Title: A sequential cross-product knowledge accumulation, extraction and transfer framework for machine learning-based production process modelling
Abstract: ABSTRACTMachine learning is a promising method to model production processes and predict product quality. It is challenging to accurately model complex systems due to data scarcity, as mass customisation leads to various high-variety low-volume products. This study conceptualised knowledge accumulation, extraction, and transfer (KAET) to exploit the knowledge embedded in similar entities to address data scarcity. A sequential cross-product KAET (SeqTrans) is proposed to conduct KAET, integrating data preparation and preprocessing, feature selection (FS), feature learning (FL), and transfer learning (TL). The FS and FL modules conduct knowledge extraction and help address various practical challenges such as changing operating conditions and unbalanced datasets. In this paper, sequential TL is introduced to production modelling to conduct knowledge transfer among multiple entities. The first case study of auxetic material performance prediction demonstrates the effectiveness of sequential TL. Compared with conventional TL, sequential TL can achieve the same test mean square errors with 300 fewer training examples when facing data scarcity. In the second case study, balancing anomaly detection models were constructed for two gas turbines in the same series using real-world production data. With SeqTrans, the F1-score of the anomaly detection model of the data-poor engine was improved from 0.769 to 0.909.KEYWORDS: Machine learningfeature selectionfeature learningtransfer learninggas turbineauxetic material AcknowledgementsPart of this paper is derived from Jiarui Xie's master's thesis (Xie Citation2022) at the Department of Mechanical Engineering at McGill University. A more comprehensive analysis and an additional case study have been provided in this paper compared with the thesis.Disclosure statementNo potential conflict of interest was reported by the author(s).Data availability statementThe data that support the findings of the first case study, auxetic material performance prediction, are available from the corresponding author, Y. Zhao, upon reasonable request. Due to the nature of the second case study, gas turbine balancing anomaly detection, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.Additional informationFundingThis work is funded by Mathematics of Information Technology and Complex Systems (MITACS) Accelerate program [grant number IT13369]; McGill Engineering Doctoral Award (MEDA); and McGill University Graduate Excellence Fellowship Award. Financial supports from the National Research Council of Canada [NRC INT-015-1] for Mutahar Safdar are acknowledged with gratitude. This work is also supported by the Additive Design and Manufacturing Lab (ADML) at McGill University and Siemens Energy in Montreal.Notes on contributorsJiarui XieJiarui Xie is a Mechanical Engineering PhD student at McGill University. He is affiliated with Additive Design and Manufacturing Lab under Dr. Yaoyao Fiona Zhao's supervision. He specializes in the application of artificial intelligence in intelligent manufacturing, focusing on deep learning and feature learning. His current projects focus on data readiness in design and manufacturing, and diagnosis and prognosis of unmanned aerial vehicles.Chonghui ZhangChonghui Zhang is a PhD Candidate in Additive Design and Manufacturing Laboratory (ADML) at the Department of Mechanical Engineering in McGill University, in Montreal, Canada. His research interests include design, model reduction, and metamaterials. He is researching the applications of machine learning to support the inverse design of auxetic metamaterials.Manuel SageManuel Sage is a PhD Candidate in the Additive Design and Manufacturing Laboratory (ADML) at the Department of Mechanical Engineering in McGill University, in Montreal, Canada. His research interests include machine learning and reinforcement learning in general, and their applications to energy systems in particular. He is researching the applicability of deep reinforcement learning algorithms to control problems in hybrid energy systems.Mutahar SafdarMutahar Safdar is a PhD Candidate in Additive Design and Manufacturing Laboratory (ADML) at the Department of Mechanical Engineering in McGill University, in Montreal, Canada. His research interests include engineering informatics, design manufacturing, and digital twin. He is researching the applications of machine learning to transform metal additive manufacturing into a reliable and high-volume production technology.Yaoyao Fiona ZhaoDr. Yaoyao Fiona Zhao is an Associate Professor and William Dawson Scholar at the Department of Mechanical Engineering in McGill University, in Montreal, Canada. Her research expertise lies in the general field of design and manufacturing including the exploration of new design methods, developing efficient numerical simulation method for additive manufacturing processes, manufacturing informatics, application of machine learning in design and manufacturing, sustainable product development and intelligent manufacturing.
Publication Year: 2023
Publication Date: 2023-09-07
Language: en
Type: article
Indexed In: ['crossref']
Access and Citation
Cited By Count: 3
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot