Vibration based preload estimation in machine tool spindles

Prevention of catastrophic bearing failures caused by excessive thermally induced preload is one of the key issues in the design and operation of high speed spindles. Temperature monitoring and shutting the spindle off in hazardous conditions is at present the most common method of avoiding bearing seizure. However, due to a rapid heat built-up and measurement delay, this solution in not reliable. An improved protection can be achieved either by custom systems with springs or hydraulic actuators that maintain a steady preload, or by monitoring the instantaneous preload and appropriate, predictive adjustment of machining conditions. While the former approach increases the cost of spindles and degrades their potential reliability, monitoring and proactive control of the actual preload eliminates these disadvantages. Its feasibility is predicated upon fast, reliable and accurate in-process preload estimation. A suitable estimation algorithm proposed in this paper employs vibrations of the spindle housing measured by means of accelerometer(s) and analyzed in the context of mechanical multi-degree-of-freedom model of the spindle assembly treated as a component of the entire machine tool. Representative experimental results are presented. (C) 2001 Published by Elsevier Science Ltd.