Ashwin Kumar Rajagopalan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ashwin Kumar Rajagopalan has a total of 85 co-authors. Here are the top co-authors by number of shared publications:
- Marco Mazzotti
ETH Zurich
13 shared publications - Daniel Biri
Institute of Energy and Process Engineering, ETH Zurich, Zurich, Switzerland
7 shared publications Manfred Morari
Department of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia 19104, United States
6 shared publications- Stefan Bötschi
Institute of Process Engineering, ETH Zurich, 8092 Zurich, Switzerland
6 shared publications Marco Mazzotti
5 shared publications- Aurora J. Cruz Cabeza
Durham University
4 shared publications
Latest Publications
- Open Access
- 6 August 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Solution of Multidimensional Population Balance Equation Models via the Sectional Method of Moments
Population balance equations (PBEs) provide a rigorous framework for modelling particulate processes, but their numerical solution becomes increasingly demanding for multidimensional systems. This work introduces the Sectional Method of Moments (SecMoM). Through an extension of the widely used Method of Moments (MoM), the population distribution is divided into moving sections and local moments are calculated within each section, providing distribution-level information with moderate additional computational overhead. By tracking only the populated region of the internalcoordinate domain, SecMoM avoids requiring a fixed global grid. The formulation is extended to multiple internal coordinates, property-dependent transport, nucleation, and complete dissolution. SecMoM is demonstrated for one-and two-dimensional crystallisation PBEs involving size-dependent growth, nucleation, dissolution, and temperature cycling. Comparisons with finite volume method (FVM) solutions show excellent agreement in liquid-phase concentration, total particle number, average particle dimensions, and population-distribution evolution, with mean relative absolute deviations of 1.26×10 −3 or lower. The method captures distribution broadening, the formation and removal of fine particles, and anisotropic growth in two dimensions. For the 2D crystallisation case, the corresponding FVM data structure contained approximately 14 times more stored values than the SecMoM representation. For the simplified benchmark, both codes were implemented in JAX, enabling execution on CPU and GPU hardware. SecMoM was approximately 1.4 times faster on the CPU and 2.5 times faster on the GPU than the optimised FVM implementation. SecMoM therefore combines distribution-level resolution with efficient computation and compact data storage, making it well suited to multidimensional PBEs and repeated-simulation applications.
- 6 August 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessSolution of Multidimensional Population Balance Equation Models via the Sectional Method of Moments
Population balance equations (PBEs) provide a rigorous framework for modelling particulate processes, but their numerical solution becomes increasingly demanding for multidimensional systems. This work introduces the Sectional Method of Moments (SecMoM). Through an extension of the widely used Method of Moments (MoM), the population distribution is divided into moving sections and local moments are calculated within each section, providing distribution-level information with moderate additional computational overhead. By tracking only the populated region of the internalcoordinate domain, SecMoM avoids requiring a fixed global grid. The formulation is extended to multiple internal coordinates, property-dependent transport, nucleation, and complete dissolution. SecMoM is demonstrated for one-and two-dimensional crystallisation PBEs involving size-dependent growth, nucleation, dissolution, and temperature cycling. Comparisons with finite volume method (FVM) solutions show excellent agreement in liquid-phase concentration, total particle number, average particle dimensions, and population-distribution evolution, with mean relative absolute deviations of 1.26×10 −3 or lower. The method captures distribution broadening, the formation and removal of fine particles, and anisotropic growth in two dimensions. For the 2D crystallisation case, the corresponding FVM data structure contained approximately 14 times more stored values than the SecMoM representation. For the simplified benchmark, both codes were implemented in JAX, enabling execution on CPU and GPU hardware. SecMoM was approximately 1.4 times faster on the CPU and 2.5 times faster on the GPU than the optimised FVM implementation. SecMoM therefore combines distribution-level resolution with efficient computation and compact data storage, making it well suited to multidimensional PBEs and repeated-simulation applications.
