Joint project multi-axis milling

 

Model based predictive active force control strategy for multi-axis milling

 

Motivation

Milling is a flexible and highly dynamic machining process for manufacturing products with simple geometries up to complex free-form surfaces. In the milling process, the milling tool performs a rotary cutting motion on a firmly clamped workpiece. This cutting motion results in a functional geometry with high surface quality.

Industrially established control systems have so far focused on the control of the machining center’s parameters, for example the feed rate. However, these parameters have an indirect influence on the quality of the machined good. For example, excessively high feed rates lead to large forces in the tool's engagement zone. The consequences are a displacement of the tool and thus a reduction of the component quality and an increased tool wear. Conservative feed rates, on the other hand, result in a large process time and, therefore, in a reduction in productivity. A force control systems taking into account the production conditions online makes it possible to achieve an optimum process time while maintaining the required surface quality. However, in current force control systems the process parameters are identified in advanced. Influences such as tool wear are not taken into account explicitly, but implicitly by means of a conservative estimate, so that the cutting force in the engagement zone does not exceed a maximum force. This results in optimization potentials for milling with regard to process time and scrap rate.

 

Project Goals and Methods

The objective of this research project is to establish a model predictive force control strategy in multi-axis rough milling. The active force serves as the control variable, since it determines the resulting bending and torsional moment on the tool. The measured process force is acquired using a table dynamometer. The active force is determined by removing dynamic influences resulting from material removal, inertial forces and the transmission behavior of the measurement system acting on the measurement.

In a previous funded project, the success of a model predictive force control strategy in planar milling has been demonstrated. The maximum feed rate is calculated based on the engagement conditions for maintaining the allowed active force, and set by a model predictive controller (MPC). Unknown machine parameters are estimated using a linear Kalman filter. Using the force measurement, the parameters of an active force model are continuously adapted.

For multi-axis milling, the model has to be extended by additional degrees of motion of the tool and the machine table. The gravitational force becomes relevant due to the deflection of the measuring system from the horizontal. This exerts a non-linear influence on the measuring system. The extension of the model by the influence of inertial forces and gravity on the force measurement leads to a reformulation of the Kalman filter and the MPC. It is investigated whether linear methods are still suitable for controlling the nonlinear process. For this purpose, the complexity of the Kalman filter and MPR components is successively increased and the performance of the components is compared. In addition, an extended Kalman filter and a moving horizon estimator are developed, which estimate the active force in engagement based on the measured value of the force platform. The developed methods will be developed under consideration of strict real-time requirements and validated on a CNC machining center.

 

Innovations and Perspectives

The project enables the use of active force control for multi-axis milling and, thus, makes it possible to exploit existing potentials in terms of process time and component quality for complex milling tasks. The methods investigated increase productivity while maintaining high manufacturing quality and reducing the need for manual calibrations, even for complex processes.

 

Project partner
RWTH Aachen University, WZL - Laboratory for Machine Tools and Production Engineering