Integrated Diagnostic System for Mechanical Drives
Project duration
2009 - 2012
Project Category
SLO
Contact Information
Partners
- Institut Jožef Stefan (Slovenia)
- VTT Manufacturing Technology (Finland)
- University of Cranfield (United Kingdom)
- UTIA Institute of Information Theory and Automation, Department of Adaptive Systems (Czech Republic)
Because of wear, material stress and environmental influences, mechanical drives are more inclined to failures than any other item of equipment. Unexpected failures can result in partial or total breakdown of a production line, destroyed equipment and even catastrophes. That is why proper maintenance of such equipment is so important.
Project presentation
The reason reside in inappropriate maintenance paradigms in which still prevailing is reactive (post mortem) technology and (in some cases) preventive strategies at best. However, it turns to be very costly, as surveys over the last 20 years reveal that direct maintenance costs in European industry amount to 4 - 8% of the whole companies' income. In power generation sector these figures go even higher, up to 11% [Mechefske]. In addition, the indirect costs caused by degraded product quality, reduced production efficiency, loss of customers etc. are at least of the same range of magnitude.
A way to reduce the figures above is to abandon the current maintenance paradigms (reactive and preventive) and make room for cost-efficient condition-based (predictive) maintenance. Trends in the world clearly follow this direction. For example, manufacturers of advanced process equipment already provide a new generation of products with embedded diagnostic solutions for automatic on-line condition monitoring.
However, an open problem in industry is that - according to the estimates about Slovenia - over 95% of installed drives belong to the older generation with no embedded diagnostic functionality. This means that most of them are poorly monitored or even not monitored at all. As many of such machines will still be operating for some time, an upgrade with low-cost intelligent condition monitoring module would significantly improve surveillance, reduce maintenance costs, decrease failure rates and increase equipment availability. Namely, timely localization of the root cause would allow for more efficient pro-active maintenance as well as less costly and better planed fault accommodation.
Intensive research in the area of fault diagnosis and prognosis has been running for several decades. However, available commercial solutions are still mainly tailored for the particular classes of drives and are usually limited to fault diagnosis (without prognosis). Some additional weaknesses extend to: (a) high prices, (b) utilization of a restricted set of diagnostic techniques (mainly analysis of vibrations or oil analysis) and (c) in some cases limited capability of fault localization.
Project contribution
The aim of the project is to develop hardware and software prototype called Diagnostic and Prognostic Processor (DPP) for rotational machines and drives. Main innovative features of the system are:
a. the ability to timely detect, isolate and identify incipient faults,
b. the ability to predict the residual life time of the drive,
c. insensitivity to the variable operating conditions and external disturbances,
d. simple implementation owing to the of self-tuning options,
e. simple integration with the process information system and
f. almost no need for wiring due to wireless communication.
Economic price range will make the system affordable for a broad range of users in manufacturing, energy and transport sector, which are the most important for the Slovenian economy.
The system will allow fusion of features from diverse sources like vibrations sensors, on-line oil analyzers, noise sensors, thermal sensors and motor current and voltage. Feature extraction will rely on advanced signal processing methods. Clear relationships between signal processing and in-deep insight into the tribological processes will contribute notable added-value to the project. Fault localization and prognosis of the residual life span will be based on the probabilistic models derived from available experimental data. Operating version of the prototype will be built on a DSP platform, which guarantees high performance and cost-effective industrial solutions. System validation will be done on a laboratory benchmark rig.
Reference
Mechefske, CK, (2005) Machine Condition Monitoring and Fault Diagnosis, Vibration and Shock Handbook, Edited by Clarence W. de Silva, CRC Press, Taylor and Francis Group, Boca Raton, Florida, USA, Chapter 25, p25-1 to 25-35.