Consideration of the use of artificial intelligence methods for determining the friction coefficient of lubricated sliding bearings

Abstract

Sliding bearings’ main advantages over the ball bearings are it’s load capacity and longevity. If hydrodinamic lubrication conditions are met then the sliding bearings will work forever. This is especially important when dealing with high rotation rate environments, where ball bearings see very limited use due to their inferior load capacity and longevity. The focus of this study is the examination and determination of coefficient of friction values in sliding bearings made of tin based white metal alloys conditioned with hydrodinamic lubrication. A sliding bearing is based on a kinematic pair made of a steel axle and a braided alloy TEGOTENAX V840 made by Ecka Granules. The chemical composition of the this alloy is 88.7% Sn, 7.6% Sb and 3.7% Cu. The examination was performed using one oil lubricant. Experimental data has been used for creating an adaptive neuro-fuzzy inference system model. ANFIS model provides an estimation of coefficient of friction values in function of load. Based on the ANFIS model an analytical expression, used for connecting load values with coefficient of friction values, was defined. This analytical expression is suitable for engineering applications and purposes.