Sub-macroscopic inclusion classification in bearing steels based on LFCN and ultrasonic testing

Abstract With the improvements in science and technology, the demand for advanced steel with excellent performance has gradually increased. Therefore, the evaluation of steel internal cleanliness is an important indicator for the evaluation of material quality. Sub-macroscopic inclusions, which size from 50 μm to 400 μm and cannot be detected under the domestic and international bearing steel testing standard, are bound to seriously affect the quality, stability and service life of bearing steel. Hence, research into inclusion control technology has gradually attracted attention in academia and the industrial manufacturing field. In this paper, we propose an end-to-end long short-term memory fully convolutional network classification model, and verify its effectiveness on a large-scale sub-macroscopic inclusion signal data set collected by ultrasonic experiments. To the best of our knowledge, this study is the first in this field that has acquired such a large amount of experimental sub-macroscopic signal data and that has solved the classification task using a fully convolutional network. In particular, our framework can accurately detect the features of sub-macroscopic inclusions, which meets the urgent need of the metallurgical industry. The accuracy rate of the proposed model is 88.97%, which is a state-of-the-art experimental result among other strong time-series classifiers.