Smart Soft Sensor Design with Hierarchical Sampling Strategy of Ensemble Gaussian Process Regression for Fermentation Processes

Abstract

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.

Industrial Semi-Supervised Dynamic Soft-Sensor Modeling Approach Based on Deep Relevant Representation Learning

“…This metric measures the overall expected deviation between predicted and actual values in a squared error sense [ 56 ]. Therefore, RMSE highly reflects the prediction performance and reliability of soft sensors to be tested [ 12 ]. A small RMSE score indicates better generalization and prediction performance.…”

“…Basically, soft-sensing uses secondary variables (i.e., easy-to-measure variables) to estimate primary variables (i.e., hard-to-measure variables) [ 4 , 5 ]. Countless soft sensors have been designed using traditional methods: principal component regression (PCR) [ 6 , 7 ], partial least square (PLS) [ 8 , 9 ], support vector machine (SVM) [ 10 , 11 ], gaussian process regression (GPR) [ 12 , 13 ], artificial neural network (ANN) [ 14 , 15 ], and so on.…”

A Fermentation State Marker Rule Design Task in Metabolic Engineering

“…The use of model-based correction strategies to avoid the shift from the desired steady state (Section 5) becomes valuable or even irreplaceable in the case of unstable processes. It is important that marker rules are derived from genome-scale mechanistic models of cellular metabolism in contrast to soft sensors, where black-box approaches dominate [2,3,37]. A mechanistic explanation of marker rules also enables better chances to interpret deviations from the rule, and may lead to improvements in the genome scale model if necessary.…”

Machine Learning Algorithms Used in PSE Environments: A Didactic Approach and Critical Perspective

“…Different variations of random trees and random forest are coupled for addressing the Diels–Alder reaction considering the influence of solvent structure on the rate of the reaction in Dev et al Moreover, Wang et al propose an ensemble using Gaussian process regression as a base model for addressing the quality variable prediction applied to the TEP case. Polymerization processes are addressed using kernel learning strategies for obtaining nonlinear modeling of properties in Liu et al In addition, Sheng et al present an hierarchical approach based on the Gaussian regression process for modeling penicillin fermentation process.…”