PhD Defence by Dimitrios Papageorgiou

High Accuracy Nonlinear Control and Estimation for Machine Tool Systems
By Dimitrios Papageorgiou

Professor Mogens Blanke, DTU Electrical Engineering
Associate Professor Hans Henrik Niemann, DTU Electrical Engineering
Dr.-Ing. Jan H. Richter, Siemens AG

External Examiners:
Professor Ilmar Santos, DTU Mechanical Engineering
Professor Christopher Edwards, Exeter University
Professor Tor Arne Johansen, Norwegian University of Science and Technology

Associate Professor Roberto Galeazzi, DTU Electrical Engineering

All are welcome.

High accuracy axis positioning in machine tool systems is essential for ensuring that workpiece tolerances are maintained. Several mechanical phenomena that develop over time due to wear and tear, such as incipient friction and backlash degrade the performance of conventional control loops and, by extension, the overall quality of the manufactured products. 

This project investigated various positioning algorithms for machine tool systems based on sliding-mode and adaptive methodologies, while utilizing two different architectures. The control designs were tested on a real drive-train system with Siemens equipment when subject to unknown and increasing friction. 

Systematic evaluation of the experimental data using quantitative and qualitative indices facilitated the comparison of the considered methods. The results showed that the majority of the nonlinear controllers have superior friction-resilience properties in comparison to the state-of-the-art P-PI solutions used in the industry. 

Robust estimation of backlash in machine tools was also addressed in this research project. A modular estimation scheme for online identification of developing deadzones between motor and load was designed based on sliding mode and adaptive principles. 

The proposed algorithm performance was experimentally evaluated in various scenarios with different levels of uncertainty. The results showed that the developed method was able to provide sufficiently accurate estimates of the deadzone width, such that can be used in backlash compensation algorithms and equipment wear assessment.


Mon 02 Oct 17
13:00 - 17:00


DTU Elektro


Technical  University of Denmark
Lyngby Campus
Building 306, Aud. 37
2800 Lyngby
19 JUNE 2018