Soft sensors have been increasingly applied for quality prediction in complex industrial processes, which often have different scales of topology and highly coupled spatiotemporal features. However, the existing soft sensing models usually face difficulties in extracting the multiscale local spatiotemporal features in multicoupled complex process data and harnessing them to their full potential to improve the prediction performance. Therefore, a multiscale attention-based CNN (MSACNN) is proposed in this article to alleviate such problems. In MSACNN, convolutional kernels of different sizes are first designed in parallel in the convolutional layers, which can generate feature maps containing local spatiotemporal features at different scales. Meanwhile, a channel-wise attention mechanism is designed on the feature maps in parallel to get their attention weights, representing the significance of the local spatiotemporal feature at different scales. The superiority of the proposed MSACNN over the other state-of-the-art methods is validated through the performance evaluation in two real industrial processes.

In complex process industries, multivariate time sequences are omnipresent, whose nonlinearities and dynamics present two major challenges for soft sensing of important quality variables. Consequently, due to the potent representational capabilities, nonlinear dynamic models like gated recurrent unit (GRU) and long short-term memory (LSTM) networks have been used for data sequence modeling. Though it is a common occurrence in many industrial plants, data series with heterogeneous sample intervals and missing values cannot be directly handled by these dynamic algorithms. To this end, attention-based interval-aided networks (AIA-Net) are proposed in this article to adaptively model the temporal information for heterogeneous sampling sequences with missing values in the processes industry. It includes two main mechanisms, which are named attention-based time-aware dynamic imputation and interval-aided time-aware network, respectively. The reduction rate is introduced by the attention-based time-aware dynamic imputation to apply the effects of time intervals and is used in the imputation of missing data. The interval-aided time-aware network includes time intervals in the model structure and uses a sampling interval gate to correct the temporal correlations in time series. The proposed AIA-Net is successfully applied to a real hydrocracking process to predict the C5 and C6 content in the light naphtha.

Soft sensor technology is widely used in the industries to handle highly nonlinear, dynamic, time-dependent sequence data of industrial processes for predicting the key variables associated with auxiliary process variables. Many existing soft sensor algorithms based on deep-learning are able to build complex nonlinear models but ignore the dynamic characteristics of processes. The long short-term memory (LSTM) neural network is exploited to solve the modeling issue which is related to strong time-varying features. In this paper, a novel multi-step sequence-to-sequence model based on attention LSTM (MA-LSTM) neural networks is proposed to improve the soft sensor modeling performance of industrial processes with strong dynamics and nonlinearity. The LSTM-based encoder-decoder architecture with the attention mechanism is applied to extract inherent characteristics relevant to the quality variables and capture both the long-term and short-term dependences of sequence data. Instead of the traditional sequence-to-point quality prediction architecture, the improved architecture is designed to predict multi-step quality variables, and the intermediate results are fed into an additional one-dimensional weighted convolution module to obtain accurate prediction results. The superiority of the proposed framework is demonstrated through a debutanizer column case and a sulfur recovery process.

In process industries, accurate prediction of critical quality variables is particularly important for process control and optimization. Usually, soft sensors have been developed to estimate the quality variables via process variables. However, there are often process variables that are far apart in topology but have high correlations, presenting challenges for such feature learning in soft sensor. To solve these difficulties, a variable correlation analysis based convolutional neural network (VCA-CNN) is proposed in this paper for far topological feature extraction, which generates three-channel data samples containing natural topology features, linear topology features, and nonlinear topology features through Pearson and Spearman correlation to adequately focus on far-topology features. In the feature extraction stage, the VCA-CNN employs a special row convolution kernel to extract far-topological features. Then, normal convolutional and pooling operations are alternately processed on the output feature maps. This approach overcomes the limitation that traditional convolutional operations mainly extract local correlations of process variables. The effectiveness and superiority of the VCA-CNN are validated in the hydrocracking process and debutanizer column.

For industrial processes, it is significant to carry out the dynamic modeling of data series for quality prediction. However, there are often different sampling rates between the input and output sequences. For the most traditional data series models, they have to carefully select the labeled sample sequence to build the dynamic prediction model, while the massive unlabeled input sequences between labeled samples are directly discarded. Moreover, the interactions of the variables and samples are usually not fully considered for quality prediction at each labeled step. To handle these problems, a hierarchical self-attention network (HSAN) is designed for adaptive dynamic modeling. In HSAN, a dynamic data augmentation is first designed for each labeled step to include the unlabeled input sequences. Then, a self-attention layer of variable level is proposed to learn the variable interactions and short-interval temporal dependencies. After that, a self-attention layer of sample level is further developed to model the long-interval temporal dependencies. Finally, a long short-term memory network (LSTM) network is constructed to model the new sequence that contains abundant interactions for quality prediction. The experiment on an industrial hydrocracking process shows the effectiveness of HSAN.

Fault detection and fault identification are consecutive steps of multivariate statistical process monitoring. In recent years, increasing attention has been paid to process dynamics. For dynamic process modeling, canonical variate analysis (CVA) extracts process dynamics effectively. However, process noises are not well analyzed in traditional CVA and corresponding fault identification methods are less studied. To solve these issues, a variational Bayesian CVA (VBCVA) model is proposed for dynamic process monitoring. Through a probabilistic perspective, the inevitable noises in realistic industrial processes can be captured in the new model. Moreover, the proposed model is further extended in the variational Bayesian framework to overcome the common problems in probabilistic methods. Besides, an improved fault identification approach based on fault relevance is introduced, which avoids the smearing effect caused by data reconstruction. Finally, the feasibility of the proposed process monitoring scheme is verified on the TE benchmark and a real wastewater treatment process.