Towards an Automatic System to Spindle Faults Detection ⁎

Abstract

The diagnosis of faults has allowed to evolve the maintenance strategies in the industries, optimizing the production stops. In the case of machining systems, timely fault diagnosis avoids products that are out of specification and/or extreme damage. Optimum machining is highly dependent on the performance and condition of the spindle, within which the bearing system represents the mechanical components with the greatest likelihood of failure. The advances in the use of the Wavelet Transform (WT) was analyzed and a fault detection method for spindles was proposed. This method automatically detects the frequency range where most information of the fault is located and separates it from other noisy frequencies. Furthermore, faults can be detected at early stages. Early results, validated with experimental data, are promising for an automatic system.