Fazli Abdul Samad | Universiti Malaysia Pahang (UMP) (original) (raw)
Uploads
Papers by Fazli Abdul Samad
This paper discusses the development of constrained optimization strategies for a fedbatch penici... more This paper discusses the development of constrained optimization strategies for a fedbatch penicillin fermentation process. To facilitate the study, a mathematical model of the system is developed based on published materials, and simulated using MATLAB software. Good agreement is obtained when the results are validated against published work. To provide on-line estimates of the difficult to measure penicillin concentration, Partial Least Squares model is employed. Using these estimates, good control of product concentration is established thus enabling it to be implemented as part of product concentration control loop. Further improvements are introduced using dynamic optimization aiming at increasing the achievable product concentration while satisfying all process constraints. Two strategies are considered. These are optimal control policy using direct-shooting algorithm and unconstrained Dynamic Matrix Control (DMC). From this two optimization approaches, it is possible to estimate the optimal operating conditions as substrate feed rate so that the systems presents high performance within threshold value limit. The result also revealed that the use of DMC approach is superior the direct shooting method in term of the penicillin concentration as well as penicillin purity. Results obtained in this study have exposed the potentials of dynamic optimization schemes in improving the product purity in a fed-batch fermentation process.
Computers & Chemical Engineering, 2011
A generic multi-dimensional modelling framework for crystallization processes has been developed ... more A generic multi-dimensional modelling framework for crystallization processes has been developed to study various aspects of batch cooling crystallization operations and modelling options. The framework contains a generic crystallizer model from which a wide range of problem-system specific models can be created through a model generation procedure. The modelling framework allows one to study a wide range of chemical (crystallization)
Internal Model Control (IMC) yields very good performance for set point tracking, but gives slugg... more Internal Model Control (IMC) yields very good performance for set point tracking, but gives sluggish response for disturbance rejection problem. Feedback/feedforward controller has been widely used to overcome the disturbance rejection. However, feedback/feedforward IMC is suffering with the difficulty of measuring disturbance and/or making disturbance model in real plant due to the possibility of more than one disturbance. The present study has developed a controller for disturbance rejection based on feedback/feedforward IMC structure. The controller could give a good performance without model disturbance. The results show that the proposed method is better to the existing two-degree-of-freedom IMC (2DOF IMC) controller in the presence of model mismatch.
A generic model-based framework has been developed for crystallization processes, with applicatio... more A generic model-based framework has been developed for crystallization processes, with applications aiming at the control of process operations and the monitoring of product quality. This generic model-based framework allows the systematic development of a wide range of crystallization models for different operational scenarios. This enables the design and control engineers to analyze various crystallization operations and conditions, thus facilitating the development of process control and monitoring systems (PAT systems) for crystallization processes. The generic framework has been implemented in the ICAS-PAT software which allows the user to design and validate PAT systems through a systematic computer-aided framework. The application of the framework is highlighted for batch cooling crystallization of paracetamol where the framework was applied for design of a process monitoring and control system to obtain a desired crystal size distribution (CSD).
A systematic procedure to achieve a target crystal size distribution (CSD) under generated operat... more A systematic procedure to achieve a target crystal size distribution (CSD) under generated operational policies in batch cooling crystallization is presented. An analytical CSD estimator has been employed in the systematic procedure to generate the necessary operational policies to achieve the target CSD. Furthermore, this systematic procedure has been integrated with a generic multi-dimensional model-based framework. The generic nature of the model-based framework allows the study of a wide range of chemical systems under different operational scenarios, enabling thereby, the analysis of various crystallization operations and conditions. Therefore this generic multidimensional model-based framework can be used to generate “specific” models for crystallization processes and further verify the operational policies generated by the analytical CSD estimator for achieving the targeted CSD consistently. The application of the systematic procedure is illustrated for a potassium dichromate case study.
A generic multi-dimensional modeling framework for studying batch cooling crystallization process... more A generic multi-dimensional modeling framework for studying batch cooling crystallization processes under generated operational policies is presented. The generic nature of the modeling allows the study of a wide range of chemical systems under different operational scenarios, enabling thereby, the analysis of various crystallization operations and conditions. Furthermore, a systematic procedure for generating operational policies through available analytical crystal size
The control system of AFPT (air flow pressure and temperature) pilot plant has been used successf... more The control system of AFPT (air flow pressure and temperature) pilot plant has been used successfully to demonstrate the design of some controllers such as PID, cascade PID and fuzzy logic controller. It is important to develop a gas density control system in the AFPT pilot plant since gas density is one of the gas properties which need to be set precisely in few processes. The control system is designed based on internal model control (IMC) structure. Internal model control (IMC) yields very good performance for set point tracking, but gives sluggish response for disturbance rejection problem. Feedback/feedforward controller and cascade control configuration have been widely used to overcome the disturbance rejection. However, feedback/feedforward IMC is suffering with the difficulty of measuring disturbance and/or making disturbance model in real plant due to the possibility of more than one disturbances. The present study has developed a controller for disturbance rejection based on feedback/feedforward IMC structure and combined with cascade control. The proposed controller is also implemented to a long time delay process i.e. Semino and Brambilla proces model. The controller structure could give a good performance compared to standard 1DOF IMC.
