Semi-supervised learning for data-driven soft-sensing of biological and chemical processes

Esche, Erik; Talis, Torben; Weigert, Joris; Rihm, Gerardo Brand; You, Byungjun; Hoffmann, Christian; Repke, Jens‐Uwe

doi:10.1016/j.ces.2022.117459

Cited by 12 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Carneiro [33], Wang [34] 2021-2023 Prediction of steel properties Zou [35], Feng [36], Liu [37], Wang [38], Qian [39] 2021-2023 Prediction of molten steel composition Wang [40] 2022 Energy efficiency Lee [41] 2021 Motor equipment load Huang [42], Yu [43] 2022-2023 Modeling and prediction of inventory change Zhou [44], Esche [45], Zhu [46], Li [47], Bouaswaig [48],…”

Section: Review Of Dynamic Problems In Complex Industrial Processesmentioning

confidence: 99%

Dynamic Operation Optimization of Complex Industries Based on a Data-Driven Strategy

Tian,

Zhao,

Xie

et al. 2024

Processes

View full text Add to dashboard Cite

As industrial practices continue to evolve, complex process industries often exhibit characteristics such as multivariate correlation, dynamism, and nonlinearity, making traditional mechanism modeling inadequate in terms of addressing the intricacies of complex industrial problems. In recent years, with advancements in control theory and industrial practices, there has been a substantial increase in the volume of industrial data. Data-driven dynamic operation optimization techniques have emerged as effective solutions for handling complex industrial processes. By responding to dynamic environmental changes and utilizing advanced optimization algorithms, it is possible to achieve dynamic operational optimization in industrial processes, thereby reducing costs and emissions, improving efficiency, and increasing productivity. This correlates nicely with the goals set forth by conventional process operation optimization theories. Nowadays, this dynamic, data-driven strategy has shown significant potential in complex process industries characterized by multivariate correlations and nonlinear behavior. This paper approaches the subject from a data-driven perspective by establishing dynamic optimization models for complex industries and reviewing the state-of-the-art time series forecasting models to cope with changing objective functions over time. Meanwhile, aiming at the problem of concept drift in time series, this paper summarizes new concept drift detection methods and introduces model update methods to solve this challenge. In addressing the problem of solving dynamic multi-objective optimization problems, the paper reviews recent developments in dynamic change detection and response methods while summarizing commonly used as well as the latest performance measures for dynamic multi-objective optimization problems. In conclusion, a discussion of the research progress and challenges in the relevant domains is undertaken, followed by the proposal of potential directions for future research. This review will help to deeply understand the importance and application prospects of data-driven dynamic operation optimization in complex industrial fields.

show abstract

Section: Review Of Dynamic Problems In Complex Industrial Processesmentioning

confidence: 99%

Dynamic Operation Optimization of Complex Industries Based on a Data-Driven Strategy

Tian,

Zhao,

Xie

et al. 2024

Processes

View full text Add to dashboard Cite

show abstract

“…Jin et al proposed an ensemble evolutionary optimization-based pseudo-labeling method (EnEOPL), wherein the Gaussian process regression (GPR) was assigned as base learner, being further enhanced by an ensemble framework to optimize and generate the pseudo-labels [23]. Esche et al proved that when the time interval between two individual samples is noticeably too large, the SSL effect delivered by proposed deepkernel learning is significant, as illustrated in Williams-Otto simulation and bioethanol production process [24]. However, these methods only utilized a single base learner, of which the monotone learning characteristic can obscure the real X-Y mapping that could be more perplexed underlying the process, thus easily leading to under-fitting or over-fitting issues during the modeling.…”

Section: Introductionmentioning

confidence: 99%

Collaborative Multiple Players to Address Label Sparsity in Quality Prediction of Batch Processes

Zhao,

Zhang,

Zhu

et al. 2024

Sensors

View full text Add to dashboard Cite

For decades, soft sensors have been extensively renowned for their efficiency in real-time tracking of expensive variables for advanced process control. However, despite the diverse efforts lavished on enhancing their models, the issue of label sparsity when modeling the soft sensors has always posed challenges across various processes. In this paper, a fledgling technique, called co-training, is studied for leveraging only a small ratio of labeled data, to hone and formulate a more advantageous framework in soft sensor modeling. Dissimilar to the conventional routine where only two players are employed, we investigate the efficient number of players in batch processes, making a multiple-player learning scheme to assuage the sparsity issue. Meanwhile, a sliding window spanning across both time and batch direction is used to aggregate the samples for prediction, and account for the unique 2D correlations among the general batch process data. Altogether, the forged framework can outperform the other prevalent methods, especially when the ratio of unlabeled data is climbing up, and two case studies are showcased to demonstrate its effectiveness.

show abstract

“…It can address the issue of the data imbalance between labeled and unlabeled data in industrial data sets. Therefore, the method of semisupervised can improve the prediction performance of the soft sensor model. − However, the applications in wastewater treatment processes are still limited and have not received sufficient attention. Additionally, as the operating conditions change, the prediction model performance usually degrades.…”

Section: Introductionmentioning

confidence: 99%

Modeling an Adaptive Hybrid Soft Sensor with Co-training Learning toward Applications in Wastewater Treatment

Wu,

Tang,

Jin

et al. 2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

In wastewater treatment processes, sampling delays, resulting from difficult-to-measure but important variables, usually shape wastewater treatment processes (WWTPs) as multirate systems. In addition, when soft sensor models are applied to the wastewater treatment process, the performance of the model prediction always degrades because of the time-varying characteristics, uncertainty in process variables, and inaccurate measurement instruments. In this paper, an adaptive hybrid soft sensor model with semisupervised learning is proposed to deal with multiple rate issues in the WWTPs. First, the method uses a semisupervised learning approach, called the co-training framework algorithm, to refine useful information hidden in partially sampling-rate-inconsistent data and then solves the multirate system problem with the optimized model. Second, two different types of regression models, Recursive Partial Least Squares and Recursive Extreme Learning Machine, are used to model and predict the hard-to-measure variables as well as to update the prediction model adequately and finally to solve the performance degradation of the soft sensor model. Additionally, we validate the proposed method with two cases, namely, a simulation data set and a full-scale WWTP data set. The experimental results show that the proposed method in this paper can achieve a shorter time consumption and better prediction accuracy than standard soft sensor models.

show abstract

Semi-supervised learning for data-driven soft-sensing of biological and chemical processes

Cited by 12 publications

References 34 publications

Dynamic Operation Optimization of Complex Industries Based on a Data-Driven Strategy

Dynamic Operation Optimization of Complex Industries Based on a Data-Driven Strategy

Collaborative Multiple Players to Address Label Sparsity in Quality Prediction of Batch Processes

Modeling an Adaptive Hybrid Soft Sensor with Co-training Learning toward Applications in Wastewater Treatment

Contact Info

Product

Resources

About