Publications

Authors: Anwesha Chaudhury, Huiquan Wu, Mansoor Khan, Rohit Ramachandran

Paper Link: Link

Abstract: This work is concerned with the theoretical development of a semi-mechanistic aggregation kernel in a population balance model (PBM) which takes into account the effect of droplet spreading on a particle surface that aids in the coalescence of particles. Empirical aggregation kernels are more commonly used in simulations however they do not reflect the true physics of the system. The proposed kernel is computationally less expensive, yet it takes into account the various key operating parameters that affect the process. The kernel has been validated qualitatively and the lumped and distributed properties show good agreement with the expected behavior of the process. The various empirical parameters present in the granulation model have been identified and expressed as a function of the measurable operating quantities, thus providing a better knowledge regarding the effect of the process parameters on the final product with the help of a more fundamental, first principle based model. A detailed sensitivity analysis (involving viscosity, impeller speed, contact angle and liquid spray rate) has also been conducted in order to study the influence of the process parameters on the final granule properties. This knowledge provides a theoretical basis for the high-shear wet granulation process design space development. The model has also been able to successfully capture the steady and induction behavior of the process under the expected operating conditions.

Authors: Ravendra Singh, Marianthi Ierapetritou, Rohit Ramachandran

Paper Link: Link

Abstract: The direct compaction continuous pharmaceutical tablet manufacturing process considered in this study is highly interactive and has time delays for several process variables due to sensor placements, that indicate that a simple feedback regulatory control system (e.g. PI(D)) by itself may not be sufficient to achieve the tight process control as imposed by regulatory authorities. This process comprises of coupled dynamics involving slow and fast responses indicating the requirement of a hybrid control scheme such as an advanced hybrid MPC-PID control scheme. In this article, an efficient plant-wide hybrid MPC-PID control strategy for an integrated continuous pharmaceutical tablet manufacturing process via direct compaction has been designed in silico. The designed hybrid control system has been implemented in a first principle model-based flowsheet that was simulated in gPROMS (PSE). Results demonstrated enhanced performance of critical quality attributes (CQAs) under the hybrid control scheme compared to only PID or MPC control schemes thus illustrating the potential of hybrid control scheme in improving pharmaceutical manufacturing operations. A systematic methodology for design and implementation of hybrid MPC-PID control system has been also developed that can be employed for other processes.

Authors: Anwesha Chaudhury, Ivan Oseledets, Rohit Ramachandran

Paper Link: Link

Abstract: Multi-dimensional population balance equations (PBEs) are commonly used to describe the dynamics of particulate processes such as granulation. Such a class of equations are numerically complex and computationally intensive to solve due to the multiple internal coordinates involved. A computationally efficient model reduction technique would overcome the computational overheads associated with the solution of multi-dimensional PBEs. Moreover, this enables the process model to be used efficiently in process control and optimization. This study is concerned with the development of a novel reduced order model for a three-dimensional population balance model (PBM) for granulation, using a tensor decomposition technique in combination with separation of variables and singular value decomposition. These techniques were used to decompose the complex aggregation and breakage integrals. The developed model is faster by two orders of magnitude, requires less memory allocation for the storage of variables and results in negligible error when compared with the full model.

Authors: Ondřej Kašpar, Viola Tokárová, Sarang Oka, Koushik Sowrirajan, Rohit Ramachandran, František Štěpánek

Paper Link: Link

Abstract: The X-ray micro-tomography (micro-CT) technique has been used to visualize the microstructure of granules produced by high shear wet granulation and the dynamic evolution of porosity during granule dissolution. Using acetaminophen (paracetamol) as the active pharmaceutical ingredient (API) and microcrystalline cellulose (Avicel PH-200) as an excipient, the porosity of the granules was systematically influenced by the granulation process parameters (binder/solids ratio, impeller speed and wet massing time). An increase of granule porosity from 7% to 10% and 18% lead to a decrease of the dissolution time t90 from 435 min to 98 min and 37 min, respectively. The combination of time-resolved micro-CT imaging with UV/vis detection of the quantity dissolved made it possible to evaluate the effective diffusion coefficient of the API through the granule structure, and thus establish a quantitative structure–property relationship for dissolution. A power-law dependence of the effective diffusivity on porosity (Archie’s law) was found to hold.

