Data-driven soft-sensors for online monitoring of batch processes with different initial conditions

Abstract

A soft-sensing methodology applicable to batch processes operated under changeable initial conditions is presented. These cases appear when the raw materials specifications differ from batch to batch, different production scenarios should be managed, etc. The proposal exploits the capabilities of the machine learning techniques to provide practical soft-sensing approach with minimum tuning effort in spite of the fact that the inherent dynamic behavior of batch systems are tracked through other online indirect measurements. Current data modelling techniques have been also tested within the proposed methodology to demonstrate its advantages. Two simulation casestudies and a pilot-plant case-study involving a complex batch process for wastewater treatment are used to illustrate the problem, to assess the modelling approach and to compare the modelling techniques. The results reflect a promising accuracy even when the training information is scarce, allowing significant reductions in the cost associated to batch processes monitoring and control.

Dynamic Surrogate Modeling for Multistep-ahead Prediction of Multivariate Nonlinear Chemical Processes

“…[6,7]. Also, since process FPMs typically do not take into account the physical characteristics of mechanical and electrical components, connections and piping, which remarkably influence the real process, the accuracy of the FPMs predictions are reduced [8,9]. In other cases, the development of a detailed analytical FPM is conceptually difficult or even unaffordable, due to the limited knowledge about the nonlinear behaviors and complex phenomena characterizing the process, such as reaction kinetics, thermodynamic relationships, heat and mass transfer, etc.…”

“…A set of multivariate dynamic models is to be constructed, which describes the step-ahead evolution of the tanks levels ℎ 1( +1) , ℎ 2( +1) , ℎ 3( +1) , see Eqs. (9). The same general procedure described in Section 3 and the application details illustrated in Section 4.1 are systematically followed in this case, too.…”

Modeling, instrumentation, automation, and optimization of water resource recovery facilities (Review of 2018 Literature) DIRECT

“…The kinetic energy of vibrations induced in water pipes and converted by electromagnetic or piezoelectric transducers were one of the proposed solutions along with a thermoelectric generator that produces electrical energy using the Seebeck effect. A soft‐sensing methodology applicable to batch processes operated under changeable initial conditions was developed by Shokry et al (2018). The approach used the capabilities of the machine learning techniques to provide practical soft‐sensing methodology with minimum tuning effort.…”

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

“…Data-driven models named soft sensors have been developed as a successful alternative to the above-mentioned issue [ 3 ]. 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.…”