Chemical Engineering and Processing: Process Intensification, 2003
In the literature, several correlations have been proposed for gas holdup prediction in bubble co... more In the literature, several correlations have been proposed for gas holdup prediction in bubble columns. However, these correlations fail to predict gas holdup over a wide range of conditions. Based on a databank of around 3500 measurements collected from the open literature, a correlation for gas holdup was derived using a combination of Dimensional Analysis and artificial neural network (ANN) modeling. The overall gas holdup was found to be a function of four dimensionless groups: Re g , Fr g , Eo /Mo , and r g /r L. Statistical analysis showed that the proposed correlation has an average absolute relative error (AARE) of 15% and a standard deviation of 14%. A comparison with selected correlations in the literature showed that the developed ANN correlation noticeably improved prediction of overall gas holdup. The developed correlation also shows better prediction over a wide range of operating conditions, physical properties, and column diameters, and it predicts properly the trend of the effect of the operating and design parameters on overall gas holdup.
The right of Sandip Kumar Lahiri to be identified as the author of this work has been asserted in... more The right of Sandip Kumar Lahiri to be identified as the author of this work has been asserted in accordance with law.
International Journal of Chemical Reactor Engineering, 2010
This study is motivated by a growing popularity of artificial neural network for process modeling... more This study is motivated by a growing popularity of artificial neural network for process modeling and regression problems. However, many ANN regression application studies are performed by ‘expert' users who have a good understanding of ANN methodology. Since the quality of ANN models depends on a proper setting of ANN architecture and ANN meta-parameters, the main issue for practitioners trying to apply ANN regression is how to set these parameter values (to ensure good generalization performance) for a given data set. Non-expert users face a difficulty in finding an optimum ANN architecture and are often confused about how to choose the ANN meta parameters. The present paper addresses this issue and develops a new hybrid procedure to find the optimum ANN architecture and tunes the ANN parameters, thus relieving the ‘non expert' users. This method incorporates hybrid artificial neural network and differential evolution technique (ANN-DE) for efficient tuning of ANN meta par...
This paper presents artificial intelligence-based process modeling and optimization strategies, n... more This paper presents artificial intelligence-based process modeling and optimization strategies, namely, support vector regression – differential evolution (SVR-DE) for modeling and optimization of catalytic industrial ethylene oxide (EO) reactor. In the SVR-DE approach, a support vector regression model is constructed for correlating process data comprising values of operating and performance variables. Next, model inputs describing process operating variables are optimized using Differential Evolution (DE) with a view to maximize the process performance. DE possesses certain unique advantages over the commonly used gradient-based deterministic optimization algorithms. The SVR-DE is a new strategy for chemical process modeling and optimization. The major advantage of the strategy is that modeling and optimization can be conducted exclusively from the historic process data wherein the detailed knowledge of process phenomenology (reaction mechanism, kinetics, etc.) is not required. Us...
Often it is time consuming to monitor the plant condition in modern complex process industries as... more Often it is time consuming to monitor the plant condition in modern complex process industries as there is abundance of instrumentation that measure thousands of process variables in every few seconds. This has caused a "data overload" and due to the lack of appropriate analyses very little is currently being done to utilize this wealth of information. Fortunately, in process, groups of variables often moves together because more than one variable may be measuring the same driving principle governing the behavior of the process. Multivariate statistical methods such as Principal Component Analysis (PCA) are capable of compressing the information down into low dimensional spaces which retain most of the information. Using this method of statistical data compression a multivariate monitoring procedure has been developed to efficiently monitor the performance of large processes and to rapidly detect and identify important process changes. A graphical interface was developed w...
International Journal of Chemical Reactor Engineering
The present work emphasizes the development of a generic methodology that addresses the core issu... more The present work emphasizes the development of a generic methodology that addresses the core issue of any running chemical plant, i.e., how to maintain a delicate balance between profit and environmental impact. Here, an ethylene oxide (EO) production plant has been taken as a case study. The production of EO takes place in a multiphase catalytic reactor, the reliable first principle-based model of which is still not available in the literature. Artificial neural network (ANN) was therefore applied to develop a data-driven model of the complex reactor with the help of actual industrial data. The model successfully built up a correlation between the catalyst selectivity and temperature with other operational parameters. A hybrid multi-objective metaheuristic optimization technique, namely ANN-multi-objective genetic algorithm (MOGA) algorithm was used to develop a Pareto diagram of selectivity versus reactor temperature. The Pareto diagram will help the plant engineers to make a stra...
