Modified neural network based cascaded control for product composition of reactive distillation

Abstract

In this research work, neural network based single loop and cascaded control strategies, based on Feed Forward Neural Network trained with Back Propagation (FBPNN) algorithm is carried out to control the product composition of reactive distillation. The FBPNN is modifi ed using the steepest descent method. This modifi cation is suggested for optimization of error function. The weights connecting the input and hidden layer, hidden and output layer is optimized using steepest descent method which causes minimization of mean square error and hence improves the response of the system. FBPNN, as the inferential soft sensor is used for composition estimation of reactive distillation using temperature as a secondary process variable. The optimized temperature profi le of the reactive distillation is selected as input to the neural network. Reboiler heat duty is selected as a manipulating variable in case of single loop control strategy while the bottom stage temperature T9 is selected as a manipulating variable for cascaded control strategy. It has been observed that modifi ed FBPNN gives minimum mean square error. It has also been observed from the results that cascaded control structure gives improved dynamic response as compared to the single loop control strategy.

Inferential Control of a Distillation Column through the Successive Update of the Soft-sensor and Control Algorithm

“…Most studies on inferential control have not considered the problems mentioned above, nor have they clearly specied the operating conditions under which the model construction data was obtained (Abusnina et al, 2014;Jiang et al, 2015;Bosca et al, 2016;Cao et al, 2017;Escalante-Sánchez et al, 2018;Jana and Banerjee, 2018;Guan et al, 2019). In other studies, the model construction data were collected under impractical conditions (Pannocchia and Brambilla, 2003;Bahar and Özgen, 2010;Vijaya Raghavan et al, 2011;Chiewchanchairat et al, 2015;Sakhre et al, 2016;Bidar et al, 2018;Clementi et al, 2018;Qian et al, 2019). Some authors have obtained the data used for model construction from plant operations without a control system (Pannocchia and Brambilla, 2003;Bahar and Özgen, 2010;Vijaya Raghavan et al, 2011;Chiewchanchairat et al, 2015;Sakhre et al, 2016).…”

“…In other studies, the model construction data were collected under impractical conditions (Pannocchia and Brambilla, 2003;Bahar and Özgen, 2010;Vijaya Raghavan et al, 2011;Chiewchanchairat et al, 2015;Sakhre et al, 2016;Bidar et al, 2018;Clementi et al, 2018;Qian et al, 2019). Some authors have obtained the data used for model construction from plant operations without a control system (Pannocchia and Brambilla, 2003;Bahar and Özgen, 2010;Vijaya Raghavan et al, 2011;Chiewchanchairat et al, 2015;Sakhre et al, 2016). Qian et al (2019) obtained data for model construction under persistent excitation.…”