Software prevents reoccurring bearing failures.
Press Release Summary:
Based on probabilistic network, BearingDetective decision support system is structured by modeling from possible causes to symptoms. Knowledge-based system searches its information base and provides suspected causes of bearing damage for application parameters that have been input into program; causes are ranked in order of highest probability to lowest. Network for bearing failure analysis has 4 node categories: conditions, internal mechanisms, failure modes, and observed symptoms.
Original Press Release:
SKF BearingDetective; Preventing Reoccurring Bearing Failures
SKF has developed an advanced 'expert' system for identifying bearing failure modes. This system will help customers to prevent reoccurring bearing damage and failures, thus saving on time and costs associated with unplanned machine stoppages.
SKF BearingDetective is a decision support system that allows more consistent, fast and accurate assessments of rolling bearing damage or failure. It is a knowledge-based system that searches its information base and offers a number of possible causes of bearing damage for application parameters that have been input into the program. The causes are ranked in order of highest probability rating (in %) to the lowest.
Based on a probabilistic network the system overcomes the shortcomings of previous expert systems that were often structured as decision trees that led from symptom to possible causes. SKF BearingDetective is structured by modeling in an opposite way to earlier methods; from possible causes to symptoms. By modeling in this way and attaching a degree of uncertainty to possible failure states, a much better fit is achieved with the physical phenomena that occur during bearing service life. With the aid of state-of-the-art computational intelligence this assessment process has been made to be user friendly and very fast.
At the core of the system is a wealth of knowledge gathered from basic rolling bearing principles to practical engineering and application results. The probabilistic network is a visual network in which nodes are connected by causal relationships, and probability calculations are applied. The network for bearing failure analysis has four node categories;
o conditions (speeds, bearing types, load, temperature etc)
o internal mechanisms (sliding contact, lubricant film disruption etc)
o failure modes (sub-surface initiated fatigue, fretting corrosion etc)
o observed symptoms (rust, discolouration, spalling etc)
By applying probability analysis to the input data and correlating with the basic principles and applications knowledge base, the probable causes are derived.
SKF BearingDetective is a web-enabled system allowing SKF engineers all over the world instant and flexible access when working with customers to assist in bearing failure analysis.
SKF, April 2003
For further information, please contact:
Colin T. Roberts, SKF Group Technical Press Coordinator
SKF Engineering and Research Centre, Netherlands,
Tel: +31 (0)30 60 75 608, e-mail: colin.roberts@skf.com
Wolfgang Gläntz, SKF Technical Press Germany
SKF GmbH, Schweinfurt
Tel +49 (0)9721 56 3343, e-mail: wolfgang.glaentz@skf.com
SKF, www.skf.com
