TY - CHAP

T1 - Faulty Process Detection Using Machine Learning Techniques

AU - Dastgerdi, Amin Karimi

AU - Lange, Dirk

AU - Mercorelli, Paolo

N1 - Publisher Copyright: © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

PY - 2023/7/11

Y1 - 2023/7/11

N2 - These days, a shortage of resources leads to the adaptation of new digital manufacturing systems. These systems are playing a critical role in improving productivity and quality assurance in real-time process monitoring tasks and are being designed to monitor a production process and aim to save resources, energy, and time. This becomes even more vital in drilling processes to be as optimized and accurate as possible since it often comes last in the processes and a tiny error could lead to undesirable drilling part breakage. It could be even more costly in case the drilling part itself is in an inappropriate condition or damaged. Especially, if the part has a long production time and, in the end, the drilling or threading process goes wrong. Therefore, faulty processes must be forecasted sufficiently in advance to prevent damage and further costs. In this study, a set of machine learning algorithms have been used to develop an analysis of industrial manufacturing processes to detect faulty processes with the purpose of tool and machine protection as well as product quality assurance. The results of this study show that machine learning algorithms can detect faulty processes in the production process with high accuracy.

AB - These days, a shortage of resources leads to the adaptation of new digital manufacturing systems. These systems are playing a critical role in improving productivity and quality assurance in real-time process monitoring tasks and are being designed to monitor a production process and aim to save resources, energy, and time. This becomes even more vital in drilling processes to be as optimized and accurate as possible since it often comes last in the processes and a tiny error could lead to undesirable drilling part breakage. It could be even more costly in case the drilling part itself is in an inappropriate condition or damaged. Especially, if the part has a long production time and, in the end, the drilling or threading process goes wrong. Therefore, faulty processes must be forecasted sufficiently in advance to prevent damage and further costs. In this study, a set of machine learning algorithms have been used to develop an analysis of industrial manufacturing processes to detect faulty processes with the purpose of tool and machine protection as well as product quality assurance. The results of this study show that machine learning algorithms can detect faulty processes in the production process with high accuracy.

KW - Drill tool breakage

KW - Faulty process detection

KW - Machine learning

KW - Tool condition monitoring

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=85168996986&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/0f74059d-3666-3c6a-a64f-0f390f1ab81e/

U2 - 10.1007/978-981-99-2468-4_25

DO - 10.1007/978-981-99-2468-4_25

M3 - Article in conference proceedings

AN - SCOPUS:85168996986

SN - 978-981-99-2467-7

SN - 978-981-99-2470-7

T3 - Smart Innovation, Systems and Technologies

SP - 321

EP - 333

BT - Congress on Smart Computing Technologies

A2 - Bansal, Jagdish Chand

A2 - Sharma, Harish

A2 - Chakravorty, Antorweep

PB - Springer Singapore

T2 - Congress on Smart Computing Technologies, 2022

Y2 - 11 December 2022 through 12 December 2022

ER -