Adaptive Model Predictive Control as a Prospect for Control of Machine Tools with Significant Flexibility

In the presented paper technique for derivation of adaptive model-based predictive control is introduced. The derived controller is built up by using the methods of subspace identification, receding window technique and model-based predictive control. The presented control concept is equipped by inspection mechanism increasing reliability of the concept. The controller is meant to be used in the area of machine tools, where machines with significantly reduced masses are taking the new trend. Those machines can easily suffer from low stiffness or nonlinearities due to new kinematic structures. For such machines presented controller could be an alternative replacing the cascade controller that has been used in the area of machine tools for decades. Capabilities of the new controller are evaluated by simulation on a simple machine tool axis with reduced masses represented by a two-mass model. Results of simulation using derived adaptive predictive controller are compared with results obtained using typical cascade controller.