Fahim Uddin is currently pursuing his doctorate in chemical engineering at Universiti Teknologi Petronas. He has served as a lecturer in chemical engineering department at NED University of engineering and technology, Karachi, Pakistan for almost four years, from which he is on study leave. He received his B.E. and M.E. degrees from NED University as well. He has represented the institution at and contributed to some national and international conferences. His expertise in modeling & simulation combined with his competence i.e. 91 percentile in quantitative reasoning at GRE 2014 indicate that he will make his mark in this field. Supervisors: Dr. Lemma Dendena Tufa and Dr. Abdulhalim Shah Maulud
Fault detection and diagnosis (FDD) in process industries is an important task for efficient proc... more Fault detection and diagnosis (FDD) in process industries is an important task for efficient process monitoring and plant safety. It is also essential for improving product quality and reducing production cost by reducing process downtime. Real-time multiscale classification of faults plays a vital role in process monitoring. However, some major issues such as high correlation, complexity, and nonlinearity of data are yet to be addressed. In this paper, a fault diagnosis approach based on multikernel support vector machines is proposed to classify the internal and external faults such as reflux failure, change in reboiler duty, column tray upsets, and change in feed composition, flow, and temperature in a distillation column. The data set is generated using Aspen plus dynamics simulation at normal and faulty states. The classification has been done by various methods such as decision tree, K-nearest neighbors, linear discriminant analysis, artificial neural network, subspace discriminant, and multikernel support vector machine. It is observed that the SVM based diagnostic system provides more accurate root cause isolation. The proposed MK-SVM method was evaluated by using the confusion matrix as the performance evaluator. The result showed that the proposed model has a high FDR which is 99.77% and a very low FAR, i.e., 0.23%.
Early detection of anomalies can assist to avoid major losses in term of product degradation, mac... more Early detection of anomalies can assist to avoid major losses in term of product degradation, machines' damages as well as human health issues. This research aims to use Artificial Neural Network to recognize anomalies in the distillation column. The pilot scale distillation column for the ethanol-water system is selected for the study. Faults are generated by variation in feed rate, feed composition and reboiler duty using Aspen Plus® dynamic simulation. The effect of these faults on process variables i.e. changes in distillate and bottom composition, distillate and bottom temperature, bottom flow rate, and the pressure drop is observed. The network is trained using back propagation algorithm to determine root mean square error (RMSE). Based on RMSE minimization, the (6-8-6) net serves as the best choice for the case studied for efficient fault detection. The presented techniques are general in nature and easily applicable to various other industrial problems.
Fault detection in the process industries is one of the most challenging tasks. It requires timel... more Fault detection in the process industries is one of the most challenging tasks. It requires timely detection of anomalies which are present with noisy measurements of a large number of variable, highly correlated data with complex interactions and fault symptoms. This study proposes the robust fault detection method for the distillation column. Fault detection and diagnosis (FDD) for process monitoring and control has been an effective field of research for two decades. This area has been used widely in sophisticated engineering design applications to ensure the proper functionality and performance diagnosis of advanced and complex technologies. Robust fault detection of the realistic faults in distillation column in dynamic condition has been considered in this study. For early detection of faults, the model is based on nonlinear autoregressive with exogenous input (NARX) network. Tapped delays lines (TDLs) have been used for the input and output sequences. A case study was carried out with three different fault scenarios, i.e., valve sticking at reflux and reboiler, and tray upset. These faults would cause the product degradation. The normal data (no fault) is used for the training of neural network in all three cases. It is shown that the proposed algorithm can be used for the detection of both internal and external faults in the distillation column for dynamic system monitoring and to predict the probability of failure.
Knowledge of dynamic behavior of process plants is essential in designing, commissioning and oper... more Knowledge of dynamic behavior of process plants is essential in designing, commissioning and operating process plants. Several advanced model based technologies require explicit dynamic models of the process plant. However, even the development of the dynamic model of a simple unit operation is very daunting, costly and time-consuming. This paper presents a rigorous dynamic simulation using Aspen Plus. The data from the dynamic simulation is then used to develop transfer function model of the column. This case study considers a distillation column with an ethanol-water feed. A 2×3 distillation column transfer function model is developed from step responses of the Aspen plus dynamic simulation, i.e. with feed valve opening %, reflux flow rate and reboiler duty as input and top and the bottom compositions as an output. A separate SISO transfer function from condenser duty to column pressure is also developed.
The Artificial Neural Network (ANN) modelling is presented for the steam gasification of palm ker... more The Artificial Neural Network (ANN) modelling is presented for the steam gasification of palm kernel shell using CaO adsorbent and coal bottom ash as a catalyst. The effect of the parameters such as; temperature, CaO/biomass ratio and Coal bottom ash wt.% at fixed steam/biomass ratio and steam/ biomass ratio at the fixed temperature on product gas composition of H 2 , CO, CO 2 , and CH 4 are modelled using ANN. The effect of parameters is used as an input, while the gas compositions, syngas yield, LHV gas and HHV gas of gas as the output of the network. Back propagation algorithm has been used for the training with 7 neurons in the hidden layer. Hence, the selected ANN architecture was (2-7-1). The gas composition predicted by the ANN model are compared with experimental results obtained from pilot scale gasification system that has been reported in our previous study. The ANN predicted results show high agreement with the published experimental values with the coefficient of determination R 2 ¼ 0.998 for almost all the cases, i.e., the effect of parameters. RMSE, MAD, and AARE have been reported to be very insignificant for the predicted and experimental values.
Industrial & Engineering Chemistry Research, Oct 2018
Many control-relevant systems present the challenges of nonlinearity, directionality and ill-cond... more Many control-relevant systems present the challenges of nonlinearity, directionality and ill-conditioning for the control systems, and exhibit poor controller performance. This study proposes a Nonlinearity Index to quantify the extent of nonlinearity of such systems. A dynamic nonlinear model of a pilot-scale distillation column operating near the azeotropic region was simulated using Aspen Plus Dynamics. A comparison of the results is made with the Nonlinearity Measure proposed by Du et al. Results show that a significant increase is exhibited in the proposed Nonlinearity Index values of the system as the system moves toward the azeotropic region. Prediction errors of the linear models are also shown to be correlated to the proposed index. Therefore, the proposed Nonlinearity Index is consistent with the existing indicators of nonlinearity, and thus its measurements of system nonlinearity are reliable. Controller performance for the system at higher values of the proposed index further presents its efficacy.
High-purity distillation columns present challenges of nonlinearity, interactions, directionality... more High-purity distillation columns present challenges of nonlinearity, interactions, directionality, and ill-conditioning for the control systems and thus exhibit poor controller performance. Effects of these issues on the performance of model pre-dictive control in operations in near-azeotropic regions are investigated. A dynamic model of a pilot-scale distillation column is simulated and controlled by a linear model predictive controller. As the system moves towards the azeotropic region, the nonlinearity, directionality, and interactions of the system increase. A significant decline is observed in the system control performance since the controller incurs overshoots and offsets for negative and positive setpoint changes. The integral time absolute errors are substantially higher near the azeotropic region. Only a small operating range is able to ensure sufficient controllability near the azeotropic region.
Objectives: Closed-loop identification is reported to provide better results for identification o... more Objectives: Closed-loop identification is reported to provide better results for identification of systems for control applications. This study conducted closed-loop identification on an Aspen Plus® dynamic simulation based on a pilot-plant distillation column to develop discrete-time linear time-invariant models. Methods/Statistical Analysis: Identification data was generated using set-point perturbations in control variables under proportional-integral control. Identified models were compared with a model identified using open-loop data using 20-step ahead predictions. Findings: Results indicate that closed-loop identification provides more precise prediction models than open-loop identification in this case study. 20-step predictions for closed loop models exceeded 90% fit, whereas the open loop model predictions provided a 70% fit and missed the steady-state values. Application/Improvements: Thus closed-loop identification is more appropriate for applications in model-based controllers.
