Session: 02-13-02: Industry 4.0 Aspects

Paper Number: 70682

Start Time: Tuesday, 01:20 PM

70682 - Tool Remaining Useful Life Prediction in Robotic Machining of Composite Materials Based on Mechanical Vibrations 

The development of materials and methods used in the aircraft manufacturing industry has been advancing fast in order to provide a reliable and light aircraft. In this sense, the use of composite materials becomes indispensable, meanwhile, the processing of this kind of material must be studied to obtain the higher manufacturing efficiency and the best quality of the final product.

So, the focus of this study is to create a remaining useful life prediction model for the tools used on the machining of composite materials with robotic manipulators. This task is performed by monitoring and analyzing the mechanical vibrations of the motor assembly and the cutting tool, hence optimizing the usage of these tools, then reducing the consumption of this material and ensuring the quality and surface integrity of the finished parts. To make it possible, a proper algorithm has been developed in order to simulate and evaluate the prediction method based on the digital manufacturing concept. The self-awareness of the process is improved by combining signal processing algorithms and statistical techniques to assist the constant monitoring of the tool wear. In this sense, a digital model is constantly updated aiming the optimization of the cutting process.

For the proposed research, a three-axis low-cost ADXL335 accelerometer is installed on the motor spindle that drives the milling cutter during the machining process. Assisted by Industry 4.0 trends of IoT and digitalization, a fast-prototyping ESP32 board is used to sample the mechanical vibration signals and provide an important signature of the process dynamics. Thus, the data is captured during the cutting of a sandwich panel made with facings of fiberglass cloth reinforced epoxy laminate and Gillcore® HD meta-aramid honeycomb core with the void filled with 3M® Scoth-Weld® structural void filling lightweight compound. Until now, several authors have been dealing with the tool remaining useful life prediction during the machining of metallic materials and using traditional CNC machines. However, the present study deal with the same issue regarding the tool life prediction, but looking for the behaviors of anthropomorphic robot for machining a composite raw material, besides making use of a low cost hardware to sample the mechanical vibrations during the manufacturing process.

The milling process is performed by a six-axis robot Stäubli TX200 suspended by a seventh longitudinal axis, which allows the robot to operate with its fixation and upside-down base. The AC motor, named as spindle, is coupled to the robot through a screwed and interchangeable support. The spindle allows the coupling of a system driven by compressed air where the cutting toolls attached. Using this setup, a sheet of composite material is placed just below the robot on a vacuum table that fixes the raw material to be machined. The offline program (OLP) determines the path of the robot through the part giving its geometry is loaded in the robot controller, once the robot executes the programmed movements and cutting parameters previously programmed.

After the process completion, a dimensional inspection of the generated part is performed, as well as its finish quality, the superficial integrity. Based on the digital manufacturing tool, the level of wear of the cutting tool is checked using a microscope and digital image processing software to help quantifying it. Then regression models are employed to analyze the evolution of time and frequency-domain features of the signals with respect to the tool wear.

Faced to related works mentioned above, the novel contribution of the present research reside on the development of remaining useful life estimators for cutting tools used on the robotic machining of composite materials, favoring the replacement of components, as well as the communication of anomalies that occurred during the process with users and manufacturers.

Presenting Author: José Otávio Savazzi Federal University of São Carlos (UFSCar)

Authors:

José Otávio Savazzi Federal University of São Carlos (UFSCar)
Sidney Bruce Shiki Federal University of São Carlos (UFSCar)
Gustavo Franco Barbosa Federal University of São Carlos (UFSCar)
David Guerra-Zubiaga Kennesaw State University