This paper discusses the development of constrained optimization strategies for a fedbatch penici... more This paper discusses the development of constrained optimization strategies for a fedbatch penicillin fermentation process. To facilitate the study, a mathematical model of the system is developed based on published materials, and simulated using MATLAB software. Good agreement is obtained when the results are validated against published work. To provide on-line estimates of the difficult to measure penicillin concentration, Partial Least Squares model is employed. Using these estimates, good control of product concentration is established thus enabling it to be implemented as part of product concentration control loop. Further improvements are introduced using dynamic optimization aiming at increasing the achievable product concentration while satisfying all process constraints. Two strategies are considered. These are optimal control policy using direct-shooting algorithm and unconstrained Dynamic Matrix Control (DMC). From this two optimization approaches, it is possible to estimate the optimal operating conditions as substrate feed rate so that the systems presents high performance within threshold value limit. The result also revealed that the use of DMC approach is superior the direct shooting method in term of the penicillin concentration as well as penicillin purity. Results obtained in this study have exposed the potentials of dynamic optimization schemes in improving the product purity in a fed-batch fermentation process.
Computers & Chemical Engineering, 2011
A generic multi-dimensional modelling framework for crystallization processes has been developed ... more A generic multi-dimensional modelling framework for crystallization processes has been developed to study various aspects of batch cooling crystallization operations and modelling options. The framework contains a generic crystallizer model from which a wide range of problem-system specific models can be created through a model generation procedure. The modelling framework allows one to study a wide range of chemical (crystallization)
Internal Model Control (IMC) yields very good performance for set point tracking, but gives slugg... more Internal Model Control (IMC) yields very good performance for set point tracking, but gives sluggish response for disturbance rejection problem. Feedback/feedforward controller has been widely used to overcome the disturbance rejection. However, feedback/feedforward IMC is suffering with the difficulty of measuring disturbance and/or making disturbance model in real plant due to the possibility of more than one disturbance. The present study has developed a controller for disturbance rejection based on feedback/feedforward IMC structure. The controller could give a good performance without model disturbance. The results show that the proposed method is better to the existing two-degree-of-freedom IMC (2DOF IMC) controller in the presence of model mismatch.
A generic model-based framework has been developed for crystallization processes, with applicatio... more A generic model-based framework has been developed for crystallization processes, with applications aiming at the control of process operations and the monitoring of product quality. This generic model-based framework allows the systematic development of a wide range of crystallization models for different operational scenarios. This enables the design and control engineers to analyze various crystallization operations and conditions, thus facilitating the development of process control and monitoring systems (PAT systems) for crystallization processes. The generic framework has been implemented in the ICAS-PAT software which allows the user to design and validate PAT systems through a systematic computer-aided framework. The application of the framework is highlighted for batch cooling crystallization of paracetamol where the framework was applied for design of a process monitoring and control system to obtain a desired crystal size distribution (CSD).
A systematic procedure to achieve a target crystal size distribution (CSD) under generated operat... more A systematic procedure to achieve a target crystal size distribution (CSD) under generated operational policies in batch cooling crystallization is presented. An analytical CSD estimator has been employed in the systematic procedure to generate the necessary operational policies to achieve the target CSD. Furthermore, this systematic procedure has been integrated with a generic multi-dimensional model-based framework. The generic nature of the model-based framework allows the study of a wide range of chemical systems under different operational scenarios, enabling thereby, the analysis of various crystallization operations and conditions. Therefore this generic multidimensional model-based framework can be used to generate “specific” models for crystallization processes and further verify the operational policies generated by the analytical CSD estimator for achieving the targeted CSD consistently. The application of the systematic procedure is illustrated for a potassium dichromate case study.
A generic multi-dimensional modeling framework for studying batch cooling crystallization process... more A generic multi-dimensional modeling framework for studying batch cooling crystallization processes under generated operational policies is presented. The generic nature of the modeling allows the study of a wide range of chemical systems under different operational scenarios, enabling thereby, the analysis of various crystallization operations and conditions. Furthermore, a systematic procedure for generating operational policies through available analytical crystal size
The control system of AFPT (air flow pressure and temperature) pilot plant has been used successf... more The control system of AFPT (air flow pressure and temperature) pilot plant has been used successfully to demonstrate the design of some controllers such as PID, cascade PID and fuzzy logic controller. It is important to develop a gas density control system in the AFPT pilot plant since gas density is one of the gas properties which need to be set precisely in few processes. The control system is designed based on internal model control (IMC) structure. Internal model control (IMC) yields very good performance for set point tracking, but gives sluggish response for disturbance rejection problem. Feedback/feedforward controller and cascade control configuration have been widely used to overcome the disturbance rejection. However, feedback/feedforward IMC is suffering with the difficulty of measuring disturbance and/or making disturbance model in real plant due to the possibility of more than one disturbances. The present study has developed a controller for disturbance rejection based on feedback/feedforward IMC structure and combined with cascade control. The proposed controller is also implemented to a long time delay process i.e. Semino and Brambilla proces model. The controller structure could give a good performance compared to standard 1DOF IMC.