The National Science Center has announced the results of the Miniatura 5. This competition aims to provide financial support for a scientific activity to prepare a future research project to be submitted in NCN competitions or other national and international competitions. Among the winners, there was Dr.BEng. Marcin Paprocki was awarded funding for a research trip to the University of Belgrade and the Research and Development Institute LOLA Ltd.in Serbia as part of the scientific activity: Predictive Maintenance on a CNC machine with parallel kinematics (H-Bot) using the EtherCAT bus. Research trip. Congratulations!
The purpose of the research is to investigate the possibility of Predictive Maintenance (PM) in CNC machines (Computerized Numerical Control) with parallel kinematics - H-Bot machines. The accuracy of the H-Bot machine operation largely depends on the synchronous operation of the drives and the mechanical structure (including the tension of the toothed belts on both machine axes). Therefore, the quality of the synchronous movement of mechanical axes is influenced by both the communication bus (synchronous data exchange between the drives and the CNC controller) and the ability to detect backlash and elasticity in the machine's axes.
As part of the scientific activity, it is planned to research the laboratory standwith the H-Bot machine. In the control system of the H-Bot machine, it is intended to develop a software predictive maintenance module using elements of artificial neural networks. A modern industrial real-time bus - EtherCAT will be used for communication between the controller and other items of the numerical control system of machines, including sensors.
It is planned to carry out basic research in the field of detection of backlash and elasticity of the mechanical axes of the H-Bot machine during operation. This is especially important in the H-Bot design, where high dynamics and susceptibility to deformation characterize the toothed belt's movement.
It is intended to develop a software module using elements of artificial neural networks (CNN networks) in the control system. This module will process the feedback from the sensors (strain gauges and accelerometers) to detect undesirable vulnerabilities of the mechanical system. The correct operation of the module will be ensured by cyclical learning and verification of the implemented CNN artificial neural network in the CNC controller system. The network's learning will be carried out fromthe sensors' receivedinformation. In particular, research will be carried out on the possibility of detecting unwanted contour errors resulting from the toothed beltcompliance.
Scientists from the Faculty of Mechanical Engineering of the University of Belgrade and the Research and Development Institute LOLA Ltd. in Serbia are interested in this research. These institutions have extensive experience in constructing CNC machines(machine tools, machining robots - including delta kinematics). These institutionsare interested in joint research in developing and implementing advanced control algorithms and predictive maintenance in CNC machine control systems.
As part of the MINIATURE 5 grant, a one-week researchtrip to these institutions is planned to exchange experiences and outline common research areas. As a result of the cooperation, it is intended to disseminate joint research results in the form of at least one scientific article. For the performance of the above scientific activity, Dr. Marcin Paprocki received funding of PLN 6,490.
