Tennessee Eastman Process Research Articles

A novel dynamic machine learning-based explainable fusion monitoring: application to industrial and chemical processes

Abstract The complexity and fusion dynamism of the modern industrial and chemical sectors have been increasing with the rapid progress of IR 4.0–5.0. The transformative characteristics of Industry 4.0–5.0 have not been fully explored in terms of the fundamental importance of explainability. Traditional monitoring techniques for automatic anomaly detection, identifying the potential variables, and root cause analysis for fault information are not intelligent enough to tackle the intricate problems of real-time practices in the industrial and chemical sectors. This study presents a novel dynamic machine learning based explainable fusion approach to address the issues of process monitoring in industrial and chemical process systems. The methodology aims to detect faults, identify their key causes and feature variables, and analyze the root path of fault propagation with the time and magnitude of one cause variable to another impact. This study proposed using a time domain multivariate granger-entropy-aided dynamic independent component analysis (DICA)—distributed canonical correlation analysis approach, incorporating the dynamics time wrapping supported time delay-signed directed graph. The proposed methodology utilized the application to industrial and chemical processes and verified using the continuous stirred tank reactor and Tennessee Eastman process as practical application benchmarks. The framework’s validations and efficiency are evaluated using established techniques such as classic computed ICA and DICA as standard model scenarios. The outcomes and results showed that the newly developed strategy is preferable to previous approaches regarding explainability and robust detection and identification of the actual root causes with high FDRs and low FARs.

MOLA: Enhancing Industrial Process Monitoring Using a Multi-Block Orthogonal Long Short-Term Memory Autoencoder

In this work, we introduce MOLA, a multi-block orthogonal long short-term memory autoencoder paradigm, to conduct accurate, reliable fault detection of industrial processes. To achieve this, MOLA effectively extracts dynamic orthogonal features by introducing an orthogonality-based loss function to constrain the latent space output. This helps eliminate the redundancy in the features identified, thereby improving the overall monitoring performance. On top of this, a multi-block monitoring structure is proposed, which categorizes the process variables into multiple blocks by leveraging expert process knowledge about their associations with the overall process. Each block is associated with its specific orthogonal long short-term memory autoencoder model, whose extracted dynamic orthogonal features are monitored by distance-based Hotelling’s T2 statistics and quantile-based cumulative sum (CUSUM) designed for multivariate data streams that are nonparametric and heterogeneous. Compared to having a single model accounting for all process variables, such a multi-block structure significantly improves overall process monitoring performance, especially for large-scale industrial processes. Finally, we propose an adaptive weight-based Bayesian fusion (W-BF) framework to aggregate all block-wise monitoring statistics into a global statistic that we monitor for faults. Fault detection speed and accuracy are improved by assigning and adjusting weights to blocks based on the sequential order in which alarms are raised. We demonstrate the efficiency and effectiveness of our MOLA framework by applying it to the Tennessee Eastman process and comparing the performance with various benchmark methods.

Predicting and monitoring faults in intricate processes through the utilization of an ensemble of machine learning regression models: a case study on the Tennessee Eastman Process

Predicting and monitoring faults in intricate processes through the utilization of an ensemble of machine learning regression models: a case study on the Tennessee Eastman Process

Design of model fusion learning method based on deep bidirectional GRU neural network in fault diagnosis of industrial processes

This paper proposes an end-to-end model fusion feature learning method based on deep bidirectional gated recurrent unit (MCNN-DBiGRU) for fault diagnosis in industrial processes. First, a feature-aligned multi-scale feature extraction model (MCNN) is designed by analyzing the working principles of convolutional and pooling layers of convolutional neural networks. Secondly, a deep bidirectional mechanism is proposed to better extract the time series features in the process data. This mechanism makes the recurrent neural network not only present the forward processing input features from the past to the future, but also the reverse processing from the future to the past. By integrating these features, the diagnostic performance of the network model is improved. To verify that the proposed model has effective diagnostic accuracy for fault diagnosis, we conduct simulation experiments on the Tennessee-Eastman (TE) process and a chemical coking furnace, and compare with several conventional network models. In the end, not only the effectiveness of the model is proved, but it is also confirmed that the model is superior to other conventional neural networks in both diagnostic accuracy and feature robustness.

