Tool wear is one of the main issues in machining. Knowing the health state of the tool is of primary importance since it would permit to reduce down-times and to improve surface quality of the manufactured parts. In literature, one of the main challenges in this contest is the estimation of the residual tool life during non-stationary conditions. In this thesis an algorithm for properly monitoring the tool condition in milling is developed. The approach is based on the real time estimation of the Specific Cutting Coefficients that can be roughly considered not related to the adopted cutting parameters. It was demonstrated that a synthetic index, combination of the estimated cutting coefficients, can be used for monitoring the tool wear. In order to deal with unpredictable process and machine variability, a Growing Self Organizing Map is developed and integrated in the method. Tuned control charts are used for monitoring tool wear trend. This algorithm is shown to be efficient in tool-state monitoring. Out-of-control appears with a deviation with respect to direct measurements of tool wear of maximum 100 seconds, differently from a more classical approach where the out-of-control appears at the early stages of the process, as a consequence of high process variability.
L'usura utensile è uno dei principali problemi nelle lavorazioni. Conoscere lo stato di salute di un utensile è di primaria importanza poichè permette di ridurre i tempi di inutilizzo e di aumentare la qualità superficiale dei pezzi lavorati. In letteratura, una delle sfide principali in questo ambito è la stima della vita rimanente dell'utensile durante condizioni non stazionarie. In questa tesi verrà proposto un algoritmo per il monitoraggio in fresatura della condizione dell'utensile. L'approccio è basato sulla stima in tempo reale dei coefficienti di taglio che possono essere considerati indipendenti dai paramentri di processo adottati. Verrà dimostrato che un indice sintetico, ottenuto come combinazione dei coefficienti di taglio, può essere usato nel monitoraggio dell'usura utensile. Per rapportarsi con variabilità del processo e della macchina, verrà costruita una Growing Self Organizing Map e verrà applicata al metodo. Apposite carte di controllo verranno utilizzate per il monitoraggio dell'usura utiensile. Questo algoritmo si rivelerà efficace nel monitoraggio dello stato dell'utensile. I segnali di out of control si mostreranno con un errore di massimo 100 secondi rispetto alle misure dirette dell'usura, diversamente da un metodo più tradizionale in cui il segnale di out-of-control si presenterà alle fasi iniziali del processo, causa la grande variabilità del processo.
Online robust tool wear monitoring in milling under variable cutting conditions
Arfini, Andrea
2020/2021
Abstract
Tool wear is one of the main issues in machining. Knowing the health state of the tool is of primary importance since it would permit to reduce down-times and to improve surface quality of the manufactured parts. In literature, one of the main challenges in this contest is the estimation of the residual tool life during non-stationary conditions. In this thesis an algorithm for properly monitoring the tool condition in milling is developed. The approach is based on the real time estimation of the Specific Cutting Coefficients that can be roughly considered not related to the adopted cutting parameters. It was demonstrated that a synthetic index, combination of the estimated cutting coefficients, can be used for monitoring the tool wear. In order to deal with unpredictable process and machine variability, a Growing Self Organizing Map is developed and integrated in the method. Tuned control charts are used for monitoring tool wear trend. This algorithm is shown to be efficient in tool-state monitoring. Out-of-control appears with a deviation with respect to direct measurements of tool wear of maximum 100 seconds, differently from a more classical approach where the out-of-control appears at the early stages of the process, as a consequence of high process variability.File | Dimensione | Formato | |
---|---|---|---|
2021_12_Arfini.pdf
non accessibile
Descrizione: Tesi
Dimensione
39.92 MB
Formato
Adobe PDF
|
39.92 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/182648