Process model is the kernel element of Model Predictive Control (MPC) system. It is always desira... more Process model is the kernel element of Model Predictive Control (MPC) system. It is always desirable to get a model as accurate as the actual facility or plant to reduce the built-in mismatch. With the passage of time, the mismatch between model and plant increases, which results in degradation of MPC performance. To rectify mismatches through plant re-identification is exorbitant and time consuming. Hence, mismatch detection is critical to isolate the faulty sub models to avoid complete re-identification. Badwe et al. proposed a method using partial correlation to isolate and detect plant-model mismatch which uses dynamic models in the decorrelation step. This study extends his work by comparing the performances of Autoregressive Exogenous Input (ARX) model and Auto-Regressive Moving Average with Exogenous Input (ARMAX) model for detection of model-plant mismatch. Wood and Berry binary distillation column is used as a case study to demonstrate the application of the ARX and ARMAX models in mismatch detection. Results show that ARMAX models provide higher accuracy with less model order as compared to ARX. This results in less computational complexity and less processing power required in the MPC, hence improving its efficiency.
Boilers are amongst the most important industrial utilities with their performance crucial in smo... more Boilers are amongst the most important industrial utilities with their performance crucial in smooth functioning of any industrial facility. Poor boiler performance and frequent breakdowns can signiiicantly reduce equipment efficiency translating to overall detrimental performance and prooit reduction. This paper focuses on achieving superior feed water treatment to maximize boiler efficiency. The work is to explore the adaptability of ion-exchange enabled water treatment in both small and large setups, particularly those working with limited throughputs, for which membrane technologies are not much useful both in terms of capital costs and unavailability of high pressures as needed with such systems. A mini water treating facility comprising of iltration and ion exchange units was designed and run with water to determine the extent of puriiication. Comparative analysis of three major properties: pH, hardness and TDS levels of raw and product water streams revealed a signiiicant reduction in values in the product demonstrating the practicality and effectiveness of the method.
A steady state Aspen Plus® simulation is used to propose and analyse a novel indirect gasificatio... more A steady state Aspen Plus® simulation is used to propose and analyse a novel indirect gasification model using steam as the gasifying medium. A sensitivity study is carried out by varying the steam-to-coal ratio. Part of the Syngas produced is combusted in the furnace to provide flue gases at high temperature which is then passed through the bayonets to provide reaction heat. RK-Soave method is used for the evaluation of physical properties of mixed conventional components and CISOLID components in the simulation. The feedstock used in the simulation is indigenous Thar coal from Pakistan, where large coal reserves are present. It was found that a lower steam-to-coal ratio increases the heat content of the syngas produced. Also, the carbon conversion undergoes a maximum at the ratio 2.0. As we increase the steam-to-coal ratio, the yields are better but inevitably reduce the quality of syngas produced. Carbon conversions and H2/CO ratios as high as 95.76% and 3.09 respectively were observed with higher steam to carbon ratios. However, lower ratios provide high yields (69.00%) and cold gas efficiencies (54.41%). Based on these results, it can be said that this parameter significantly influences the syngas quality and processing costs and the diversity of trends suggest a more detailed analysis for optimisation of the process.
In view of limited liquid fuels, in terms of crude oil reserves and to reduce the use of constant... more In view of limited liquid fuels, in terms of crude oil reserves and to reduce the use of constantly and rapidly diminishing natural gas reserves, researchers are attracted towards Fisher Tropsch reaction. Aspen Plus® has become reliable, acquainted and recognized processes modeling software, extensively in practice for coal and biomass gasification processes. It contains different physical property packages that are useful for solid handling. Aspen Plus® model has been proposed to develop a better understanding of the process for geometric analysis of gasifier. This simulation presents an alternate technology for conventional coal gasification to improve the performance of process by varying geometry of gasifier. The Purpose of this study, is entirely focus on the production of synthesis gas from coal, through a process of indirect gasification and using only steam as the gasifying medium. The book serves as reference material for students, engineers and scientists working in the area of syngas production and coal gasification. .
