A Bayesian approach to design of adaptive multi-model inferential sensors with application in oil sand industry

Khatibisepehr, Shima; Huang, Biao; Xu, Fangwei; Espejo, Aris

doi:10.1016/j.jprocont.2012.09.006

Cited by 50 publications

(40 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This application highlights the need for multiple models and the ability to detect when a system has shifted from one operating mode to another. PCA could be used in an on-line manner for this purpose [7] as could a Bayesian approach [8]. …”

Section: Linear Regression Modelsmentioning

confidence: 99%

Modelling of a Naphtha Recovery Unit (NRU) with Implications for Process Optimization

Cluett

2018

Processes

View full text Add to dashboard Cite

Abstract:The naphtha recovery unit (NRU) is an integral part of the processes used in the oil sands industry for bitumen extraction. The principle role of the NRU is to recover naphtha from the tailings for reuse in this process. This process is energy-intensive, and environmental guidelines for naphtha recovery must be met. Steady-state models for the NRU system are developed in this paper using two different approaches. The first approach is a statistical, data-based modelling approach where linear regression models have been developed using Minitab ® from plant data collected during a performance test. The second approach involves the development of a first-principles model in Aspen Plus ® based on the NRU process flow diagram. A novel refinement to this latter model, called "withdraw and remix", is proposed based on comparing actual plant data to model predictions around the two units used to separate water and naphtha. The models developed in this paper suggest some interesting ideas for the further optimization of the process, in that it may be possible to achieve the required naphtha recovery using less energy. More plant tests are required to validate these ideas.

show abstract

Section: Linear Regression Modelsmentioning

confidence: 99%

Modelling of a Naphtha Recovery Unit (NRU) with Implications for Process Optimization

Cluett

2018

Processes

View full text Add to dashboard Cite

show abstract

“…The first strategy computes the final model output as the weighted combination of all the local models' outputs. The weights can be either equal for all the local models [31] or different for different local models [22], [23], [25]- [27], [39]. The second strategy is to switch to different local models online according to some model adaptation criteria, such as fuzzy membership [24] or posterior probability [10], [40], [41].…”

Section: Accepted Manuscriptmentioning

confidence: 99%

“…The schematic structure of an exothermic irreversible first-order CSTR, which are widely used for testing soft sensors' performance [27], [40], is shown in Fig. 3.…”

Section: A Continuous Stirred Tank Reactormentioning

confidence: 99%

See 1 more Smart Citation

Online soft sensor design using local partial least squares models with adaptive process state partition

Shao

Tian

Wang

et al. 2015

Chemometrics and Intelligent Laboratory Systems

View full text Add to dashboard Cite

We propose a soft sensing method using local partial least squares models with adaptive process state partition, referring to as the LPLS-APSP, which is capable of effectively handling time-varying characteristics and nonlinearities of processes, the two major adverse effects of common industrial processes that cause low-performance of soft sensors. In our proposed approach, statistical hypothesis testing is employed to adaptively partition the process state into the unique local model regions each consisting of certain number of consecutive-time data samples, and partial least squares is adopted to construct local models. Advantages of this adaptive strategy are that the number of local models does not need to be pre-defined and the local model set can be augmented online without retraining from scratch. Moreover, to improve the prediction accuracy, a novel online model adaptation criterion is proposed, which not only takes the current process dynamics into account, but also enables mining the information contained in the neighborhood of the query sample. The guidelines for tuning the model parameters are also presented. The LPLS-APSP scheme is applied to develop the dynamic soft sensors for a simulated continuous stirred tank reactor and a real industrial debutanizer column, and the results obtained demonstrate the effectiveness of this proposed approach, in comparison to several existing state-of-the-art methods, for online soft sensor design.

show abstract

“…The most commonly used solutions in the soft sensing field to this problem are the clustering based methods, such as fuzzy C-means (FCM) [Liu 2007;Fu et al, 2008;Liu et al, 2014] and expectation maximization (EM) based finite mixture models [Grbić et al, 2013;Khatibisepehr et al, 2012;Yu 2012b]. The main problems of these methods are that appropriate number of clusters is not easy to determine and it is difficult to add new clusters online when new process modes appear.…”

Section: Introductionmentioning

confidence: 99%

Adaptive soft sensor for quality prediction of chemical processes based on selective ensemble of local partial least squares models

Shao

Tian

2015

Chemical Engineering Research and Design

126

View full text Add to dashboard Cite

A Bayesian approach to design of adaptive multi-model inferential sensors with application in oil sand industry

Cited by 50 publications

References 19 publications

Modelling of a Naphtha Recovery Unit (NRU) with Implications for Process Optimization

Modelling of a Naphtha Recovery Unit (NRU) with Implications for Process Optimization

Online soft sensor design using local partial least squares models with adaptive process state partition

Adaptive soft sensor for quality prediction of chemical processes based on selective ensemble of local partial least squares models

Contact Info

Product

Resources

About