Just-in-time framework for robust soft sensing based on robust variational autoencoder

Modeling with high-dimensional data subject to abnormal observations have always been a practical interest. In this paper, under the just-in-time learning (JITL) framework, a robust soft sensor modeling approach is developed based on robust Variational Autoencoder (VAE). Unlike the vanilla VAE that extracts features from the given dataset under the Gaussian prior assumption, robust VAE employs Student’s t-distribution as prior distribution to handle abnormal data. Under assumption of the Student’s t-prior, the proposed robust VAE model is capable of describing collected data contaminated with outliers. Once the robust VAE model is trained, each robust feature variable in the latent space can be determined. Subsequently, similarity measure is calculated using robust Kullback-Leibler divergence between two Student’s t-distributions, that is, the distribution of a new data sample and that of each historical data sample. After completing similarity measurement for a query sample, the weights for input-output historical data can be determined. Based on these weighted historical data samples, a robust probabilistic principal component regression (PPCR) is utilized to perform local modeling for prediction. Numerical simulations, including the Tennessee Eastman and Penicillin fermentation benchmark processes, are utilized to validate the proposed JITL-based robust soft sensor modeling method.

Advance industrial monitoring of physio-chemical processes using novel integrated machine learning approach

With the rapid transition of Industry 4.0 to 5.0, modern industrial physio-chemical processes are characterized by two critical challenges: process safety and the quality of the final product. Traditional industrial monitoring methods have low reliability in accuracy and robustness, and they are inefficiently providing satisfactory results. This paper introduces a novel integration technique that employs machine learning (ML) to tackle the challenges associated with real industrial monitoring in physical and industrial processes. The proposed framework integrates distributed canonical correlation analysis - R-vine copula (DCCA-RVC), global local preserving projection (GLPP), and 2-Dimensional Deng information entropy (2-DDE). The framework's ability and productivity are assessed utilizing existing approaches such as wavelet-PCA, MRSAE, and DALSTM-AE and the new proposed novel integrated machine learning-based (DCCA-RVC) approach as benchmarks for model performance. The proposed novel approach has been validated by testing it on the ethanol-water system distillation column (DC) and Tennessee Eastman Process (TEP), utilizing it as actual industrial benchmarks. The results demonstrate that the novel integration ML-technique (DCCA-RVC) T22 – GLP monitoring graphs for the fault class type 1 in the distillation column showed a (FAR) of 0 %, a (FDR) of 100 %, a precision of 100 %, F1-score of 100 % and an accuracy of 100 %. However, for the TEP process failure event 13, the (FAR) was 0 %, the (FDR) was 99 %, the accuracy was 100 %, the F1-score was 99.5 %, and the accuracy was 99.5 %.

Novel modified weights and cosine similarity based maximum marginal projection and its application in fault diagnosis

Abstract Facing the problem that the data generated in industrial processes have few labeled samples and the local manifold learning dimensionality reduction method ignores the local spatial structure of sample points and the distance relationship in constructing different weights. To solve the above problems, this paper presents a novel modified weights and cosine similarity based maximum marginal projection named MCMMP. In MCMMP, cosine similarity is used to consider the space feature of sample points, which enhances the performance of dimensionality reduction. The new modified weights are applied to measure the between-class and the within-class sample points, which enhance the divisibility of sample points. After MCMMP dimensionality reduction, the classifier is used to classify the dimensionality reduction sample points. Finally, the proposed new method is used in two cases Tennessee Eastman Process (TEP) and Three-phase Flow Facility (TFF) to test the fault diagnosis performance. The results of the simulation process indicated that the new fault diagnosis method based on MCMMP, compared with other related diagnosis methods, has good performance.

A novel label-aware global graph construction method and spiking-coded graph neural network for intelligent process fault diagnosis

Fault diagnosis plays a crucial role in ensuring the safety and efficiency of industrial processes. However, traditional techniques often face difficulties in handling large-scale data characterized by complex structures and relationships. To efficiently represent industrial data as graphs and develop a low-energy-cost feature extraction model, a novel label-aware global graph construction method and a spiking graph convolutional network (SGCN) are proposed in this study to achieve intelligent process fault diagnosis. The label-aware method enhances graph data representation by capturing intrinsic correlations and global features. The SGCN integrates graph convolutional layers with spiking encoding, enabling effective feature extraction while offering computational efficiency advantages. A weighted loss function is introduced to mitigate data imbalance issues. Experiments on the Tennessee Eastman process, the Three-phase Flow Facility, and the de-propanizer distillation process demonstrate SGCN’s superior performance over baseline models in various fault scenarios, while significantly reducing computational costs. The proposed method offers promising potential for reliable and efficient fault diagnosis in complex real-world industrial environments.

