Abrasion Loss Prediction of Trapezoidal Thread Rod Based on Denoised Data

Abstract

There exists a direct wear influence of trapezoidal screw thread on its service life. It is necessary to determine the wear law to forecast the wear distance in order to design trapezoidal threaded rod. Therefore, the test bed is built to obtain a large amount of abrasion data at different speeds. And the wear data at 300 rpm is denoised by wavelet packet to smooth. And according to the data curvilinear trend, a variety of functions such as polynomial functions, power functions etc. are used to fit the data. Several fit goodness parameters show the best fitting function is a power function; and moreover, the fitting effect using the noised data is better than the original data by comparison. In addition, correlation analysis of data on both sides of the threaded rod proves that the two rods are of the high relevancy in wear. And during running-in preliminary stage, the difference between data from both side rods becomes gradually the biggest, but it tends to be stable since then.