MODEL BASED FAULT DIAGNOSIS AND SUPERVISION OF THE MAIN AND FEED DRIVES OF A FLEXIBLE MILLING CENTER

The early detection and localization of faults in machines and drives is of primary interest to make further improvement of the reliability and saftey. This paper presents a new approach in fault diagnosis and supervision of main and feed drives with elastic components. The supervision of drives with additional sensors for temperature, pressures or vibrations is usually expensive. Currently, internal process faults are only detected partially and at a rather late stage, by generating alarms if certain limits of the measured signals are exceeded or limit switches stop the machine tool. A process model based approach for fault diagnosis was developed. Dynamic models of each drive are tried and together with easily measurable signals of the drive, like current and speed, physical parameters are estimated. Changes of those process parameters are the symptoms for a diagnostic inference mechanism. Experimental results with least squares parameter estimation in the form of disrete square root filters are shown.