Industrial Process Fault Detection Based on IGA‐Combinatorial Model Decision Mechanism

ABSTRACTTo address the challenges of extracting features from complex industrial process data, the reliance of numerous fault detection methodologies on presupposed data distribution types, and the limited generalization capacity of fault detection, this manuscript introduces a sophisticated algorithm for industrial process fault detection. This algorithm harnesses the information gain adaptive (IGA) technique for feature selection and a synergistic model decision mechanism. Initially, the process involves the computation of information gain via decision trees, coupled with the determination of the value through cross‐validation. This strategy enables the adaptive selection of features, thereby facilitating data dimensionality reduction and effective feature extraction. The subsequent phase introduces a ternary statistical measure monitoring group for the detection of linear faults, while autoencoders and one‐class SVM methodologies are applied for the monitoring of nonlinear faults. The culmination of this approach is the development of an innovative weighted decision mechanism, designed to amalgamate the findings from both linear and nonlinear detection avenues, yielding more dependable detection results. The validation of this algorithm employs datasets from the water chillers process and Tennessee Eastman (TE) process, demonstrating the IGA‐combined model's superior performance over isolated linear or nonlinear detection algorithms in terms of detection accuracy and robustness. Notably, the efficacy of this method is not contingent upon specific assumptions regarding data distribution, rendering it a versatile and efficacious tool for the fault detection in industrial processes.

SIMTSeg: A self-supervised multivariate time series segmentation method with periodic subspace projection and reverse diffusion for industrial process

Subsequences with varied regimes in the industrial multivariate time series (MTS) are closely associated with the dynamic status of the multi-phased industrial process. Time series segmentation (TSS) provides insights into the underlying behavior of industrial systems. However, the complexity of industrial data poses significant challenges to the conventional TSS methods. Motivated by this, a Self-supervised Industrial Multivariate Time-series Segmentation method (SIMTSeg) is presented in this work. An MTS folding module based on Ramanujan periodic subspace projection is first proposed, where the MTS is reshaped into the 3D feature map to realize the compact representation of the intricate data dependencies. Subsequently, a self-supervised module based on the encoder-decoder architecture is adopted to address the problem of deficient and task-specific annotations in industrial data. The folded feature map is denoised step by step following the reverse diffusion process, and finally turns into the segmentation mask without redundant details. The proposed SIMTSeg has been validated by a popular industrial benchmark, the Tennessee Eastman Process, and outperforms the unsupervised data-driven baselines in terms of various performance metrics. SIMTSeg has no prerequisite on the number of segmentation points or regime types, and is capable of giving more meaningful segmentation results that are in line with the high-level semantics.

A novel multi-label classification deep learning method for hybrid fault diagnosis in complex industrial processes

Hybrid Fault Detection and Diagnosis, encompassing both individual and simultaneous faults, is an important solution needed in chemical process safety practice. We systematically examine the two perspectives of simultaneous faults in previous studies: multi-class and multi-label classification, and highlight the limitations of the former while demonstrating the efficacy of the latter. Then, a novel multi-label classification Hybrid Fault Transformer (mcHFT) model was put forward to address hybrid faults. Our model is capable of learning not only intrinsic features of individual faults but also their coupled relationships. Importantly, this work constitutes the first comprehensive evaluation of Hybrid FDD on the Tennessee Eastman (TE) process to our knowledge. The mcHFT model significantly enhances key performance indicators over existing models and introduces an adaptive strategy to reduce false positives. The dataset developed for this research is made available under an MIT license, contributing a valuable resource for future exploration in this field.

An efficient multimodal attentional principal component analysis for continual learning-based dynamic process monitoring

Traditional multimode process monitoring methods extract features from time series data. Due to the catastrophic forgetting effect, data-driven multimode dynamic process monitoring is challenging based on a single monitoring model paradigm, i.e. the learned knowledge from previous modes may diminish as operating conditions undergo changes between modes, yet it is impractical to access all past data to retrain the model. In this work, a novel efficient method of multimodal attentional principal component analysis (M-APCA) with continual learning ability is introduced. Under the assumption that data from successive modes are received sequentially, dynamic process data are modeled using an attention mechanism to capture the relationship between data and the latent space, whereby meaningful information is concentrated as dynamic features which are extracted via a vector autoregressive model. In order to overcome the catastrophic forgetting problem, the idea of replay continual learning is employed. Specifically, past modes’ data which are significant to reflect the operating conditions, are selected and stored. These are repeatedly used in tandem with sequential data as replay data. Two types of attention mechanisms are considered and analyzed, each of which is specifically designed to learn from data in an unsupervised manner, so the overall algorithm is efficient both in time and storage costs. The proposed attentional principal component analysis and M-APCA are analyzed against several state-of-the-art methods to highlight the virtues of the proposed method. Compared with multimode monitoring methods, the effectiveness is demonstrated through case studies of: a continuous stirred tank heater, the Tennessee Eastman process and a practical coal pulverizing system.