- 31 March 2026
- Published by Springer Nature
- p. 191-238
Image Analysis Methodologies
This chapter compiles established practices for characterizing crystal populations to characterize the crystallization of biomolecules using image-based systems. It starts with an introduction to available measurement tools, comparing their advantages and disadvantages. This is followed by a detailed description of the different steps from image acquisition to data analysis and visualization, with considerations offered for each step. The chapter ends with a discussion of the challenges and limitations currently facing imaging tools. By providing a comprehensive overview of crystal characterization using imaging techniques, this chapter aims to equip practitioners with the knowledge necessary to effectively apply these methods in both research and industrial contexts.
- 31 March 2026
- Published by Springer Nature
- p. 191-238
Image Analysis Methodologies
This chapter compiles established practices for characterizing crystal populations to characterize the crystallization of biomolecules using image-based systems. It starts with an introduction to available measurement tools, comparing their advantages and disadvantages. This is followed by a detailed description of the different steps from image acquisition to data analysis and visualization, with considerations offered for each step. The chapter ends with a discussion of the challenges and limitations currently facing imaging tools. By providing a comprehensive overview of crystal characterization using imaging techniques, this chapter aims to equip practitioners with the knowledge necessary to effectively apply these methods in both research and industrial contexts.
- Open Access
- 16 December 2025
- Published by Elsevier in Chemical Engineering Science
- Vol. 323
Unveiling the interplay between crystallisation process models and data: Development of a diagnostic tool
- 16 December 2025
- Published by Elsevier in Chemical Engineering Science
- Vol. 323
Open AccessUnveiling the interplay between crystallisation process models and data: Development of a diagnostic tool
- Open Access
- 21 November 2025
- Published by Elsevier in Chemical Engineering Research and Design
- Vol. 224, 452-466
Toward predictive modeling of industrial crystallizers: A compartmental framework for coupling hydrodynamics and population dynamics
- 21 November 2025
- Published by Elsevier in Chemical Engineering Research and Design
- Vol. 224, 452-466
Open AccessToward predictive modeling of industrial crystallizers: A compartmental framework for coupling hydrodynamics and population dynamics
- Open Access
- 19 September 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Toward Predictive Modeling of Industrial Crystallizers: A Compartmental Framework for Coupling Hydrodynamics and Population Dynamics
Accurate prediction of particle size distribution (PSD) is critical for the optimal design of batch crystallization processes. The interplay between crystallization kinetics and local hydrodynamics complicates predictive process modeling. Computational fluid dynamics (CFD) provides hydrodynamic characterization of the system, which when integrated with population balance equation (PBE) models, leads to a comprehensive representation of the system. The goal of this work is two-fold. First, to develop a computationally efficient CFD–PBE framework using a compartmental modeling approach. Second, to conduct a systematic evaluation of the impact of hydrodynamics on crystallization processes. CFD simulations of a batch crystallizer provide mean flow field and its turbulence properties, which are mapped onto a coarse-grained structure of well-mixed compartments. PBEs are solved over the compartments, capturing mixing effects on the evolution of the PSD, at a fraction of computational cost. Crystallization case studies are conducted to examine mixing effects under varying kinetics, operating conditions, and impeller types. Results show that mixing-driven phenomena are system-specific, and hydrodynamic effects must be considered when transferring insights between systems and scales. The framework is system and geometry-agnostic, making it a valuable tool for robust process design and scale-up, offering a practical alternative to models that assume perfect mixing.
- 19 September 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessToward Predictive Modeling of Industrial Crystallizers: A Compartmental Framework for Coupling Hydrodynamics and Population Dynamics
Accurate prediction of particle size distribution (PSD) is critical for the optimal design of batch crystallization processes. The interplay between crystallization kinetics and local hydrodynamics complicates predictive process modeling. Computational fluid dynamics (CFD) provides hydrodynamic characterization of the system, which when integrated with population balance equation (PBE) models, leads to a comprehensive representation of the system. The goal of this work is two-fold. First, to develop a computationally efficient CFD–PBE framework using a compartmental modeling approach. Second, to conduct a systematic evaluation of the impact of hydrodynamics on crystallization processes. CFD simulations of a batch crystallizer provide mean flow field and its turbulence properties, which are mapped onto a coarse-grained structure of well-mixed compartments. PBEs are solved over the compartments, capturing mixing effects on the evolution of the PSD, at a fraction of computational cost. Crystallization case studies are conducted to examine mixing effects under varying kinetics, operating conditions, and impeller types. Results show that mixing-driven phenomena are system-specific, and hydrodynamic effects must be considered when transferring insights between systems and scales. The framework is system and geometry-agnostic, making it a valuable tool for robust process design and scale-up, offering a practical alternative to models that assume perfect mixing.
