Technical gazette, Vol. 29 No. 1, 2022.
Original scientific paper
https://doi.org/10.17559/TV-20190710141137
Robotic Grinding Process of Turboprop Engine Compressor Blades with Active Selection of Contact Force
Andrzej Burghardt
orcid.org/0000-0002-7358-4866
; Rzeszow University of Technology, Faculty of Mechanical Engineering and Aeronautics, Department of Applied Mechanics and Robotics, al. Powstańców Warszawy 12, 35-959 Rzeszów, Poland
Dariusz Szybicki
; Rzeszow University of Technology, Faculty of Mechanical Engineering and Aeronautics, Department of Applied Mechanics and Robotics, al. Powstańców Warszawy 12, 35-959 Rzeszów, Poland
Piotr Gierlak*
orcid.org/0000-0003-4545-8253
; Rzeszow University of Technology, Faculty of Mechanical Engineering and Aeronautics, Department of Applied Mechanics and Robotics, al. Powstańców Warszawy 12, 35-959 Rzeszów, Poland
Krzysztof Kurc
; Rzeszow University of Technology, Faculty of Mechanical Engineering and Aeronautics, Department of Applied Mechanics and Robotics, al. Powstańców Warszawy 12, 35-959 Rzeszów, Poland
Magdalena Muszyńska
; Rzeszow University of Technology, Faculty of Mechanical Engineering and Aeronautics, Department of Applied Mechanics and Robotics, al. Powstańców Warszawy 12, 35-959 Rzeszów, Poland
Abstract
The work presents a robotic system for grinding the blades of a turboprop engine compressor. The proprietary conceptual solution includes a data acquisition system based on a robotic 3D scanner, a neural decision system and a robot performing a grinding process with force control. The contact force of the tool to the blade was assumed as a variable and controlled process parameter. A neural network was used to generate the contact force on the basis of measured machining allowances on the blade. A virtual grid of several dozen regularly spaced points was placed on the surface of the blade. The neural network was learned the allowance-force dependence for the selected points, making it possible to select the proper contact force on the surface to be machined. The developed algorithm for the process of robotic grinding of the blades takes into account the necessity of ongoing quality control of the processing and the introduction of corrections in the process.
Keywords
blade grinding; force control; neural network; robotization
Hrčak ID:
269299
URI
Publication date:
15.2.2022.
Visits: 2.184 *