Fault detection and diagnosis (FDD) in process industries is an important task for efficient proc... more Fault detection and diagnosis (FDD) in process industries is an important task for efficient process monitoring and plant safety. It is also essential for improving product quality and reducing production cost by reducing process downtime. Real-time multiscale classification of faults plays a vital role in process monitoring. However, some major issues such as high correlation, complexity, and nonlinearity of data are yet to be addressed. In this paper, a fault diagnosis approach based on multikernel support vector machines is proposed to classify the internal and external faults such as reflux failure, change in reboiler duty, column tray upsets, and change in feed composition, flow, and temperature in a distillation column. The data set is generated using Aspen plus dynamics simulation at normal and faulty states. The classification has been done by various methods such as decision tree, K-nearest neighbors, linear discriminant analysis, artificial neural network, subspace discriminant, and multikernel support vector machine. It is observed that the SVM based diagnostic system provides more accurate root cause isolation. The proposed MK-SVM method was evaluated by using the confusion matrix as the performance evaluator. The result showed that the proposed model has a high FDR which is 99.77% and a very low FAR, i.e., 0.23%.
Early detection of anomalies can assist to avoid major losses in term of product degradation, mac... more Early detection of anomalies can assist to avoid major losses in term of product degradation, machines' damages as well as human health issues. This research aims to use Artificial Neural Network to recognize anomalies in the distillation column. The pilot scale distillation column for the ethanol-water system is selected for the study. Faults are generated by variation in feed rate, feed composition and reboiler duty using Aspen Plus® dynamic simulation. The effect of these faults on process variables i.e. changes in distillate and bottom composition, distillate and bottom temperature, bottom flow rate, and the pressure drop is observed. The network is trained using back propagation algorithm to determine root mean square error (RMSE). Based on RMSE minimization, the (6-8-6) net serves as the best choice for the case studied for efficient fault detection. The presented techniques are general in nature and easily applicable to various other industrial problems.
Fault detection in the process industries is one of the most challenging tasks. It requires timel... more Fault detection in the process industries is one of the most challenging tasks. It requires timely detection of anomalies which are present with noisy measurements of a large number of variable, highly correlated data with complex interactions and fault symptoms. This study proposes the robust fault detection method for the distillation column. Fault detection and diagnosis (FDD) for process monitoring and control has been an effective field of research for two decades. This area has been used widely in sophisticated engineering design applications to ensure the proper functionality and performance diagnosis of advanced and complex technologies. Robust fault detection of the realistic faults in distillation column in dynamic condition has been considered in this study. For early detection of faults, the model is based on nonlinear autoregressive with exogenous input (NARX) network. Tapped delays lines (TDLs) have been used for the input and output sequences. A case study was carried out with three different fault scenarios, i.e., valve sticking at reflux and reboiler, and tray upset. These faults would cause the product degradation. The normal data (no fault) is used for the training of neural network in all three cases. It is shown that the proposed algorithm can be used for the detection of both internal and external faults in the distillation column for dynamic system monitoring and to predict the probability of failure.
Knowledge of dynamic behavior of process plants is essential in designing, commissioning and oper... more Knowledge of dynamic behavior of process plants is essential in designing, commissioning and operating process plants. Several advanced model based technologies require explicit dynamic models of the process plant. However, even the development of the dynamic model of a simple unit operation is very daunting, costly and time-consuming. This paper presents a rigorous dynamic simulation using Aspen Plus. The data from the dynamic simulation is then used to develop transfer function model of the column. This case study considers a distillation column with an ethanol-water feed. A 2×3 distillation column transfer function model is developed from step responses of the Aspen plus dynamic simulation, i.e. with feed valve opening %, reflux flow rate and reboiler duty as input and top and the bottom compositions as an output. A separate SISO transfer function from condenser duty to column pressure is also developed.