- Open Access
- 16 September 2025
- Published by Elsevier in Powder Technology
- Vol. 468
Impact of disperse populations of non-equant crystals on filter cake resistance: Insights from filtration experiments and simulations
- 16 September 2025
- Published by Elsevier in Powder Technology
- Vol. 468
Open AccessImpact of disperse populations of non-equant crystals on filter cake resistance: Insights from filtration experiments and simulations
- Open Access
- 7 August 2025
- Published by Springer Nature in Adsorption
- Vol. 31 (6) , 1-14
Understanding errors in gas adsorption at low pressures: the case of direct air capture
Accurate experimental adsorption equilibrium measurements are necessary for benchmarking adsorbents, validating molecular simulations and setting up process simulations. Although many sources of errors in these measurements have been reported in the literature, the purity of the gas used is generally not considered a major problem as long as research grade gases are used. In this work, we propose that significant deviations in the measured isotherms can potentially arise due to the accumulation of impurities in the measurement cell, especially in the low-pressure region, which is important for systems dealing with low partial pressure, such as $$\hbox {CO}_2$$ direct air capture (DAC). We conduct numerical studies to highlight this issue. The first part of our analysis uses the Langmuir isotherm equation to generate baseline isotherms representative of adsorbents with varying affinities for $$\hbox {CO}_2$$ , enabling a parametric assessment of impurity effects. This is followed by a material-specific study examining the influence of impurities on isotherms for several zeolites and metal-organic frameworks (MOFs).- 7 August 2025
- Published by Springer Nature in Adsorption
- Vol. 31 (6) , 1-14
Open AccessUnderstanding errors in gas adsorption at low pressures: the case of direct air capture
Accurate experimental adsorption equilibrium measurements are necessary for benchmarking adsorbents, validating molecular simulations and setting up process simulations. Although many sources of errors in these measurements have been reported in the literature, the purity of the gas used is generally not considered a major problem as long as research grade gases are used. In this work, we propose that significant deviations in the measured isotherms can potentially arise due to the accumulation of impurities in the measurement cell, especially in the low-pressure region, which is important for systems dealing with low partial pressure, such as $$\hbox {CO}_2$$ direct air capture (DAC). We conduct numerical studies to highlight this issue. The first part of our analysis uses the Langmuir isotherm equation to generate baseline isotherms representative of adsorbents with varying affinities for $$\hbox {CO}_2$$ , enabling a parametric assessment of impurity effects. This is followed by a material-specific study examining the influence of impurities on isotherms for several zeolites and metal-organic frameworks (MOFs). - Open Access
- 26 July 2025
- Published by Elsevier in Powder Technology
- Vol. 466
Is there a “right” particle size and shape measurement tool?: A comparative study of seven online and offline devices on nine different particle populations
- 26 July 2025
- Published by Elsevier in Powder Technology
- Vol. 466
Open AccessIs there a “right” particle size and shape measurement tool?: A comparative study of seven online and offline devices on nine different particle populations
- Open Access
- 17 July 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Unveiling the Interplay Between Crystallisation Process Models and Data: Development of a Diagnostic Tool
Crystallisation processes play a crucial role in the pharmaceutical and agrochemical industries, where particle size and shape distributions (PSSD) significantly affect product quality and performance. Accurate process models are essential for design and optimisation, yet they are often hindered by overparameterisation stemming from mismatches between model complexity and data quality or availability. This overfitting results in models that fail to capture the underlying physics, leading to poor predictive capabilities, limited extrapolation beyond controlled laboratory conditions, and inefficient use of experimental resources. This work discusses a rigorous, valuable methodology and introduces a tool to address practical modelling challenges, guiding practitioners in academia and industry toward more predictive models. We propose a comprehensive framework for identifiability analysis to diagnose and enhance model feasibility, demonstrated through a case study of cooling crystallisation involving needle-like particles. A digital twin was developed to simulate the crystallisation process and replicate measurement devices, followed by an extensive identifiability analysis at both theoretical and practical levels, yielding actionable insights for real-world applications. Our results are threefold. First, we emphasise the constraints that identifiability imposes on modelling when parameters are not uniquely determinable (i.e., ill-posed). Second, we discuss the effects of device selection, data resolution, sampling frequency, and experimental design on model identifiability. Finally, we propose strategies for aligning model complexity with available data to prevent overparameterisation from theoretical formulation through to practical implementation.
