Prognostics and health management of mechanical drives based on novel MEMS sensor networks

Principal investigator:

Prof. Dr. Jožef Vižintin, Center for Tribology and Thermical Diagnostics, Faculty of Mechanical Engineering, University of Ljubljana
Responsible investigator at the department: Prof. dr. Đani Juričić

Duration:

July 1, 2011 – June 30, 2014

Funding:

ARRS - Slovenian Research Agency, L2-4160 (C), co-financed by Litostroj Power d.o.o. and Kalmer d.o.o.

Abstract:

Mechanical drives are the most ubiquitous item of equipment in almost all industrial branches. Wear, excessive operational loads or errors in assembly might cause premature unexpected failures resulting in partial or total production downtime, damaged equipment or even loss of lives. Proper maintenance is therefore very important.

According to the ARTEMIS report, the direct cost of maintenance in EU is estimated 4%- 8% of the total sales turnover. Moreover, 30-50% of the expenditure is wasted through ineffective maintenance! The problem is that currently prevailing reactive (react-to-failure) and preventive (periodic) maintenance paradigms are outdated and need to be replaced with more costeffective predictive maintenance based on advanced diagnostic, prognostic and health management solutions (PHM). While diagnostics tends to determine condition of the component and isolate faults, the aim of prognostics is to assess the useful life of the asset. Health management refers to the ability to make intelligent decisions about maintenance actions. Reasons that keep companies reluctant to investments in PHM are still in high capital costs, installation difficulties, and the overall complexity of the currently available monitoring systems.

The aim of the project is to respond to these challenges and come up with the prototype of a versatile, easily manageable and radically low-cost platform labelled MEMS-PHM for prognostics and health management of electro-mechanical drives that will rely on cutting edge MEMS (microelectromechanical sensor) technologies. Strong motivation for the underlying project is fuelled by several key factors:


  • PHM solutions are just emerging on the market but tailored for special target assets (e.g. military aircrafts). There is obvious need for 'general purpose' PHM solutions that would be applicable to a broader range of operating machines.
  • Almost 40% of all the machinery operating at the time being fall in the power range 5- 100kW, hence representing valuable asset. According to a review done in USA only in 1% of this machinery the condition is monitored automatically all the time. The rest is poorly monitored or monitored at best periodically.
  • Advanced generation of micro-sensors including MEMS accelerometers have recently emerged on the market. They are characterized by extremely low cost, miniature size and reliability while preserving the accuracy, bandwidth, and robustness of the traditional sensors.

This project will conduct basic and applied research leading to the anticipated major results as follows:


  • A portfolio of low cost MEMS-PHM platforms able to perform diagnostics and prognostics tasks. Autonomous energy supply via energy harvesters, no cabling and wireless communication make ground for versatile PHM functionality at the cost as low as 1/10 of the costs needed by current condition monitoring technologies.
  • Innovative tool for application SW design and automatic code generation for MEMSPHM, hence reducing the development and configuration effort.
  • New algorithms for condition assessment and prognostics based on information fusion concepts, enabling the proposed system to gain information from various sensors like accelerometers, thermocouples, sensors for oil parameters etc.
  • Robust algorithms able to detect and isolate faults under non-stationary operating conditions and external disturbances.
  • Interface to the e-maintenance platforms.

The prototype versions of the system will be validated on laboratory motor-generator test rig and demo industrial installations, already been agreed with the companies supporting the project.