Adaptive soft sensor for online prediction and process monitoring based on a mixture of Gaussian process models

Grbić, Ratko; Slišković, Dražen; Kadlec, Petr

doi:10.1016/j.compchemeng.2013.06.014

Cited by 148 publications

(74 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The process consists of five major operations: reactor, product condenser, vapor-liquid separator, recycle compressor, and product stripper, as shown in Figure 9. This process contains 41 measured variables and 12 manipulated variables, and it includes 21 faults (Grbić et al, 2013), as shown in Table 2.…”

Section: Te Processmentioning

confidence: 99%

Fault Detection Based on Probabilistic Kernel Partial Least Square Regression for Industrial Processes

Xie

Zhang

Jia

et al. 2018

J. Chem. Eng. Japan / JCEJ

View full text Add to dashboard Cite

Section: Te Processmentioning

confidence: 99%

Fault Detection Based on Probabilistic Kernel Partial Least Square Regression for Industrial Processes

Xie

Zhang

Jia

et al. 2018

J. Chem. Eng. Japan / JCEJ

View full text Add to dashboard Cite

“…Kaneko et al (2011) proposed a method to estimate the relationships between applicability domains and the accuracy of prediction of soft sensor models. Grbić et al (2013) used the confidence region corresponding to ±2 standard deviations to evaluate the reliability of adaptive soft-sensors. In addition, the reliability of soft-sensors was discussed in the review paper by Kano and Fujiwara (2013).…”

Section: Reliability Check Of Predicted Probability Distributionmentioning

confidence: 99%

Prediction of Molten Steel Temperature in Steel Making Process with Uncertainty by Integrating Gray-Box Model and Bootstrap Filter

Ahmad

Kano

Hasebe

et al. 2014

J. Chem. Eng. Japan

View full text Add to dashboard Cite

Stable operation of a continuous casting process requires precise control of molten steel temperature in a tundish (TD temp), which is a container used to feed molten steel into an ingot mold. Since TD temp is implicitly controlled by adjusting molten steel temperature in the preceding secondary re ning process (RH temp), a model relating TD temp with RH temp is required. This research proposes a procedure to predict the probability distribution of TD temp by integrating a gray-box model and a bootstrap lter to cope with uncertainties of the process. The derived probability distribution is used not only to predict TD temp but also to evaluate the reliability of prediction. The proposed method was validated through its application to real operation data at a steel work, and it was con rmed that the developed model satis ed the requirements for its industrial application.

show abstract

“…[24], [26] uses adaptation of base models and adaptive weighting with [23] additionally introducing adaptive offset correction. Another local ensemble method, with a moving window and weights change AMs is described in [13].…”

Section: Related Workmentioning

confidence: 99%

Augmenting adaptation with retrospective model correction for non-stationary regression problems

Bakirov

Gabrys

Fay

2016

2016 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Abstract-Existing adaptive predictive methods often use multiple adaptive mechanisms as part of their coping strategy in nonstationary environments. We address a scenario when selective deployment of these adaptive mechanisms is possible. In this case, deploying each adaptive mechanism results in different candidate models, and only one of these candidates is chosen to make predictions on the subsequent data. After observing the error of each of candidate, it is possible to revert the current model to the one which had the least error. We call this strategy retrospective model correction. In this work we aim to investigate the benefits of such approach. As a vehicle for the investigation we use an adaptive ensemble method for regression in batch learning mode which employs several adaptive mechanisms to react to changes in the data. Using real world data from the process industry we show empirically that the retrospective model correction is indeed beneficial for the predictive accuracy, especially for the weaker adaptive mechanisms.

show abstract

Adaptive soft sensor for online prediction and process monitoring based on a mixture of Gaussian process models

Cited by 148 publications

References 31 publications

Fault Detection Based on Probabilistic Kernel Partial Least Square Regression for Industrial Processes

Fault Detection Based on Probabilistic Kernel Partial Least Square Regression for Industrial Processes

Prediction of Molten Steel Temperature in Steel Making Process with Uncertainty by Integrating Gray-Box Model and Bootstrap Filter

Augmenting adaptation with retrospective model correction for non-stationary regression problems

Contact Info

Product

Resources

About