Authors: Maitraye Sen, Amanda Rogers, Ravendra Singh, Anwesha Chaudhury, Joyce John, Marianthi G.Ierapetritou, Rohit Ramachandran

Paper Link: Link

Abstract: In this work, a continuous API purification process has been optimized using an integrated flowsheet model. The simulation is dynamic in nature and includes an API purification step (crystallization), followed by filtration, drying and mixing of the API with an excipient. For the first time, this study demonstrates the use of a reduced order model (ROM) within the mixing unit for prediction of particle velocities that is coupled with a population balance model (PBM) of the mixer to quantify macroscopic properties. The main objective is to optimize the integrated flowsheet model such that there is an overall improvement in process operation. The optimum cooling schedule during crystallization has been obtained. The optimum values of filter pressure gradient, drying gas temperature and mixer RPM (speed) have also been determined. It is seen that the optimized operating conditions give a narrower CSD of the API crystals and lower RSD (Relative standard deviation) of the final mixed product, compared to nonoptimal operating condition. The developed model can be used as an effective tool in control and optimization and can have future implementation in design of a Process Analytical Technology (PAT) system which can lead to improved operation of the manufacturing process.

Authors: Dana Barrasso, Sarang Oka, Ariel Muliadi, James D. Litster, Carl Wassgren & Rohit Ramachandran

Paper Link: Link

Abstract: Continuous tablet manufacturing has been investigated for its potential advantages (e.g., cost, efficiency, and controllability) over more conventional batch processes. One avenue for tablet manufacturing involves roller compaction followed by milling to form compactible granules. A better understanding of these powder processes is needed to implement Quality by Design in pharmaceutical manufacturing. In this study, ribbons of microcrystalline cellulose were produced by roller compaction and milled in a conical screen mill. A full factorial experiment was performed to evaluate the effects of ribbon density, screen size, and impeller speed on the product size distribution and steady-state mass holdup of the mill. A population balance model was developed to simulate the milling process, and a parameter estimation technique was used to calibrate the model with a subset of experimental data. The calibrated model was then simulated at other processing conditions and compared with additional unused experimental data. Statistical analyses of the results showed good agreement, demonstrating the model’s predictive capability in quantifying milled product critical quality attributes within the experimental design space. This approach can be used to optimize the design space of the process, enabling Quality by Design.

Authors: Dana Barrasso, Samjit Walia, Rohit Ramachandran

Paper Link: Link

Abstract: In the recent few years, continuous processing has been considered as a promising process alternative to batch processing in pharmaceutical manufacturing. Via a novel population balance model framework, a multi-dimensional multi-component model for a continuous granulation process was developed, describing time evolutions of distributions with respect to granule size, liquid distribution and granule composition. A parametric study was performed to analyze the effects of various process and design parameters, including granulator size and configuration, liquid spray rate and particle velocity, on evolutions and distributions of key granule properties. Simulation results capture experimentally observed sensitivities and trends thus demonstrating the use of a model-based framework for granulation process design, control and optimization to enable QbD in drug product development.

Authors: Ravendra Singh, Marianthi Ierapetritou, Rohit Ramachandran

Paper Link: Link

Abstract: The next generation of QbD based pharmaceutical products will be manufactured through continuous processing. This will allow the integration of online/inline monitoring tools, coupled with an efficient advanced model-based feedback control systems, to achieve precise control of process variables, so that the predefined product quality can be achieved consistently. The direct compaction process considered in this study is highly interactive and involves time delays for a number of process variables due to sensor placements, process equipment dimensions, and the flow characteristics of the solid material. A simple feedback regulatory control system (e.g., PI(D)) by itself may not be sufficient to achieve the tight process control that is mandated by regulatory authorities. The process presented herein comprises of coupled dynamics involving slow and fast responses, indicating the requirement of a hybrid control scheme such as a combined MPC–PID control scheme. In this manuscript, an efficient system-wide hybrid control strategy for an integrated continuous pharmaceutical tablet manufacturing process via direct compaction has been designed. The designed control system is a hybrid scheme of MPC–PID control. An effective controller parameter tuning strategy involving an ITAE method coupled with an optimization strategy has been used for tuning of both MPC and PID parameters. The designed hybrid control system has been implemented in a first-principles model-based flowsheet that was simulated in gPROMS (Process System Enterprise). Results demonstrate enhanced performance of critical quality attributes (CQAs) under the hybrid control scheme compared to only PID or MPC control schemes, illustrating the potential of a hybrid control scheme in improving pharmaceutical manufacturing operations.

