Sparse PCA Support Exploration of Process Structures for Decentralized Fault Detection

Abstract

With the ever-increasing use of sensor technologies in industrial processes and more data becoming available to engineers, the fault detection and isolation activities in the context of process monitoring have gained significant momentum in recent years. A statistical procedure frequently used in this domain is principal component analysis (PCA), which can reduce the dimensionality of large data sets without compromising the information content. While most process monitoring methods offer satisfactory detection capabilities, understanding the root cause of malfunctions and providing the physical basis for their occurrence have been challenging. The relatively new sparse PCA techniques represent a further development of the PCA in which not only the data dimension is reduced but also the data are made more interpretable, revealing clearer correlation structures among variables. Hence, taking a step forward from classical fault detection methods, in this work, a decentralized monitoring approach is proposed based on a sparse algorithm. The resulting control charts reveal the correlation structures associated with the monitored process and facilitate a structural analysis of the occurred faults. The applicability of the proposed method is demonstrated using data generated from the simulation of the benchmark vinyl acetate process. It is shown that the sparse principal components, as a foundation to a decentralized multivariate monitoring framework, can provide physical insight toward the origins of process faults.

A novel fault detection scheme for a nonlinear dynamic process based on generalized non‐negative matrix projection‐maximum mean discrepancy: Application on the DAMADICS benchmark process

“…It should be noted that GNMP is an improved version of PNMF, [27] which relaxes the non-negative constraints on the original data X to a certain degree. Therefore, the matrix decomposition form in GNMP can be represented as in Equation (11). To remove the non-negative constraint on the original data X, two matrices are introduced in Equation ( 12) that transform positive and negative elements of X differently.…”

“…With the availability of diverse sensors for collecting and recording production process status data, the field of data-driven process fault detection has witnessed rapid advancements to ensure optimal safety in manufacturing processes. [7,8] The classic methods for process monitoring include principal component analysis (PCA), [9][10][11][12][13] partial least squares (PLS), [14][15][16][17][18] and non-negative matrix factorization (NMF). [19][20][21][22] A significant amount of research shows that PCA demonstrates excellent process monitoring performance when the collected process data only contain linear characteristics and follow a multivariate Gaussian distribution.…”

Fluid catalytic cracking process quality-driven fault detection based on partial least squares and deep feedforward neural network

“…Process monitoring has been highly valued for ensuring the long-term reliable operation of such complicated processes. Classical monitoring methods based on models or prior knowledge have failed to work well; thus, data-driven methods are bound to develop rapidly because of the wide use of new measurement technologies and the considerable progress in data mining (Bounoua et al, 2019; Ge, 2017; Jiang et al, 2019; Li and Feng, 2020; Theisen et al, 2021; Yao et al, 2022). Multivariate statistical process monitoring (MSPM) methods are fairly representative of data-driven methods.…”

Actuator and Sensor Fault Detection and Isolation System Applied to a Distillation Column

“…by failures in some physical components such as sensors and actuators [3], [5], [7] and failures in the data acquisition and monitoring systems [5], [8], [9], [10], [11], [12]. These drawbacks can result in an unstable, dangerous, and disastrous operation state as in the cases of the refinery accidents in BP in Texas, USA; Kuwait Petrochemical's Mina Al-Ahmedi refinery; Union Carbide in Bhopal, India; and Piper Alpha of Occidental Petroleum [3], [5], [12].…”

“…The third approach is to analyze previous knowledge of the process and the relationship between the faults and the parameters or states of the system [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. Based on these three approaches, different methods of FD, diagnosis and isolation have been proposed: principal component analysis (PCA) [8], [11], independent component analysis (ICA) [10], [11], statistical process monitoring (SPM) [26], [27], partial least squares (PLS) [11], multivariate statistical process monitoring (MSPM) [11], [26], [28], [29], nonlinear PCA (NPCA) and kernel PCA (KPCA) to handle non-linearity [11], [30], [31], support vector machine (SVM), artificial neural network (ANN), [11], [32], [33], [34], gaussian mixture model (GMM) [5], [10], [11], [26], the Bayesian network (BN) [5], [10], [11], [26], [35], [36], dynamic BN (DBN), the Kalman filter (KF) [11], [34]; naive Bayesian classifier (NBC), [11], and model-based FD and failure prediction framework for a class of multi-input and multi-output no...…”