The Artificial Neural Network (ANN) modelling is presented for the steam gasification of palm ker... more The Artificial Neural Network (ANN) modelling is presented for the steam gasification of palm kernel shell using CaO adsorbent and coal bottom ash as a catalyst. The effect of the parameters such as; temperature, CaO/biomass ratio and Coal bottom ash wt.% at fixed steam/biomass ratio and steam/ biomass ratio at the fixed temperature on product gas composition of H 2 , CO, CO 2 , and CH 4 are modelled using ANN. The effect of parameters is used as an input, while the gas compositions, syngas yield, LHV gas and HHV gas of gas as the output of the network. Back propagation algorithm has been used for the training with 7 neurons in the hidden layer. Hence, the selected ANN architecture was (2-7-1). The gas composition predicted by the ANN model are compared with experimental results obtained from pilot scale gasification system that has been reported in our previous study. The ANN predicted results show high agreement with the published experimental values with the coefficient of determination R 2 ¼ 0.998 for almost all the cases, i.e., the effect of parameters. RMSE, MAD, and AARE have been reported to be very insignificant for the predicted and experimental values.
Industrial & Engineering Chemistry Research, Oct 2018
Many control-relevant systems present the challenges of nonlinearity, directionality and ill-cond... more Many control-relevant systems present the challenges of nonlinearity, directionality and ill-conditioning for the control systems, and exhibit poor controller performance. This study proposes a Nonlinearity Index to quantify the extent of nonlinearity of such systems. A dynamic nonlinear model of a pilot-scale distillation column operating near the azeotropic region was simulated using Aspen Plus Dynamics. A comparison of the results is made with the Nonlinearity Measure proposed by Du et al. Results show that a significant increase is exhibited in the proposed Nonlinearity Index values of the system as the system moves toward the azeotropic region. Prediction errors of the linear models are also shown to be correlated to the proposed index. Therefore, the proposed Nonlinearity Index is consistent with the existing indicators of nonlinearity, and thus its measurements of system nonlinearity are reliable. Controller performance for the system at higher values of the proposed index further presents its efficacy.
High-purity distillation columns present challenges of nonlinearity, interactions, directionality... more High-purity distillation columns present challenges of nonlinearity, interactions, directionality, and ill-conditioning for the control systems and thus exhibit poor controller performance. Effects of these issues on the performance of model pre-dictive control in operations in near-azeotropic regions are investigated. A dynamic model of a pilot-scale distillation column is simulated and controlled by a linear model predictive controller. As the system moves towards the azeotropic region, the nonlinearity, directionality, and interactions of the system increase. A significant decline is observed in the system control performance since the controller incurs overshoots and offsets for negative and positive setpoint changes. The integral time absolute errors are substantially higher near the azeotropic region. Only a small operating range is able to ensure sufficient controllability near the azeotropic region.
Objectives: Closed-loop identification is reported to provide better results for identification o... more Objectives: Closed-loop identification is reported to provide better results for identification of systems for control applications. This study conducted closed-loop identification on an Aspen Plus® dynamic simulation based on a pilot-plant distillation column to develop discrete-time linear time-invariant models. Methods/Statistical Analysis: Identification data was generated using set-point perturbations in control variables under proportional-integral control. Identified models were compared with a model identified using open-loop data using 20-step ahead predictions. Findings: Results indicate that closed-loop identification provides more precise prediction models than open-loop identification in this case study. 20-step predictions for closed loop models exceeded 90% fit, whereas the open loop model predictions provided a 70% fit and missed the steady-state values. Application/Improvements: Thus closed-loop identification is more appropriate for applications in model-based controllers.
