Accurate and real-time quality prediction to realize the optimal process control at a competitive price is an important issue in Industrial 4.0. This paper shows a successful engineering application of how smart soft sensors can be combined with machine learning technique to significantly save human resources and improve performance under complex industrial conditions. Ensemble learning based soft sensors succeed in capturing complex nonlinearities, frequent dynamic changes, as well as time-varying characteristics in industrial processes. However, local model regions under traditional ensemble modelling methods are highly dependent on labeled data samples and, hence, their prediction accuracy might get affected when labeled samples are limited. A novel active learning (AL) framework upon the ensemble Gaussian process regression (GPR) model is proposed for smart soft sensor design in order to overcome this drawback. Firstly, to iteratively select the most informative unlabeled samples for labeling with hierarchical sampling based AL strategy, to then apply Gaussian mixture model (GMM) technique to autonomously identify operation phases, to further construct local GPR models without human involvement, and finally to integrate the base predictors by applying the Bayesian fusion strategy. Comparative studies for the penicillin fermentation process demonstrate the reliability and superiority of the recommended smart soft sensing. The cost of human annotation can be dramatically reduced by at least half while the prediction performance simultaneously keeps high.

This study proposes a multi-block fault detection method based on fault-sensitive slow features for large-scale dynamic industrial processes. Firstly, slow feature analysis (SFA) can effectively extract the process dynamic information. However, the slowest changing features may not contain more fault information. Thus, through the analysis of 2 T statistic in SFA-based process monitoring model, a fault sensitivity coefficient is defined as a new slow feature sorting criterion to select the most sensitive slow features to fault in each variable direction. Then, considering the unknown characteristics of the fault in the real-time monitoring process, the monitoring model is established for each dimension of variables based on the multi-block strategy. Finally, the support vector data description is used as a fusing method to integrate the statistics calculated in each sub-block to obtain an intuitive detection result. The effectiveness and superiority of the proposed strategy are demonstrated by the experiments on Tennessee Eastman benchmark process and an actual blast furnace ironmaking process.