Tool condition monitoring in the dressing process through electromechanical impedance and machine learning

• Autor(a) principal: Conceição Junior, Pedro de Oliveira
• Data de Publicação: 2020
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/191936
• topic: Tool condition monitoring; Grinding; Dressing process; Electromechanical impedance; PZT; Machine learning; Monitoramento da condição da ferramenta; Retificação; Processo de dressagem; Impedância eletromecânica; Aprendizagem de máquina

The electromechanical impedance (EMI) has attracted increasing attention as an effective sensor monitoring technique for applications in many engineering sectors. Due to the considerable potential of lead zirconate titanate (PZT) diaphragm transducers in terms of excellent electromechanical coupling properties, low implementation cost and wide-band frequency response, this technique provides a new alternative approach for tool condition monitoring (TCM) in grinding processes competing with the conventional and expensive indirect sensor monitoring methods. This research work aimed to develop a new approach for TCM in dressing operation using the PZT-EMI-based technique in cooperation with machine learning algorithms. The research activities proposed in this work involved in diagnosing different wear and failure ranges on dressing tools through experimental tests of dressing operation into different dressing conditions. Further, the proposed system was tested for sensory position independence and for different types of dressing tools to ensure its capability for real-time application. The proposed approach was validated on the basis of the dressing tool condition information obtained from representative damage indices computed at different frequency bands. The proposed intelligent diagnosis system was able to select the most damage-sensitive features, such as damage classification and location, based on the optimal selection of the suitable frequency band. The best results showed less than 2% of general average errors to determine the actual tool condition. This thesis contributes to an effective monitoring system of the dressing operation capable of avoiding replacement of the dressing tool before the end of its service life as well as the operation being performed with damaged tools, since the proposed approach can identify different sates of the tool life span