Nonlinear Batch Process Monitoring Using Phase-Based Kernel-Independent Component Analysis−Principal Component Analysis (KICA−PCA)

Fault detection for NOx emission process in thermal power plants using SIP-PCA

With the era of big data, data-driven models are increasingly vital to just-in-time decision support in pollution emission management and planning. This article aims to evaluate the usability of the proposed data-driven model to monitor NOx emission from a coal-fired boiler process using easily measured process variables. As the emission process is highly complex, process variables interact with each other, and they cannot guarantee that all the variables in the actual operation obey the Gaussian distributions. As conventional principal component analysis (PCA) can only extract variance information, a novel data-driven model is proposed, called survival information potential-based PCA (SIP-PCA) model, is proposed in this work. First, an improved PCA model is established based on the SIP performance index. SIP-PCA can extract more information in the latent space from the process variables following the non-Gaussian distributions. Then, the control limits for fault detection are determined based on the kernel density estimation method. Finally, the proposed algorithm is successfully applied to a real NOx emission process. By monitoring the operation of process variables, possible failures can be detected as soon as possible. Fault isolation and system reconstruction can be implemented in time, preventing NOx emissions from exceeding its standard.

LSTMED: An uneven dynamic process monitoring method based on LSTM and Autoencoder neural network

Due to the complicated production mechanism in multivariate industrial processes, different dynamic features of variables raise challenges to traditional data-driven process monitoring methods which assume the process data is static or dynamically consistent. To tackle this issue, this paper proposes a novel process monitoring method based on the long short-term memory (LSTM) and Autoencoder neural network (called LSTMED) for multivariate process monitoring with uneven dynamic features. First, the LSTM units are arranged in the encoder-decoder form to construct an end-to-end model. Then, the constructed model is trained in an unsupervised manner to capture long-term time dependency within variables and dominant representation of high dimensional process data. Afterward, the kernel density estimation (KDE) method is performed to determine the control limit only based on the reconstruction error from historical normal data. Finally, effective online monitoring for uneven dynamic process can be achieved. The performance and advantage of the process monitoring method proposed are explained through typical cases, including the numerical simulation and Tennessee Eastman (TE) benchmark process, and comparative experimental analysis with state-of-the-art methods.

Quality Prediction Model of KICA-JITL-LWPLS Based on Wavelet Kernel Function

To obtain quality variables that cannot be measured in real time during the production process but reflect information on the quality of the final product, the batch production process has the characteristics of a strong time-varying nature, non-Gaussian data distribution and high nonlinearity. A locally weighted partial least squares regression quality prediction model (KICA-JITL-LWPLS), based on wavelet kernel function independent meta-analysis with immediate learning, is proposed. The model first measures the similarity between the normalized input data and the historical data and assigns the input data to the group of historical data with high similarity to it, based on the posterior probability of the Bayesian classifier; subsequently, wavelet kernel functions are selected and kernel learning methods are introduced into the independent meta-analysis algorithm. An independent meta-analysis, based on the wavelet kernel function, is performed on the classified input data to obtain probabilistically significant independent sets of variables. Finally, a real-time learning-based LWPLS regression analysis is performed on this variable set to construct a local prediction model for the current sample by calculating the similarity between the local input data. The local predictions from the PLS output are fused with the posterior probability output from the Bayesian classifier to produce the final prediction. The method was used to predict the product concentration and bacteriophage concentration during penicillin fermentation through a simulation platform. The prediction results were basically consistent with the real values, proving that the proposed KICA-JITL-LWPLS quality prediction model, based on wavelet kernel functions, has reliable prediction results.

A Review on Data-Driven Process Monitoring Methods: Characterization and Mining of Industrial Data

Safe and stable operation plays an important role in the chemical industry. Fault detection and diagnosis (FDD) make it possible to identify abnormal process deviations early and assist operators in taking proper action against fault propagation. After decades of development, data-driven process monitoring technologies have gradually attracted attention from process industries. Although many promising FDD methods have been proposed from both academia and industry, challenges remain due to the complex characteristics of industrial data. In this work, classical and recent research on data-driven process monitoring methods is reviewed from the perspective of characterizing and mining industrial data. The implementation framework of data-driven process monitoring methods is first introduced. State of art of process monitoring methods corresponding to common industrial data characteristics are then reviewed. Finally, the challenges and possible solutions for actual industrial applications are discussed.

Nonstationary Discrete Convolution Kernel for Multimodal Process Monitoring

Data-driven process monitoring has benefited from the development and application of kernel transformations, especially when various types of nonlinearity exist in the data. However, when dealing with the multimodality behavior that is frequently observed in the process operations, the most widely used radial basis function (RBF) kernel has limitations in describing process data collected from multiple normal operating modes. In this article, we highlight this limitation via a synthesized example. In order to account for the multimodality behavior and improve the fault detection performance accordingly, we propose a novel nonstationary discrete convolution kernel, which derives from the convolution kernel structure, as an alternative to the RBF kernel. By assuming the training samples to be the support of the discrete convolution, this new kernel can properly address these training samples from different operating modes with diverse properties and, therefore, can improve the data description and fault detection performance. Its performance is compared with RBF kernels under a standard kernel principal component analysis framework and with other methods proposed for multimode process monitoring via numerical examples. Moreover, a benchmark data set collected from a pilot-scale multiphase flow facility is used to demonstrate the advantages of the new kernel when applied to an experimental data set.

A Review of Kernel Methods for Feature Extraction in Nonlinear Process Monitoring

Kernel methods are a class of learning machines for the fast recognition of nonlinear patterns in any data set. In this paper, the applications of kernel methods for feature extraction in industrial process monitoring are systematically reviewed. First, we describe the reasons for using kernel methods and contextualize them among other machine learning tools. Second, by reviewing a total of 230 papers, this work has identified 12 major issues surrounding the use of kernel methods for nonlinear feature extraction. Each issue was discussed as to why they are important and how they were addressed through the years by many researchers. We also present a breakdown of the commonly used kernel functions, parameter selection routes, and case studies. Lastly, this review provides an outlook into the future of kernel-based process monitoring, which can hopefully instigate more advanced yet practical solutions in the process industries.

Fault detection in dynamic plant-wide process by multi-block slow feature analysis and support vector data description

This study describes a dynamic large-scale process fault detection algorithm based on multi-block slow feature analysis by taking advantages of both multi-block algorithms in highlighting the local information and slow feature analysis in extracting the different dynamics of process data. A mutual information-based relevance matrix is first calculated to measure the correlation between any two variables, and then K-means clustering is used to cluster the original variables into several blocks by gathering the variables with similar relevance vectors into the same block. Slow feature analysis is applied in each block. A support vector data description is utilized to give a final decision. The proposed algorithm is tested with a well-known Tennessee Eastman (TE) process. The fault detection results show the efficiency and the superiority of the proposed method as compared to other related methods.