Enhancing fault detection in multivariate industrial processes: Kolmogorov–Smirnov non-parametric statistical approach

The accurate detection of abnormal events in modern process plants is extremely important. The detection of faults of small magnitude and in a noisy process environment is still a challenge that many industries face. In this paper, a Kolmogorov–Smirnov (KS) based non-parametric statistical fault indicator is proposed to identify a variety of sensor faults in process plants. The data collected from most modern process plants are randomly varying and non-Gaussian, due to which the multivariate scheme based on Independent Component Analysis (ICA) is considered. The KS-based indicator is amalgamated with the ICA multivariate model, which yields a novel ICA-KS-based fault detection (FD) scheme. The KS statistical indicator compares any two distributions and checks if they are similar or dissimilar. The potential of the KS indicator is extended in the FD domain, where the residuals of training and testing data are compared in a sliding window. The ability of the proposed ICA-KS-based FD strategy is validated on a simulated Distillation column (DC) process and the benchmark Tennessee Eastman (TE) process to identify a variety of sensor faults. The simulation results demonstrate the effectiveness of the detection performance of the ICA-KS scheme, which outperforms conventional FD-based methods with a high detection rate.

Process monitoring method based on vine copula and transfer learning strategy

In practical industrial processes, the limited sampling time and other factors result in a scarcity of historical data for certain modes, leading to diminished generalization and accuracy of process monitoring models. To solve this problem, a process monitoring method based on vine copula-based dependence description (VCDD) and transfer learning strategy (TLVCDD) is proposed in this study. The proposed method constructs a VCDD model by utilizing target domain data and subsequently selects suitable candidate sample groups from the abundant source domain data based on this model. Candidate sample groups are transferred sequentially according to their priorities, which are quantified based on the maximum mean discrepancies between candidate sample groups and the target domain data. The VCDD model exhibiting the most superior overall performance during the transfer process is chosen and employed for online process monitoring. The effectiveness of the proposed method is demonstrated through a numerical example and Tennessee Eastman (TE) process.

An Adaptive Spatiotemporal Decouple Graph Convolutional Network Based Quality‐Related Fault Detection Method for Complex Industrial Processes

ABSTRACTWith the rapid development of industrial technology, industrial processes become increasingly complex, presenting characteristics of large‐scale and multi‐unit collaboration. However, most current fault detection methods focus on nonlinearity, dynamics, and other characteristics, while neglecting spatiotemporal information. To address this issue, an adaptive spatiotemporal decouple graph convolutional network based quality‐related fault detection method is proposed in this article. First, the temporal graph convolutional network and spatial graph convolutional network are combined organically in the form of joint training. Second, considering that fixed graph structures cannot reflect the dynamic relationships among nodes, we proposed an adaptive weighted mask mechanism to construct dynamic correlation graph embedded with priori knowledge. Multi‐attention mechanism is used to integrate spatiotemporal information, besides, we designed a decoupling layer to avoid information redundancy. Finally, the proposed spatiotemporal graph convolutional network is used to establish a regression model, the quality‐related latent variables are extracted by the decoupling layer, and the statistic is constructed based on Kullback–Leibler divergence. The effectiveness and feasibility of the proposed method are illustrated with the hot strip mill process and the Tennessee Eastman process.

Fault Monitoring Method for the Process Industry System Based on the Improved Dense Connection Network

The safety of chemical processes is of critical importance. However, traditional fault monitoring methods have insufficiently studied the monitoring accuracy of multi-channel data and have not adequately considered the impact of noise on industrial processes. To address this issue, this paper proposes a neural network-based model, DSCBAM-DenseNet, which integrates depthwise separable convolution and attention modules to fuse multi-channel data features and enhance the model’s noise resistance. We simulated a real environment by adding Gaussian noise with different signal-to-noise ratios to the Tennessee Eastman process dataset and trained the model using multi-channel data. The experimental results show that this model outperforms traditional models in both fault diagnosis accuracy and noise resistance. Further research on a compressor unit engineering instance validated the superiority of the model.

