News Technology
Optimising paint lines with machine learning
Study provides clear recommendations for action to improve surface quality using machine learing in automated paint shops.
The accuracy of the paint finish determines product quality, especially in automotive manufacturing. Since large quantities are primarily painted by robots in this industry, it is crucial to optimally coordinate the parameters. With increasing production requirements, conventional spray patterns are no longer sufficient, leading to the development of new spray pattern methods and optimisation algorithms for high-performance robot-assisted painting.
A research group at the Technical University in Thandalam, India, has recently developed a study for paint shop operators that enables them to optimise their processes and manufacture high-quality products while at the same time increasing time efficiency. To do this, they applied the Taguchi Design of Experiments (DoE) method. Developed by Japanese engineer Genichi Taguchi, this method of experimental design serves to make processes such as painting robust against disruptive influences and to systematically improve quality without having to test all possible combinations of parameters. This saves time and money by reducing the number of trials and quickly identifies critical influencing factors.
The researchers investigated spray distance, pressure, temperature, humidity, speed and viscosity as process variables in robotic painting. The experiments were carried out with an industrial robot and statistically evaluated using variance analysis (ANOVA tests) and regression calculations. It was found that viscosity in combination with temperature in particular has a decisive influence on fluctuations in layer thickness, while speed and temperature together determine surface roughness. The prediction model showed a high degree of accuracy, based on the R2 values achieved of 0.9224 for the measurement of surface roughness and 0.9707 for the determination of layer thickness deviation. The particular benefit of the study is that it provides clear recommendations for action beyond the theoretical findings.
More information: www.rajalakshmi.org