Authors: Anuj V. Prakash, Anwesha Chaudhury, and Rohit Ramachandran

Paper Link: Link

Abstract: Computer-aided modeling and simulation are a crucial step in developing, integrating, and optimizing unit operations and subsequently the entire processes in the chemical/pharmaceutical industry. This study details two methods of reducing the computational time to solve complex process models, namely, the population balance model which given the source terms can be very computationally intensive. Population balance models are also widely used to describe the time evolutions and distributions of many particulate processes, and its efficient and quick simulation would be very beneficial. The first method illustrates utilization of MATLAB’s Parallel Computing Toolbox (PCT) and the second method makes use of another toolbox, JACKET, to speed up computations on the CPU and GPU, respectively. Results indicate significant reduction in computational time for the same accuracy using multicore CPUs. Many-core platforms such as GPUs are also promising towards computational time reduction for larger problems despite the limitations of lower clock speed and device memory. This lends credence to the use of highfidelity models (in place of reduced order models) for control and optimization of particulate processes.

Authors: Anuj V.Prakash, Anwesha Chaudhury, Dana Barrasso, Rohit Ramachandran

Paper Link: Link

Abstract: Population Balance Models (PBMs), a class of integro partial differential equations, are utilized for simulating dynamics of numerous particulate systems. PBMs describe the time evolutions and distributions of many particulate processes and their efficient and quick simulation are critical for enhanced process control and optimization, especially for real-time applications. However, their intensive computational resource requirement is largely a prohibitive factor in the utility of PBMs for control and optimization. This paper describes how distributed computing systems may be leveraged to execute PBM-based simulations thus achieving time savings, using MATLAB’s Distributed Computing Toolbox. A parallel computing algorithm was developed for a three dimensional and four dimensional population balance model with built-in constructs such as SPMD that ran efficiently on a cluster of two quad-core machines linked via a gigabit ethernet cable. Speedup of 6.2 and 5.7 times were achieved with 8 workers, over an un-parallelized code, for a 3 and 4 dimensional PBM respectively. Evaluations on work efficiency and scalability further affirm the algorithms’ respectable performance on larger clusters despite significant memory transfer overheads.

Authors: Maitraye Sen, Anwesha Chaudhury, Ravendra Singh, Joyce John, Rohit Ramachandran

Paper Link: Link

Abstract: Properties of active pharmaceutical ingredients influence the critical quality attributes (CQAs) of final solid dosage forms (e.g. tablets). In the last decade, continuous manufacturing has been shown to be a promising alternative to batch processing in the pharmaceutical industry. Therefore, a quantitative model-based analysis of the influence of upstream API properties on downstream processing quality metrics will lead to enhanced QbD in pharmaceutical drug product manufacturing (Benyahia et al., 2012). In this study, a dynamic flowsheet simulation of an integrated API purification step (crystallization), followed by filtration and drying, with a downstream process (powder mixing) is presented. Results show that the temperature profile of a cooling crystallization process influences the crystal size distribution which in turn impacts the RSD and API concentration of the powder mixing process, which in turn has a direct effect on tablet properties (Boukouvala et al., 2012). A hybrid PBM–DEM model is also presented to demonstrate the coupling of particle-scale information with process-scale information leading to enhanced elucidation of the dynamics of the overall flowsheet simulation.

Authors: Anwesha Chaudhury, Avi Kapadia, Anuj V.Prakash, Dana Barrasso, Rohit Ramachandran

Paper Link: Link

Abstract: This paper focuses on obtaining the numerical solution to a three-dimensional population balance model (PBM) of granulation using the cell-average technique first proposed by [22]. Conventionally, linear grids are used for the solution of PBMs, but the ability to incorporate non-linear grids would be more advantageous given that a larger size range can be covered using fewer number of grids, thus reducing computational overhead. Furthermore, the use of linear representation of grids in PBMs to represent industrial granulation processes that span a wide granule size range is computationally prohibitive and results show that a non-linear grid representation is computationally more efficient with comparable accuracy. Parallelization of the PBM via a multi-core strategy has also been incorporated in order to reduce the simulation time of the model. Incorporating the cell average technique along with parallelization of the overall model lends credence to the overall use of the model for effective granulation process design and analysis.

