Unsupervised Hybrid Models Integrating Deep Autoencoders and Process Controllers’ Models for Enhanced Process Monitoring and Fault Detection DOI
Mohammad Aghaee,

Stéphane Krau,

Ibrahim Melih Tamer

и другие.

Industrial & Engineering Chemistry Research, Год журнала: 2024, Номер 63(33), С. 14748 - 14760

Опубликована: Авг. 12, 2024

This paper introduces a novel hybrid process monitoring model that integrates long short-term memory autoencoders with controllers' models. The parameters of the are optimized by minimizing loss function, which combines mean square error (MSE) between controlled variables and their reconstructions from LSTM-AE model, along MSE manipulated obtained numerically implemented exactly priori known controller equations. effectiveness proposed method is evaluated on benchmark an industrial-scale penicillin as batch case study Tennessee Eastman plant under decentralized control strategy continuous study. A comparative analysis equivalent nonhybrid does not utilize equations, highlights superiority in fault detection. These improvements result use network fewer parameters, thus making it less susceptible to overfitting.

Язык: Английский

Integrated decision-making with adaptive feature weighting adversarial network for multi-target domain compound fault diagnosis of machinery DOI
Xuepeng Zhang, Jinrui Wang, Zongzhen Zhang

и другие.

Advanced Engineering Informatics, Год журнала: 2024, Номер 62, С. 102730 - 102730

Опубликована: Июль 24, 2024

Язык: Английский

Процитировано

6

A rolling bearing fault diagnosis method under insufficient samples condition based on MSLSTM transfer learning DOI Open Access
Ping Zhang, Debo Liu

Journal of Vibroengineering, Год журнала: 2025, Номер 27(1), С. 93 - 107

Опубликована: Янв. 22, 2025

It usually affects the accuracy and reliability of deep learning based intelligent diagnosis methods under condition insufficient samples. Existing for handling samples often have problems such as requiring rich expert experience or consuming a lot time. To solve above problems, rolling bearing fault method on multi-scale long-term short-term memory network (MSLSTM) transfer is proposed, which mainly consists an improved named MSLSTM learning. By introducing convolution operation into traditional LSTM to improve its drawback that only extracts single type feature information, leads poor diagnostic performance in noisy environments. Besides, pooling layer global average are replaced with avoid problem information loss. Subsequently, combined learning, fine tunes model parameters using small amount target domain data. Feasibility proposed verified through two kinds experiments. The has stronger extraction ability training efficiency compared other models.

Язык: Английский

Процитировано

0

A Perspective on Artificial Intelligence for Process Manufacturing DOI Creative Commons
Vipul Mann, Jingyi Lu, Venkat Venkatasubramanian

и другие.

Engineering, Год журнала: 2025, Номер unknown

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0

Spectroscopic Advances in Real Time Monitoring of Pharmaceutical Bioprocesses: A Review of Vibrational and Fluorescence Techniques DOI Open Access
Abhishek Mishra, Mohammad Aghaee,

Ibrahim Melih Tamer

и другие.

Spectroscopy Journal, Год журнала: 2025, Номер 3(2), С. 12 - 12

Опубликована: Апрель 1, 2025

The pharmaceutical industry has witnessed exponential growth in production volumes, driven by factors such as an aging global population and the COVID-19 pandemic. To meet demand for high product quality alongside increased productivity, there is a growing emphasis on developing innovative Fermentation Analytical Technology (FAT) Process (PAT) tools real-time performance monitoring, modeling, measurement, control. Building our earlier work involving in-line monitoring of Bordetella pertussis fermentations using fluorescence spectroscopy, this review explores compares applications vibrational spectroscopy bioprocess monitoring. We examine recent technological advancements ongoing challenges field. Various spectroscopic techniques are evaluated terms cost-effectiveness practical applicability, with particular focus promising, low-cost solution effective

Язык: Английский

Процитировано

0

Editorial overview: Digital design of pharmaceutical manufacturing processes DOI
Kimberley B. McAuley, Jonathan P. McMullen, Salvador García‐Muñoz

и другие.

Current Opinion in Chemical Engineering, Год журнала: 2025, Номер 48, С. 101108 - 101108

Опубликована: Март 4, 2025

Язык: Английский

Процитировано

0

Unsupervised Hybrid Models Integrating Deep Autoencoders and Process Controllers’ Models for Enhanced Process Monitoring and Fault Detection DOI
Mohammad Aghaee,

Stéphane Krau,

Ibrahim Melih Tamer

и другие.

Industrial & Engineering Chemistry Research, Год журнала: 2024, Номер 63(33), С. 14748 - 14760

Опубликована: Авг. 12, 2024

This paper introduces a novel hybrid process monitoring model that integrates long short-term memory autoencoders with controllers' models. The parameters of the are optimized by minimizing loss function, which combines mean square error (MSE) between controlled variables and their reconstructions from LSTM-AE model, along MSE manipulated obtained numerically implemented exactly priori known controller equations. effectiveness proposed method is evaluated on benchmark an industrial-scale penicillin as batch case study Tennessee Eastman plant under decentralized control strategy continuous study. A comparative analysis equivalent nonhybrid does not utilize equations, highlights superiority in fault detection. These improvements result use network fewer parameters, thus making it less susceptible to overfitting.

Язык: Английский

Процитировано

0