Lin Sui

Chemical processes involve complex operational mechanisms and high-dimensional variables with multiscale dependencies and temporal covariate shifts, posing significant challenges for soft sensing implementation. To address these issues, a self-constrained weighted garrote ordered memory network with adaptive distribution learning is proposed for industrial soft sensor modeling. Specifically, an improved long short-term ordered memory (LSTOM) architecture with a composite hierarchical information update mechanism is designed to capture multiscale dynamic features from industrial time-series data. Meanwhile, an adaptive distribution learning module is embedded into the LSTOM architecture to match high-order feature distributions across different temporal periods via maximally dissimilar segments. With the composite hierarchical update mechanism and embedded adaptive distribution learning, the proposed approach improves generalization in soft sensor modeling under multi-timescale dependencies and temporal covariate shifts. Finally, the shrinkage coefficients of the self-constrained weighted garrote are embedded into the LSTOM’s input weights to eliminate redundant variables and promote structural sparsity. The effectiveness of the proposed approach is validated on a public industrial-scale penicillin fermentation process and a real-world flue gas desulfurization case.

With the development of industrial intelligence, deep learning-based soft sensors are widely used in complex industrial systems to estimate key quality variables in real-time, with long short-term memory (LSTM) networks excelling in nonlinear dynamic modeling. However, conventional LSTM networks face the challenge of redundant memory weight allocation, which hinders efficient capture of critical temporal dependencies. Moreover, the probability distribution of industrial process variables often exhibits dynamic temporal shifts, significantly deteriorating model generalization performance. In this study, a novel long short-term complementary memory network based on temporal distribution matching (TDM-LSTCM) is proposed for product quality prediction in industrial processes. First, a novel memory update rate gate is designed based on LSTM to obtain adaptive update rate factors for different time steps, regulating the information flow rate. Second, the memory update rate gate is embedded into the input and forget gates, leading to the development of a complementary input-forget regulation gate structure to optimize the cell state update mechanism. Furthermore, the TDM algorithm is incorporated into the training process through a dual-stage optimization strategy. During the distribution characterization stage, the training dataset is partitioned into distinct periods with the most pronounced distributional discrepancies. In the distribution matching stage, cross-period discrepancies in hidden states are minimized to facilitate the extraction of common information. Finally, the proposed methodology is validated through a penicillin fermentation process and flue gas desulfurization system. Experimental results demonstrate that the developed approach outperforms other modeling methods, showcasing its promising application potential.

Gated recurrent units (GRUs) have been successfully applied to various industrial soft-sensor tasks. However, their linear coupling constraint in the hidden-state limits information propagation, while high-dimensional inputs and structural redundancies further challenge modeling. To address these issues, a speed-adaptive gated recurrent unit (SAGRU) network with weights dual-garrote regularization (WDG-SAGRU) is proposed for product quality prediction. Specifically, a speed-adaptive gating mechanism is integrated into the GRU update gate through a nonlinear exponential transformation. This mechanism is incorporated into the hidden-state update to break the linear coupling constraint, enhancing information flow and feature extraction. Meanwhile, the dual-garrote shrinkage coefficients are embedded into SAGRU’s input and hidden weight matrices, enabling synchronized optimization of input variable selection and structural sparsity. The proposed WDG-SAGRU is validated against state-of-the-art methods on a numerical example and an operational flue gas desulfurization system in a thermal power plant.

Multi-timescale characteristics, nonlinearity and dynamic features exist extensively in modern industrial processes, impairing the predictive performance and accuracy of soft sensor modeling for primary variables. To address this challenge, a multi-timescale neural network-based soft sensor model, consisting of two modules, is designed for complex industrial processes. The designed temporal convolutional module, which integrates two attention mechanism with multiple parallel temporal convolutional layers using dilated causal convolution, efficiently enhances feature extraction at different time scales. Then, bidirectional long short-term memory networks are developed to capture time features series from the designed temporal convolutional module. The extracted feature variables are fed into the Kolmogorov-Arnold network using B-spline functions, effectively addressing nonlinear data characteristics and improving prediction performance. The superiority of the designed soft sensor model is validated through practical applications in wastewater treatment process and flue gas desulfurization, demonstrating promising potential for real-world applications.

