Model-based Predictive Force Control in Milling – System Identification

AbstractManufacturing complex products, process monitoring and process control systems can increase productivity. Today, improvements in simulation technique enable computer based process set up and optimization reducing expensive trials. The performance of offline optimization strongly depends on the selected model and does not include uncertainties such as to tool wear or deviating material properties. Sensors help to reduce uncertainties and thus increase stability.Milling is one of the most flexible metal cutting processes. The process force is a very important evaluating process performance. Closed loop control strategies for force control in milling help increasing the process performance and protecting the tool, the workpiece and the machine. The varying transfer behavior of the controlled system is challenging. Adaptive control systems only react to changing transfer behavior and adjust the controller parameter. However, they cannot prevent high overshoots of the force.In this paper a new method to identify the controlled system is presented. The controlled system's transfer behavior is split into one part resulting from the machine tool and a second part describing the manufacturing process