Authors: Anwesha Chaudhury, Rohit Ramachandran

Paper Link: Link

Abstract: This study is concerned with the development of an integrated three-dimensional population balance model (PBM) that describes the combined effect of key granulation mechanisms that occur during the course of a granulation process. Results demonstrate the importance of simulating the different mechanisms within a population balance model framework to elucidate realistic granulation dynamics. The incorporation of liquid addition in the model also aids in demarcating the dynamics in different regimes such as premixing, granulation (during liquid addition) and wet massing (after liquid addition). For the first time, the effect of primary particle size distributions and mode of binder addition on key granule properties was studied using an integrated PBM. Experimental data confirms the validity of the overall model as compared to traditional models in the literature that do not integrate the different granulation mechanisms.

Authors: Maitraye Sen, Atul Dubey, Ravendra Singh, and Rohit Ramachandran

Paper Link: Link

Abstract: This paper describes the development of a multidimensional population balance model (PBM) which can account for the dynamics of a continuous powder mixing/blending process. The PBM can incorporate the important design and process conditions and determine their effects on the various critical quality attributes (CQAs) accordingly. The important parameters considered in this study are blender dimensions and presence of noise in the inlet streams. The blender dynamics have been captured in terms of composition of the ingredients, (relative standard deviation) RSD, and (residence time distribution) RTD. PBM interacts with discrete element modeling (DEM) via one-way coupling which forms a basic framework for hybrid modeling. The results thus obtained have been compared against a full DEM simulation which is a more fundamental particle-level model that elucidates the dynamics of the mixing process. Results show good qualitative agreement which lends credence to the use of coupled PBM as an effective tool in control and optimization of mixing process due to its relatively fewer computational requirements compared to DEM.

Authors: Fani Boukouvala, Anwesha Chaudhury, Maitraye Sen, Ruijie Zhou, Lukasz Mioduszewski, Marianthi G. Ierapetritou & Rohit Ramachandran

Paper Link: Link

Abstract: In this work, a dynamic flowsheet model for the production of pharmaceutical tablets through a continuous wet granulation process is developed. The unit operation models which are integrated to compose the process line form a hybrid configuration which is comprised of a combination of mechanistic models, population balance models, and empirical correlations, based on the currently available process knowledge for each individual component. The main objective of this study is to provide guidance in terms of the necessary steps which are required in order to move from the unit operation level to the simulation of an integrated continuous plant operation. Through this approach, not only significant process conditions for each individual process are identified but also crucial interconnecting parameters which affect critical material properties of the processed powder stream are distinguished. Through the integration of the dynamic flowsheet with a final component of tablet dissolution, the connection of the processing history of a set of powders which undergo wet granulation and are contained in each produced tablet to the release rate of the pharmaceutical ingredient is enabled. The developed flowsheet is used for the simulation of different operating scenarios and disturbances which are often encountered during operation for the assessment of their effects towards critical material attributes, product properties, and the operation of further downstream processes. Simulation results demonstrate that granulation and milling which control the particle size distribution of the processed powder mixture highly affect the hardness and dissolution of the produced tablets.

Authors: Anwesha Chaudhury, Alexander Niziolek, Rohit Ramachandran

Paper Link: Link

Abstract: This paper is concerned with the integration of a heat and mass model with a population balance model of a fluid bed granulation process. The heat/mass model helps to track the liquid and porosity changes that occur within the granulation process due to simultaneous drying in the granulator. A mechanistic kernel proposed by [22] was used to model the aggregation process, which is a function of mechanistic properties such as liquid thickness and particle size. The model takes into account various mechanisms that significantly affect the granulation outcome, such as aggregation, drying/rewetting, and consolidation. Since these rate processes evolve dynamically, the aggregation kernel is updated periodically over time and for the first time, dynamically coupled with the heat/mass model to provide a more accurate qualitative description of granulation dynamics in a fluid-bed granulator.

Authors: Maitraye Sen, Rohit Ramachandran

Paper Link: Link

Abstract: This study is concerned with the development of a novel population balance model (PBM) framework that can qualitatively capture the dynamics of a continuous powder mixing process. For the first time, a PBM has been developed to model powder mixing and it accounts for key design and process parameters such as mixer RPM, processing angle in terms of powder fluxes, along with the effect of number of axial and radial compartments. Via this approach, results clearly show the qualitative validity of the PBM as a tool to capture the dynamics of the process that affect API composition, RSD and RTD. The model also demonstrates the use of the PBM as an overall multi-scale modeling tool to combine micro-level models such as DEM in a hybrid framework. Due to the relative computational simplicity of solving the PBM (as compared to DEM), the developed model can be used effectively in control and optimization of the mixing process.