Fuzzy prediction of rolling bearing friction torque using information poor system theory

Friction torque of rolling bearings with uncertain fluctuation belongs to a poor information system,wherein both probability distribution and trend are unknown.This impedes small sample analysis and overall understanding of friction torque of rolling bearings.For this end,empirical probability density function and fuzzy prediction model of rolling bearings for space applications were proposed,and upper boundary of mean friction torque and maximum friction torque were predicted,starting from small samples based on fuzzy-set theory.Experiments on friction torque parameters of two kinds of rolling bearings HKTA and HKTB show that,at the 99%～100% confidence level,errors between results of prediction and experiment are very small,maximum absolute error is only 0.081 1 unit,and maximum relative error is only 10.165%,which could satisfy the requirements of space engineering.