Process model is the kernel element of Model Predictive Control (MPC) system. It is always desira... more Process model is the kernel element of Model Predictive Control (MPC) system. It is always desirable to get a model as accurate as the actual facility or plant to reduce the built-in mismatch. With the passage of time, the mismatch between model and plant increases, which results in degradation of MPC performance. To rectify mismatches through plant re-identification is exorbitant and time consuming. Hence, mismatch detection is critical to isolate the faulty sub models to avoid complete re-identification. Badwe et al. proposed a method using partial correlation to isolate and detect plant-model mismatch which uses dynamic models in the decorrelation step. This study extends his work by comparing the performances of Autoregressive Exogenous Input (ARX) model and Auto-Regressive Moving Average with Exogenous Input (ARMAX) model for detection of model-plant mismatch. Wood and Berry binary distillation column is used as a case study to demonstrate the application of the ARX and ARMAX models in mismatch detection. Results show that ARMAX models provide higher accuracy with less model order as compared to ARX. This results in less computational complexity and less processing power required in the MPC, hence improving its efficiency.
Boilers are amongst the most important industrial utilities with their performance crucial in smo... more Boilers are amongst the most important industrial utilities with their performance crucial in smooth functioning of any industrial facility. Poor boiler performance and frequent breakdowns can signiiicantly reduce equipment efficiency translating to overall detrimental performance and prooit reduction. This paper focuses on achieving superior feed water treatment to maximize boiler efficiency. The work is to explore the adaptability of ion-exchange enabled water treatment in both small and large setups, particularly those working with limited throughputs, for which membrane technologies are not much useful both in terms of capital costs and unavailability of high pressures as needed with such systems. A mini water treating facility comprising of iltration and ion exchange units was designed and run with water to determine the extent of puriiication. Comparative analysis of three major properties: pH, hardness and TDS levels of raw and product water streams revealed a signiiicant reduction in values in the product demonstrating the practicality and effectiveness of the method.
A steady state Aspen Plus® simulation is used to propose and analyse a novel indirect gasificatio... more A steady state Aspen Plus® simulation is used to propose and analyse a novel indirect gasification model using steam as the gasifying medium. A sensitivity study is carried out by varying the steam-to-coal ratio. Part of the Syngas produced is combusted in the furnace to provide flue gases at high temperature which is then passed through the bayonets to provide reaction heat. RK-Soave method is used for the evaluation of physical properties of mixed conventional components and CISOLID components in the simulation. The feedstock used in the simulation is indigenous Thar coal from Pakistan, where large coal reserves are present. It was found that a lower steam-to-coal ratio increases the heat content of the syngas produced. Also, the carbon conversion undergoes a maximum at the ratio 2.0. As we increase the steam-to-coal ratio, the yields are better but inevitably reduce the quality of syngas produced. Carbon conversions and H2/CO ratios as high as 95.76% and 3.09 respectively were observed with higher steam to carbon ratios. However, lower ratios provide high yields (69.00%) and cold gas efficiencies (54.41%). Based on these results, it can be said that this parameter significantly influences the syngas quality and processing costs and the diversity of trends suggest a more detailed analysis for optimisation of the process.
In view of limited liquid fuels, in terms of crude oil reserves and to reduce the use of constant... more In view of limited liquid fuels, in terms of crude oil reserves and to reduce the use of constantly and rapidly diminishing natural gas reserves, researchers are attracted towards Fisher Tropsch reaction. Aspen Plus® has become reliable, acquainted and recognized processes modeling software, extensively in practice for coal and biomass gasification processes. It contains different physical property packages that are useful for solid handling. Aspen Plus® model has been proposed to develop a better understanding of the process for geometric analysis of gasifier. This simulation presents an alternate technology for conventional coal gasification to improve the performance of process by varying geometry of gasifier. The Purpose of this study, is entirely focus on the production of synthesis gas from coal, through a process of indirect gasification and using only steam as the gasifying medium. The book serves as reference material for students, engineers and scientists working in the area of syngas production and coal gasification. .
Uploads