Controller Performance Monitoring: A Survey of Problems and a Review of Approaches from a Data-Driven Perspective with a Focus on Oscillations Detection and Diagnosis

Abstract

Optimal operations of industrial control systems require rigorous monitoring to ensure safety, increase profitability, and minimize plant maintenance downtime. Thus, controller performance monitoring has been actively pursued by the research community, resulting in increased research publications over the past two decades. The availability of large data sets, the so-called big data era, has led impetus to the data-driven domain of controller performance monitoring. In this paper, a comprehensive review of the data-driven research in the CPM domain is presented. To illustrate the rationale behind the research efforts, a succinct explanation of the faults observed in control loops and their influence on overall controller performance is also presented. The paper then provides the publicly available data repositories adopted by most researchers to assess and compare their performance monitoring techniques realistically. Moreover, a review of the most eminent techniques proposed so far concerning both detection and diagnosis phases is presented.

Detecting Plant-Wide Oscillation Propagation Effects of Disturbances and Faults in a Chemical Process Plant Using Network Topology of Variance Decompositions

“…The interconnection of process units makes this analysis task even more challenging [1]. Previous works have demonstrated the adaptability of algorithms and computations in econometrics for use in chemical process oscillation analysis [1,6]. Such adaptations of econometrics techniques for chemical process dynamics analysis naturally occur due to the inherent similarities in the feedback loop mechanisms designed to achieve stability in both fields, as depicted in Figure 2.…”

Control valve stiction detection using Markov transition field and deep convolutional neural network

“…In this hybrid methodology, CNN was first used to extract features from time series data belonging to stiction and non-stiction categories, and the extracted features were utilized to build the PCA model and determine stiction detection metrics: T 2 -statistic and Q-statistic. The details of other existing methods can be found in the works of Akavalappil and Radhakrishnan [20] and Bounoua et al [21] 1.1 | Motivation and contributions More researchers are utilizing ML in their work today thanks to advancements in the field since it is more efficient, more precise, and applicable to a wider range of situations. Several publications that merged ML with conventional stiction detection techniques have been published, as was indicated in the preceding section.…”

“…In this hybrid methodology, CNN was first used to extract features from time series data belonging to stiction and non‐stiction categories, and the extracted features were utilized to build the PCA model and determine stiction detection metrics: ‐statistic and Q ‐statistic. The details of other existing methods can be found in the works of Akavalappil and Radhakrishnan [ 20 ] and Bounoua et al [ 21 ]…”

Distributed Control of an Ill-Conditioned Non-Linear Process Using Control Relevant Excitation Signals

“…The recent drive towards the digitization and automation of industries towards the industry 4.0 paradigm shift means that there is the need to improve the performance of existing control loops. Moreover, about 30% of exisiting controllers are in manual mode while 41% of control loops are considered to have fair or poor performance [17].…”