Chemical processes involve complex physical and chemical mechanisms that exhibit slow and fast-varying features and nonstationary characteristics, making it difficult for single model-based methods to satisfactorily extract both slow and nonstationary fast-varying features for soft sensing. To address this issue, we propose a two-stream slow and nonstationary fast feature (TS-SNFF) model. This model includes a slow feature stream (SF-stream) and a nonstationary fast feature stream (NFF-streama). In the SF-stream, an encoder-decoder based Siamese network and a linear mapping layer are used for slow feature extraction. It employs long-short term memory (LSTM) networks as encoder and decoder units. Meanwhile, the NFF-stream utilizes the LSTM, differential LSTM (D-LSTM), and linear mapping layers for nonstationary fast feature extraction. The D-LSTM unit is established by embedding differential operations into the LSTM cell to obtain the nonstationary information. Then, the obtained features are fused using the merging layer, followed by a multilayer perceptron as the regressor. The proposed TS-SNFF model is utilized to address the slow and fast-varying dynamics in nonstationary conditions and nonlinearity problem in chemical processes. The effectiveness and superiority of the proposed method are demonstrated in a numerical example and an industrial process case

Soft sensor, as a significant intelligent inspection technology, has been widely used in modern process industries to achieve effective monitoring and prediction of product quality. However, the redundancy of model inputs and structure in practical industrial process modeling usually results in increased modeling complexity and decreased model performance. In this study, an input variable selection and structure optimization algorithm for long short-term memory (LSTM)-based soft sensor with a dual nonnegative garrote (DNNG) approach was proposed. Firstly, a well-trained initial LSTM model is constructed using the process dataset to capture the temporal dynamic behavior of the industrial process. Secondly, the DNNG algorithm is integrated into the LSTM to reduce the redundancy of input variables and hidden nodes. The strategy efficiently selects the most consequential input variables for the model, and simultaneously simplifies the LSTM structure by eliminating redundant recurrent hidden nodes to reduce model overfitting risk. Moreover, the hyperparameters of the model are determined by combining grid search with blocked cross-validation. Finally, the developed algorithm is compared to other state-of-the-art algorithms using a numerical case and employed to forecast the SO ₂ concentration in the net flue gas emissions of a coal-fired power plant desulphurization system. Comparative results show that the proposed algorithm effectively eliminates redundant variables and streamlines the model structure while presenting better prediction performance than other algorithms.

In the study, we propose an adaptive variable selection algorithm for multi-layer perceptron (MLP)-based soft sensors. The proposed algorithm employs nonnegative garrote (NNG) to shrink the input weights of the trained MLP. To improve the shrinkage efficiency of the NNG, adaptive operators are designed using the mean impact value estimate. Moreover, the adaptive operators are data dependent, and are introduced into the constraints of NNG to make the shrinkage more efficient and effective. Cross-validation and Bayesian information criterion are used to determine the optimal garrote parameter for the NNG. The performance of the algorithm is validated using artificial datasets and a practical industrial application in coke dry quenching (CDQ) systems. The simulation results demonstrate that the adaptive mechanism improves the efficiency and precision of NNG, and has superior performance to other state-of-the-art algorithms. The result of variable selection is consistent with the experience of the field operator, and can provide technical supports for the optimisation and control of the CDQ process.

A novel optimization algorithm for multilayer perceptron- (MLP-) based soft sensors is proposed in this paper. The proposed approach integrates input variable selection and hidden layer optimization on MLP into a constrained optimization problem. The nonnegative garrote (NNG) is implemented to perform the shrinkage of input variables and optimization of hidden layer simultaneously. The optimal garrote parameter of NNG is determined by combining cross-validation with Hannan-Quinn information criterion. The performance of the algorithm is demonstrated by an artificial dataset and the practical application of the desulfurization process in a thermal power plant. Comparative results demonstrated that the developed algorithm could build simpler and more accurate models than other state-of-the-art soft sensor algorithms.