KPI-oriented process monitoring based on causal-weighted partial least squares

With the increasing demand for product quality, Key performance indicator (KPI)-oriented process monitoring plays an important role in modern industrial. Partial least squares (PLS)-based methods are widely adopted for KPI-oriented process monitoring, in which the process variable space is decomposed into a principal component subspace that has a strong correlation with KPIs and a residual subspace unrelated to KPIs, and the two spaces are monitored separately. However, if a KPI-unrelated fault occurs in a variable that has no causal relation with any KPI, but has a spurious correlation with some KPIs because of the existence of confounders, PLS based methods may mis-regard the KPI-unrelated fault as a KPI-related fault. This occurs because as a correlation analysis-based method, PLS cannot discriminate whether a variable is a real cause of KPIs or has a spurious correlation with KPIs. Motivated by this, this paper proposes two causal-weighted PLS methods for KPI-oriented process monitoring by combining the LiNGAM-based causal discovery with PLS, which first calculate causal weights based on the modified LiNGAM algorithm and bootstrap strategy, and then employ the causal weights to reweight the weight vectors of PLS to enhance the influence of causal variables in the KPI-related principal subspace and reduce the influence of spurious correlations. Case studies based on a simulated dataset, Tennessee Eastman Process dataset and field data from finishing rolling mill process show that the proposed methods can significantly reduce the false alarm rate for KPI-unrelated faults (i.e., the probability that a KPI-unrelated fault sample is mis-regarded as a KPI-related fault sample) caused by spurious correlation without significantly compromising the fault detection rate of KPI-related faults.

Cross-domain fault diagnosis for multimode green ammonia synthesis process based on DA-CycleGAN

Green ammonia is a crucial strategy for reducing carbon emissions and promoting sustainable development. However, in industrial applications, the production load of the ammonia synthesis section must be adjusted to accommodate fluctuations in renewable energy generation and hydrogen production. Consequently, the green ammonia synthesis process operates under multiple conditions with varying production loads. The operations of this multimode process introduce new challenges for process safety. Traditional fault diagnosis methods experience significant performance degradation when production conditions change. In new conditions, only a small number of normal samples can be obtained, and no fault samples are available. To address this issue, a novel transfer learning method named DA-CycleGAN is proposed for the multimode green ammonia synthesis process. This method combines a two-dimensional generation model based on CycleGAN (Cycle-Consistent Generative Adversarial Networks) with domain adaptation to enhance model performance in cross-domain tasks. The feasibility of the proposed method was initially validated using the benchmark Tennessee-Eastman process for fault diagnosis. Subsequently, a case study of the green ammonia synthesis process demonstrated that it significantly enhances performance in multimode processes, ensuring process safety and reducing losses for industrial applications.

EPMITS: An efficient prediction method incorporating trends and shapes features for chemical process variables

With the transformation of industrial production digitization and automation, process monitoring has been an indispensable technical method to realize the safe and efficient production of chemical process. Accurate prediction of process variables in chemical process can indicate the possible system change to reduce the probability of abnormal conditions. Current popular deep learning prediction methods trained with MSE or its variants may exhibit limitations in extracting shape features of chemical process data. In this paper, we proposed an efficient prediction method incorporating trends and shapes features (EPMITS) for chemical process variables. Specifically, we introduced a novel differentiable loss function Efficient Shape Error (ESE) to quantify shape differences between two time series of equal length in chemical process data. Then we trained deep learning models with MSE and ESE as loss function by two steps in training stage, to effectively acquire both trend and shape features of chemical process data. The proposed method was evaluated by the Tennessee Eastman process datasets and a real fluid catalytic cracking dataset from a petrochemical company. The results indicate that EPMITS models exhibit high prediction accuracy and short model training time across various time scales. These findings demonstrate the considerable feasibility and significant potential of EPMITS for future fault prognosis applications.

Real-time risk prediction of chemical processes based on attention-based Bi-LSTM

Refined risk prediction must be achieved to guarantee the safe and steady operation of chemical production processes. However, there is high nonlinearity and association coupling among massive, complicated multisource process data, resulting in a low accuracy of existing prediction technology. For that reason, a real-time risk prediction method for chemical processes based on the attention-based bidirectional long short-term memory (Attention-based Bi-LSTM) is proposed in this study. First, multisource process data, such as temperature, pressure, flow rate, and liquid level, are preprocessed for denoising. Data correlation is analyzed in time windows by setting time windows and moving step lengths to explore correlations, thus establishing a complex network model oriented to the chemical production process. Second, network structure entropy is introduced to reduce the dimensions of the multisource process data. Moreover, a 1D relative risk sequence is acquired by max–min deviation standardization to judge whether the chemical process is in a steady state. Finally, an Attention-based Bi-LSTM algorithm is established by integrating the attention mechanism and the Bi-LSTM network to fit and train 1D relative risk sequences. In that way, the proposed algorithm achieves real-time prediction and intelligent perception of risk states during chemical production. A case study based on the Tennessee Eastman process (TEP) is conducted. The validity and reasonability of the proposed method are verified by analyzing distribution laws of relative risks under normal and fault conditions. Also, the proposed algorithm importantly improves the prediction accuracy of chemical process risks relative to that of existing prediction technologies.