Batch Process Monitoring Based on Fuzzy Segmentation of Multivariate Time-Series

Tanatavikorn, Harakhun; Yamashita, Yoshiyuki

doi:10.1252/jcej.16we193

Cited by 5 publications

(8 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Interestingly, in the chemical context, clustering helped several applications, such as the identification of fault subclasses in contexts where faults exhibit time-varying characteristics [29]. Similarly, in penicillin production processes, clustering has been used to identify the specific production stage in the batch process [30]. This contribution also ties in with the keywords in topic 2, in which phase partitioning is applied with clustering.…”

Section: Topic 1-chemical Contextmentioning

confidence: 99%

Clustering Application for Condition-Based Maintenance in Time-Varying Processes: A Review Using Latent Dirichlet Allocation

et al. 2022

View full text Add to dashboard Cite

In the field of industrial process monitoring, scholars and practitioners are increasing interest in time-varying processes , where different phases are implemented within an unknown time frame. The measurement of process parameters could inform about the health state of the production assets, or products, but only if the measured parameters are coupled with the specific phase identification. A combination of values could be common for one phase and uncommon for another phase; thus, the same combination of values shows a high or low probability depending on the specific phase. The automatic identification of the production phase usually relies on clustering techniques. This is largely due to the difficulty of finding training fault data for supervised models. With these two considerations in mind, this contribution proposes the Latent Dirichlet Allocation as a natural language-processing technique for reviewing the topic of clustering applied in time-varying contexts, in the maintenance field. Thus, the paper presents this innovative methodology to analyze this specific research fields, presenting the step-by-step application and its results, with an overview of the theme.

show abstract

Section: Topic 1-chemical Contextmentioning

confidence: 99%

Clustering Application for Condition-Based Maintenance in Time-Varying Processes: A Review Using Latent Dirichlet Allocation

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Interestingly, in the chemical context, clustering helped several applications, such as the identification of fault subclasses in contexts where faults exhibit time-varying characteristics [29]. Similarly, in penicillin production processes, clustering has been used to identify in the batch process the specific production stage [30]. This contribution also ties in with the keywords in topic 2.…”

Section: Topicmentioning

confidence: 99%

“…In that topic phase partitioning is applied with clustering. In this contribution [30], it is highlighted that the analyzed process belongs to the world of chemical processes, while in topic 2 the focus is on phase recognition, and the production process assumes a pure case study. Moreover, this contribution [30] is the only one applying the AdTsPCA.…”

Section: Topicmentioning

confidence: 99%

“…Finally, to conclude the application areas presented in this topic of contributions, the area of gearboxes is treated by considering vibrations as a tool for fault diagnosis [35]. Focusing on the techniques in this topic, it is to mention the proposal of the AdTsPCA technique [30]. This technique is highlighted as innovative compared to other techniques, even when considering the other identified topics, while other techniques are not highlighted as they do not predominate in the sub-thematic area.…”

Section: Topicmentioning

confidence: 99%

See 1 more Smart Citation

Clustering Application for Condition-based Maintenance in Time-varying Processes: A Review

Quatrini¹,

Colabianchi²,

Costantino³

et al. 2021

Preprint

View full text Add to dashboard Cite

In the field of industrial process monitoring, more and more interest is being shown in specific process categories. These include time-varying processes, that is, those processes whereby the response one receives as output from the system depends on when the input signal is sent into it. There are many reasons for this process variability and such contexts are not always analyzed with this operational characteristic at their core. At the same time, interest in certain categories of techniques is also becoming more prominent, to meet certain application needs. Among these, clustering and unsupervised techniques in general are gaining ground. This is largely due to the difficulty of finding fault data with which to train, for example, supervised models. On the other hand, the clustering technique, on which this contribution focuses, also makes it possible to compensate for the lack of complete knowledge of the structure of the process itself. With these two considerations in mind, this contribution proposes a literature review on the topic of clustering applied in time-varying contexts, in the maintenance field. The aim is to present an overview of the main fields of study, the role of clustering in this context and the main clustering techniques used.

show abstract

“…Then, the Hotelling T 2 statistic and reconstruction error are employed as measures of the approximation errors in finding the optimal segmentation [23], [24]. A further approach extends the PCA based approach to a probabilistic framework [9], [25].…”

Section: Introductionmentioning

confidence: 99%

Segmentation of Multivariate Industrial Time Series Data Based on Dynamic Latent Variable Predictability

Huang

2020

IEEE Access

View full text Add to dashboard Cite

Time series segmentation is an important vehicle of data mining and extensively applied in the areas of machine learning and anomaly detection. In real world tasks, dynamics widely exist in time series but have been little concerned. This paper proposes an algorithm which can partition a multivariate time series into subsequences at points where the dynamics change. In process industries, the change of dynamics often relates to operation regime change, working condition shifting or faults. Therefore, to segment time series according to change of dynamics can be useful in data preprocessing and getting deep insight of the process in various industrial process monitoring tasks. The proposed algorithm recursively merges neighborhood subsequences through a heuristic bottom-up procedure. The cost of merging is defined as the mutual predictability of the subsequence models which are constructed using the dynamic-inner principal component analysis algorithm. Then, the goal becomes finding the segmentation scheme which minimizes the averaged dynamic prediction errors of each subsequence model. The proposed algorithm is evaluated on both simulated data and the time series data collected from an industrial processing plant. The results show that it outperforms the static principal component analysis based methods. INDEX TERMS time series analysis, time series prediction, time series segmentation, change point detection, dynamic inner model, dynamic principal component analysis

show abstract

Batch Process Monitoring Based on Fuzzy Segmentation of Multivariate Time-Series

Cited by 5 publications

References 28 publications

Clustering Application for Condition-Based Maintenance in Time-Varying Processes: A Review Using Latent Dirichlet Allocation

Clustering Application for Condition-Based Maintenance in Time-Varying Processes: A Review Using Latent Dirichlet Allocation

Clustering Application for Condition-based Maintenance in Time-varying Processes: A Review

Segmentation of Multivariate Industrial Time Series Data Based on Dynamic Latent Variable Predictability

Contact Info

Product

Resources

About