Intelligent Manufacturing Model and Technology Paper Recommendations

As the goal of Intelligent Manufacturing 2025 approaches, enterprises are joining the ranks of intelligent manufacturing maturity assessment to understand their own capability levels. However, due to the complexity of the intelligent manufacturing maturity assessment standards, enterprises lack an understanding of industry levels, leading to hasty applications, wasting their own time while occupying a large amount of assessment resources. In light of this, this paper designs a new assessment process, reconstructing the entire assessment process using text processing algorithms. By utilizing the intelligent manufacturing maturity assessment standards in national standard documents as the training set, an intelligent assessment algorithm combining pre-trained language models and text neural networks (BERT+TextCNN) replaces manual assessment. Validation on real enterprise intelligent manufacturing datasets shows that when the BERT+TextCNN assessment model uses convolution kernels of [2,3,4], iterates 6 times, and has a learning rate of 3e-5, the accuracy of assessing intelligent manufacturing maturity reaches 85.32%. This indicates that the designed assessment method can accurately assist enterprises in completing self-assessment of intelligent manufacturing maturity, helping them understand their own intelligent manufacturing capability levels and formulate correct development directions.

Components of Intelligent Manufacturing Maturity Model

Intelligent Manufacturing Maturity Assessment Process

Comparison of Specific Assessment Processes Before and After Reconstruction

BERT+TextCNN Maturity Assessment Model

Comparison Table of Various Evaluation Models

Comparison of Confusion Matrices of Assessment Models

Time and Personnel Comparison Required for Assessment Under the Same Level of Assessment in the Same Enterprise

Zhang Gan (1997-)

Male, from Nanyang, Henan, Master’s student at Nanjing University of Science and Technology, research direction: intelligent manufacturing, data analysis and mining, digital design and manufacturing, etc.

E-mail:[email protected]

Yuan Tangxiao (1993-)

Male, from Lianyungang, Jiangsu, PhD student at Nanjing University of Science and Technology, research direction: intelligent manufacturing theory, data collection, algorithm application risk assessment, etc.

E-mail:[email protected]

Wang Huifen (1965-)

Female, from Suzhou, Jiangsu, Professor at Nanjing University of Science and Technology, PhD, research direction: intelligent manufacturing, digital design and manufacturing, digital twin, knowledge base and knowledge management, etc., Corresponding Author.

E-mail:[email protected]

Liu Linyan (1985-)

Female, from Suzhou, Jiangsu, Associate Professor at Nanjing University of Science and Technology, PhD, research direction: digital design and manufacturing, intelligent manufacturing, digital twin, knowledge base and knowledge management, etc.

E-mail:[email protected]

Zhang Gan, Yuan Tangxiao, Wang Huifen, Liu Linyan. Intelligent Manufacturing Maturity Assessment Method Based on BERT and TextCNN [J]. Computer Integrated Manufacturing Systems, 2024, 30(3):852-863.

DOI:10.13196/j.cims.2023.0639

Scan to View Full Text