Methods for root cause diagnosis of plant‐wide oscillations

Duan, Ping; Chen, Tongwen; Shah, Sirish L.; Yang, Fan

doi:10.1002/aic.14391

Cited by 87 publications

(57 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The data driven methods are more popular owing to the fact that model based methods rely on process information/model, expert knowledge and analysis of P&IDs that may not be available or lack desired level of accuracy [8]. whereas data driven methods are based on the historical data alone and are quite flexible to use.…”

Section: Introductionmentioning

confidence: 99%

“…whereas data driven methods are based on the historical data alone and are quite flexible to use. The detailed description of both these approaches can be found in [8].…”

Section: Introductionmentioning

confidence: 99%

“…The causality is defined as: if prediction of time series y is improved by using the knowledge of other time series x, then x has a causal influence on y [8]. This definition of causality is taken up by Granger [10] who argued that x "Granger causes (G-Causes)" y if removing x from the universe of all possible causative variables U decreases the predictability of y i.e σ 2 (y|U ) < σ 2 (y|(U − x)).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Convergent cross mapping (CCM) based approach for isolating the source of plant-wide disturbances

Aftab

Hovd

Sivalingam

2017

2017 IEEE Conference on Control Technology and Applications (CCTA)

View full text Add to dashboard Cite

Abstract-Disturbances originating in one control loop of a large industrial plant can propagate far from the source, giving rise to plant-wide oscillations. The underlying interactions among the different control loops make it hard to identify the origin of such large scale disturbances. This paper studies the application of the convergent cross mapping (CCM) based technique to isolate the source of a plant-wide disturbance. The proposed scheme exploits the cause and effect relationships among the affected variables to find the source of disturbance. The states of the causative factors are estimated from the effect variable and the directionality of information flow is established using the correlation between the original and estimated signal. The method is applied to the industrial case study and is shown to be effective in isolating the disturbance origin.

show abstract

Section: Introductionmentioning

confidence: 99%

“…whereas data driven methods are based on the historical data alone and are quite flexible to use. The detailed description of both these approaches can be found in [8].…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Convergent cross mapping (CCM) based approach for isolating the source of plant-wide disturbances

Aftab

Hovd

Sivalingam

2017

2017 IEEE Conference on Control Technology and Applications (CCTA)

View full text Add to dashboard Cite

show abstract

“…For example, Duan et al extended the traditional concept of TE and made it more applicable, especially for multivariate cases [16,17]. Yang et al [18] and Duan et al [19] also summarized these methods for capturing causality in industrial processes.…”

Section: Introductionmentioning

confidence: 99%

Detection of Causality between Process Variables Based on Industrial Alarm Data Using Transfer Entropy

Yang

2015

Entropy

Self Cite

View full text Add to dashboard Cite

In modern industrial processes, it is easier and less expensive to configure alarms by software settings rather than by wiring, which causes the rapid growth of the number of alarms. Moreover, because there exist complex interactions, in particular the causal relationship among different parts in the process, a fault may propagate along propagation pathways once an abnormal situation occurs, which brings great difficulty to operators to identify its root cause immediately and to take proper actions correctly. Therefore, causality detection becomes a very important problem in the context of multivariate alarm analysis and design. Transfer entropy has become an effective and widely-used method to detect causality between different continuous process variables in both linear and nonlinear situations in recent years. However, such conventional methods to detect causality based on transfer entropy are computationally costly. Alternatively, using binary alarm series can be more computational-friendly and more direct because alarm data analysis is straightforward for alarm management in practice. The methodology and implementation issues are discussed in this paper. Illustrated by several case studies, including both numerical cases and simulated industrial cases, the proposed method is demonstrated to be suitable for industrial situations contaminated by noise.

show abstract

“…Several authors have been applying new methods that make it possible to extract process topology describing causality relationships from process knowledge and process data. The knowledge from process topology has many applications, for instance in the field of plantwide analysis, where the use of process connectivity can enhance the isolation and diagnosis of the root-cause of plantwide disturbances [2], [3]. There is also evidence of use of topology in other areas such as alarm rationalization [4], risk assessment [5] and control structure design [6].…”

Section: Introductionmentioning

confidence: 99%

Integration, navigation and exploration of plant topology networks using the property-graph model

Romero

Thornhill

2014

2014 IEEE/SICE International Symposium on System Integration

View full text Add to dashboard Cite

Abstract-Understanding the process topology is essential for many process systems engineering activities. Previous works in this area have explored the extraction of connectivity and causality from different sources by using structural and process data. A few works, however, have focused on how to integrate, store, and visualise connectivity models, which are becoming available in the literature. This paper proposes a method for integration, navigation and exploration of topology models in an efficient way, which can be applied to the study of connectivity and causality in chemical processes. I. INTRODUCTIONDisturbances in continuous processes commonly travel along the product stream through different propagation paths affecting the process performance. Abnormalities that propagate plantwide include oscillations caused by sticking valves, disturbances produced by interacting controllers, instabilities in the systems such as slugging flows, and vibrations caused by interacting units (for example in a turbine-generator train).Since large-scale chemical facilities are highly integrated systems containing recycle flows and utility systems, units are not independent anymore. Connectivity and causality have been concepts proposed for describing dependencies between units and variables [1]. Several authors have been applying new methods that make it possible to extract process topology describing causality relationships from process knowledge and process data. The knowledge from process topology has many applications, for instance in the field of plantwide analysis, where the use of process connectivity can enhance the isolation and diagnosis of the root-cause of plantwide disturbances [2], [3]. There is also evidence of use of topology in other areas such as alarm rationalization [4], risk assessment [5] and control structure design [6]. It seems now that mining the data and knowledge about topology (connectivity and causality) is a largely solved problem. What becomes challenging now is to integrate all these insights into a single topology model and make it possible for engineers to use it for exploration, navigation and more effective visualisation.This work is supported Marie Curie FP7 project ENERGY-SMARTOPS, Contract No: PITN-GA-2010-264940.David Dorantes Romero is with ABB Technology and Innovation, Norway, Ole Deviks vei 10, david.romero@no.abb.com. He is also part-time PhD student at the Department of Chemical Engineering at Imperial College.Nina F. Thornhill is with the Centre for Process Systems engineering, Department of Chemical Engineering, Imperial College London, London SW7 2AZ, U.K.The main contribution of this paper is the proposal of a method to integrate process connectivity and causality information into a plant topology network, making use of interaction and techniques to facilitate its navigation and exploration. Examples will also be given based on a novel prototype called Topoviz™, currently under development. II. STATE OF THE ARTInformation about the topology can be captured from modelbased m...

show abstract

Methods for root cause diagnosis of plant‐wide oscillations

Cited by 87 publications

References 59 publications

Convergent cross mapping (CCM) based approach for isolating the source of plant-wide disturbances

Convergent cross mapping (CCM) based approach for isolating the source of plant-wide disturbances

Detection of Causality between Process Variables Based on Industrial Alarm Data Using Transfer Entropy

Integration, navigation and exploration of plant topology networks using the property-graph model

Contact Info

Product

Resources

About