Data-driven quality prognostics for automated riveting processes

摘要: Technologies based in robotics and automatics are reshaping the aerospace industry. Aircraft manufacturers top-tier suppliers now rely on to perform most of its operational tasks. Over years, a succession implemented mobile robots has been developed with mission automating important industrial processes such as welding, material handling or assembly procedures. However, despite progress achieved, major limitation is that process still requires human supervision an extensive quality control process. An approach address this integrate machine learning methods within The idea develop algorithms can direct manufacturing experts towards critical areas requiring control. In paper we present application concrete problem involving riveting machine. proposal consists intelligent predictive model be integrated existing real time sensing pre-processing sub-systems at equipment level. framework makes use several data-driven techniques for feature engineering, combined accurate algorithms, validated through k-folds cross validation technique which also estimates prediction errors. able classify nominal anomalous according real-world data set design requirements data. Several compared linear regression, nearest neighbor, support vector machines, decision trees, random forests extreme gradient boost. Results obtained from case study suggest proposed produces predictions meet standards.