Chemical Engineering and Processing: Process Intensification, 2003
In the literature, several correlations have been proposed for gas holdup prediction in bubble co... more In the literature, several correlations have been proposed for gas holdup prediction in bubble columns. However, these correlations fail to predict gas holdup over a wide range of conditions. Based on a databank of around 3500 measurements collected from the open literature, a correlation for gas holdup was derived using a combination of Dimensional Analysis and artificial neural network (ANN) modeling. The overall gas holdup was found to be a function of four dimensionless groups: Re g , Fr g , Eo /Mo , and r g /r L. Statistical analysis showed that the proposed correlation has an average absolute relative error (AARE) of 15% and a standard deviation of 14%. A comparison with selected correlations in the literature showed that the developed ANN correlation noticeably improved prediction of overall gas holdup. The developed correlation also shows better prediction over a wide range of operating conditions, physical properties, and column diameters, and it predicts properly the trend of the effect of the operating and design parameters on overall gas holdup.
The right of Sandip Kumar Lahiri to be identified as the author of this work has been asserted in... more The right of Sandip Kumar Lahiri to be identified as the author of this work has been asserted in accordance with law.
International Journal of Chemical Reactor Engineering, 2010
This study is motivated by a growing popularity of artificial neural network for process modeling... more This study is motivated by a growing popularity of artificial neural network for process modeling and regression problems. However, many ANN regression application studies are performed by ‘expert' users who have a good understanding of ANN methodology. Since the quality of ANN models depends on a proper setting of ANN architecture and ANN meta-parameters, the main issue for practitioners trying to apply ANN regression is how to set these parameter values (to ensure good generalization performance) for a given data set. Non-expert users face a difficulty in finding an optimum ANN architecture and are often confused about how to choose the ANN meta parameters. The present paper addresses this issue and develops a new hybrid procedure to find the optimum ANN architecture and tunes the ANN parameters, thus relieving the ‘non expert' users. This method incorporates hybrid artificial neural network and differential evolution technique (ANN-DE) for efficient tuning of ANN meta par...
This paper presents artificial intelligence-based process modeling and optimization strategies, n... more This paper presents artificial intelligence-based process modeling and optimization strategies, namely, support vector regression – differential evolution (SVR-DE) for modeling and optimization of catalytic industrial ethylene oxide (EO) reactor. In the SVR-DE approach, a support vector regression model is constructed for correlating process data comprising values of operating and performance variables. Next, model inputs describing process operating variables are optimized using Differential Evolution (DE) with a view to maximize the process performance. DE possesses certain unique advantages over the commonly used gradient-based deterministic optimization algorithms. The SVR-DE is a new strategy for chemical process modeling and optimization. The major advantage of the strategy is that modeling and optimization can be conducted exclusively from the historic process data wherein the detailed knowledge of process phenomenology (reaction mechanism, kinetics, etc.) is not required. Us...
Often it is time consuming to monitor the plant condition in modern complex process industries as... more Often it is time consuming to monitor the plant condition in modern complex process industries as there is abundance of instrumentation that measure thousands of process variables in every few seconds. This has caused a "data overload" and due to the lack of appropriate analyses very little is currently being done to utilize this wealth of information. Fortunately, in process, groups of variables often moves together because more than one variable may be measuring the same driving principle governing the behavior of the process. Multivariate statistical methods such as Principal Component Analysis (PCA) are capable of compressing the information down into low dimensional spaces which retain most of the information. Using this method of statistical data compression a multivariate monitoring procedure has been developed to efficiently monitor the performance of large processes and to rapidly detect and identify important process changes. A graphical interface was developed w...
International Journal of Chemical Reactor Engineering
The present work emphasizes the development of a generic methodology that addresses the core issu... more The present work emphasizes the development of a generic methodology that addresses the core issue of any running chemical plant, i.e., how to maintain a delicate balance between profit and environmental impact. Here, an ethylene oxide (EO) production plant has been taken as a case study. The production of EO takes place in a multiphase catalytic reactor, the reliable first principle-based model of which is still not available in the literature. Artificial neural network (ANN) was therefore applied to develop a data-driven model of the complex reactor with the help of actual industrial data. The model successfully built up a correlation between the catalyst selectivity and temperature with other operational parameters. A hybrid multi-objective metaheuristic optimization technique, namely ANN-multi-objective genetic algorithm (MOGA) algorithm was used to develop a Pareto diagram of selectivity versus reactor temperature. The Pareto diagram will help the plant engineers to make a stra...
Uploads
Papers by Sandip Lahiri