcode:2017MSSP...94...73B. doi:10.1016/j.ymssp.2017.02.011.
  16. Kennedy, Sheila (2006). "New tools for PdM". Putman Media. Retrieved 19 Nov 2019.
  17. Robin, Lana (August 15, 2006). "Slick tricks in oil analysis". Plant Services.
  18. Duyar, Ahmet; Merrill, Walter (March 1992). "Fault diagnosis for the Space Shuttle main engine". Journal of Guidance, Control, and Dynamics. 15 (2): 384–9. doi:10.2514/3.20847.
  19. "Predictive Maintenance Toolbox". Retrieved 2019-07-11.
  20. Predictive maintenance benefits for the railway industry, retrieved 19 November 2016
  21. Improving safety through early track void detection, retrieved 19 November 2016
  22. Ardakani, Hossein Davari; Lucas, Christina; Siegel, David; Chang, Shuo; Dersin, Pierre; Bonnet, Benjamin; Lee, Jay (2012). "PHM for railway system — A case study on the health assessment of the point machines". 2012 IEEE Conference on Prognostics and Health Management. pp. 1–5. doi:10.1109/ICPHM.2012.6299533. ISBN 978-1-4673-0358-3.
  23. Jin, Wenjing; Shi, Zhe; Siegel, David; Dersin, Pierre; Douziech, Cyril; Pugnaloni, Michele; Cascia, Piero La; Lee, Jay (2015). "Development and evaluation of health monitoring techniques for railway point machines". 2015 IEEE Conference on Prognostics and Health Management (PHM). pp. 1–11. doi:10.1109/ICPHM.2015.7245016. ISBN 978-1-4799-1894-2.
  24. Shi, Zhe; Liu, Zongchang; Lee, Jay (April 2018). "An auto-associative residual based approach for railway point system fault detection and diagnosis". Measurement. 119: 246–58. doi:10.1016/j.measurement.2018.01.062.
  25. 5 Use Cases for Predictive Maintenance and Big Data, Oracle Corporation, CA 94065 USA., retrieved 8 November 2018
  26. Lee, Jay; Lapira, Edzel; Bagheri, Behrad; Kao, Hung-an (2013). "Recent advances and trends in predictive manufacturing systems in big data environment". Manufacturing Letters. 1 (1): 38–41. doi:10.1016/j.mfglet.2013.09.005.
  27. Oracle 2018, 22 Big Data Use Cases You Want to Know, 2nd edition, Oracle Corporation, CA 94065 USA. (PDF), retrieved 12 November 2018
  28. Lee, Jay; Kao, Hung-An; Yang, Shanhu (2014). "Service innovation and smart analytics for industry 4.0 and big data environment". Procedia CIRP. 16: 3–8. doi:10.1016/j.procir.2014.02.001. ISSN 2212-8271.
  29. 22 Big Data Use Cases You Want to Know, Oracle Corporation, CA 94065 USA., retrieved 31 October 2018
  30. Lee, Jay; Cai, Haoshu; Zhao, Ming; Jia, Xiaodong (10 August 2018). "A simplified convolutional sparse filter for impulsive signature enhancement and its application to the prognostic of rotating machinery". arXiv:1808.03587. Cite journal requires |journal= (help)
  31. Seyed Mohammad Rezvanizaniani; Jay Lee; Zongchung Liu & Yan Chen (2014-06-15). "Review and recent advances in battery health monitoring and prognostics technologies for electric vehicle (EV) safety and mobility". Journal of Power Sources. 256: 110–124. Bibcode:2014JPS...256..110R. doi:10.1016/j.jpowsour.2014.01.085.
  32. Seyed Mohammad Rezvanizaniani; Seungchul Lee; Jay Lee (2011). "A comparative analysis of techniques for electric vehicle battery prognostics and health management (PHM)". SAE Technical Paper. SAE Technical Paper Series. 2011-01-2247. doi:10.4271/2011-01-2247.
This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.