Estimation of the load torque in a hobbing machine using effective power signals

Abstract Aviation requires highly reliable parts that call for stable and safe processes. These are often achieved using process-monitoring techniques. In order to not reduce the stiffness of the spindle, effective power measurements replace force measurements. However, the inevitable low-pass filter and unknown transmission behavior make these signals inferior to force signals. In order to compensate this, knowledge about the transfer behavior between force and effective power enables a more reliant and accurate process monitoring. This paper shows a method to determine this behavior with a model to calculate the process torque of a hobbing machine using simulation models and the effective power signal. The model is based on machine learning algorithms and uses an Artificial Neural Network (ANN) to determine the relation between the two quantities.