SKF has developed an "expert system" for identifying the causes of bearing failures, thus helping users to avoid recurring bearing problems. Although the BearingDetective is the not the first expert system in this field, it works in a different way which SKF says overcomes the shortcomings of earlier attempts.

Previous systems were usually structured as decision trees that led from the symptoms to the possible causes of bearing failures. The BearingDetective works in the opposite way — from causes to symptoms.

At the heart of the system is a wealth of knowledge from basic rolling bearing principles, to practical engineering and application results. There are four node categories, connected by causal relationships. They are:

• Conditions (speeds, bearing types, load, temperatures and so on);

• Internal mechanisms (such as sliding contacts and lubricant film disruptions);

• Failure modes (including sub-surface initiated fatigue, and fretting corrosion); and

• Observed symptoms (such as rust, spalling and discolouration).

By applying probability analysis to data supplied by the user and correlating this with the basic principles and the knowledge database, the system produces a list of possible causes of a bearing problem, ranked in order of probability.

Other News:
SKF Expert System Helps to Prevent Bearing Failures
Using Vibration Analysis to Detect Early Failure of Bearings
FAG Joplin Plant Released from Long-Running Pollution Case
Linear Actuators for High-Throughput Machines
Minebea Develops LED Backlight for 15-inch PC Monitors
Ohio Steelmakers Make Big Investments
NN,INC.Reports First Quarter 2007 Results
Westwind to Exhibit at JPCA Show 2007