Intelligent sensor system for the disturbance variable invariant conditioning of residual stresses during machining of Ti-6Al-4V - Phase 2

在计划的研究项目中，Ti-6Al-4V的残余应力状态应该被设置为可靠的，并且独立于纵向车削过程中的任何刀具磨损。因此，实现了由声学传感器组成的传感器系统，其测量切屑分割频率作为刀具磨损状态的指标。通过控制切削速度、进给量、前角和切入角等工艺变量，对作为扰动量的刀具磨损进行补偿。通过使用软测量，结合基于有限元仿真的工艺知识的声学信号被处理，并且确定影响部件的表面层状态的热机械载荷谱。目标是通过控制上述过程变量，使循环载荷下的部件行为保持在一个持续的高水平，尽管磨损引起的刀具几何形状发生变化。第二个项目阶段的第一个目标是提高过程模型的性能，使其能够使用具有更复杂几何形状的工业切削刀具模拟切屑形成过程。第二个目标是增强信号处理方法的实时处理和评估（软测量）。软传感器是可靠的，并且可以识别干扰噪声，例如切屑碰撞、由金属切削液和刀具进给和出刀引起的噪声，并且可以将它们与软传感器相关的信号分离。第三个目标是调节加工后残余应力状态的过程控制。通过人工智能支持的系统辨识，软测量的动态行为以及初始材料和刀具条件被真实的实时地辨识。利用这些信息和工艺模型，可以控制和调整工艺参数（切削速度和进给），并在压电执行器的帮助下控制和调整前角和进入角，以在最终零件中获得所需的残余应力。

In the planned research project the residual stress state of Ti-6Al-4V should be set reliable and independent from any tool wear during longitudinal turning. Therefore, a sensor system consisting of acoustic sensors, which measure the chip segmentation frequency as an indicator for the tool wear state, is implemented. The tool wear as disturbance value is compensated by controlling the following process variables: cutting speed, feed, rake angle and entering angle. By using a soft sensor, the acoustic signal in combination with a FE-simulation based process knowledge is processed and the thermo-mechanic load spectrum which effects the component’s surface layer states is determined. The aim is to keep the component behavior under cyclic loading on a consistently high level despite a wear-induced tool geometry alteration, by controlling the mentioned process variables.The first objective of the second project phase is the improvement of the process model performance being able to model the chip formation process using industrial cutting tools with more complex geometry. The second objective is the enhancing of the signal processing method to real-time processing and evaluation (soft sensor). The soft sensor is reliable and can identify interfering noises such as chip collisions, noises caused by metal cutting fluids and tool infeed and outfeed and can separate them from the signals relevant for the soft sensors. The third objective is the process control to regulate residual stress states after machining. By an AI-supported system identification the dynamic behavior of the soft sensor and the initial material and tool condition are identified in real time. With this information and using the process model the process parameters (cutting speed and feed) and, with the help of piezoelectric actuators, the rake and entering angles are controlled and adjusted to obtain the desired residual stresses in the final part.