Data-driven soft sensors have been widely adopted in industrial processes to learn hidden knowledge automatically from process data, then to monitor difficult-to-measure quality variables. However, to extract and utilize useful dynamic latent features accurately for efficient quality estimations remains one of the most important research issues in soft sensor modeling. In this article, a supervised bidirectional long short-term memory (SBiLSTM) is proposed for data-driven dynamic soft sensor modeling. The SBiLSTM incorporates extended quality information with a moving window up to $k$ time steps and enhances learning efficiency by bidirectional architecture. With this novel structure, the SBiLSTM can extract and utilize nonlinear dynamic latent information from both process variables and quality variables, then further improve the prediction performance significantly. The effectiveness of the proposed SBiLSTM network-based soft sensor model is demonstrated through two case studies on a debutanizer column process and an industrial wastewater treatment process. Results show that the SBiLSTM outperforms state-of-the-art and traditional deep learning-based soft sensor models.

Proper monitoring of quality-related variables in industrial processes is nowadays one of the main worldwide challenges with significant safety and efficiency implications.Variational Bayesian mixture of canonical correlation analysis (VBMCCA)-based process monitoring method was proposed in this paper to predict and diagnose these hard-to-measure quality-related variables simultaneously. Use of Student's t-distribution, rather than Gaussian distribution, in the VBMCCA model makes the proposed process monitoring scheme insensitive to disturbances, measurement noises, and model discrepancies. A sequential perturbation (SP) method together with derived parameter distribution of VBMCCA is employed to approach the uncertainty levels, which is able to provide a confidence interval around the predicted values and give additional control line, rather than just a certain absolute control limit, for process monitoring. The proposed process monitoring framework has been validated in a wastewater treatment plant (WWTP) simulated by benchmark simulation model with abrupt changes imposing on a sensor and a real WWTP with filamentous sludge bulking. The results show that the proposed methodology is capable of detecting sensor faults and process faults with satisfactory accuracy.

In closed-loop feedback control systems, faults are propagated through the feedback link, which eventually leads to the abnormality of the entire system. Generally, it is very difficult to identify the faults of systems under the influence of the closed-loop feedback link. The existence of redundancy improves the reliability of the system. Meanwhile, it also poses new challenges to the fault diagnosis of multiple redundant systems. In this regard, a causality-based method is proposed for the fault diagnosis of closed-loop feedback control system with multiple modular redundancy. The dynamic Bayesian networks for fault diagnosis are established based on sensor data and system parameters. The networks consist of four layers, which are sensors, performances, monitors, and faults, respectively. Furthermore, the conditional probabilities of the fault nodes are calculated by Noisy-OR and Noisy-MAX models. The proposed method can dynamically evaluate system performance and integrate other monitoring information as evidence to assist faults diagnosis and location. A double modular redundant control system for a subsea blowout preventer is used as a case to demonstrate the proposed method, and the results show that the proposed method has high accuracy. The influence of sampling frequency, noise, and redundancy mode on diagnosis results is studied and discussed in the case study.