- 17 July 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessUnveiling the Interplay Between Crystallisation Process Models and Data: Development of a Diagnostic Tool
Crystallisation processes play a crucial role in the pharmaceutical and agrochemical industries, where particle size and shape distributions (PSSD) significantly affect product quality and performance. Accurate process models are essential for design and optimisation, yet they are often hindered by overparameterisation stemming from mismatches between model complexity and data quality or availability. This overfitting results in models that fail to capture the underlying physics, leading to poor predictive capabilities, limited extrapolation beyond controlled laboratory conditions, and inefficient use of experimental resources. This work discusses a rigorous, valuable methodology and introduces a tool to address practical modelling challenges, guiding practitioners in academia and industry toward more predictive models. We propose a comprehensive framework for identifiability analysis to diagnose and enhance model feasibility, demonstrated through a case study of cooling crystallisation involving needle-like particles. A digital twin was developed to simulate the crystallisation process and replicate measurement devices, followed by an extensive identifiability analysis at both theoretical and practical levels, yielding actionable insights for real-world applications. Our results are threefold. First, we emphasise the constraints that identifiability imposes on modelling when parameters are not uniquely determinable (i.e., ill-posed). Second, we discuss the effects of device selection, data resolution, sampling frequency, and experimental design on model identifiability. Finally, we propose strategies for aligning model complexity with available data to prevent overparameterisation from theoretical formulation through to practical implementation.
- Open Access
- 9 June 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Is There a "Right" Particle Size and Shape Measurement Tool?: A Comparative Study of Seven Online and Offline Devices on Nine Different Particle Populations
The accurate characterization of the particle size and shape distribution (PSSD) of ensembles of crystals is essential for effective design, modeling, and control of crystallization processes. Various techniques for measuring PSSDs have been introduced at both laboratory and pilot scales. Despite the wide range of techniques available for the measurement of PSSDs, there is a noted gap in the literature concerning a comparative analysis of these methods when characterizing one, two, and three-dimensional particle size (and shape) distributions (PS(S)D) of population of crystals. To this end, five commercial and two bespoke techniques were used to characterize nine particle ensembles, with some obtained with supplier-specified properties, and others specifically prepared for this study. The evaluated instruments include the Mettler Toledo's FBRM, BlazeMetrics' probe, Mettler Toledo's EasyViewer, Malvern's Laser Diffraction, and Malvern's Morphologi; in addition to a stereoscopic imaging device (DISCO) and a confocal microscopy device (Petroscope). Test samples include small, medium, and large spherical glass beads with known size ranges; small and large sodium chloride cubes obtained through sieving; two populations of needle-like D-mannitol crystals; plate-like adipic acid crystals; and agglomerated L-Glutamic acid crystals. The results show that online devices generally disagree with each other, with offline devices, and with independent size references such as sieve fractions. For equant particles, offline devices exhibit good agreement with each other and independent size references. However, discrepancies arise for non-equant crystals (i.e. needles and plates). Only the Morphologi, DISCO, and Petroscope offer 2D characterization (measuring length and width), while only the non-commercial bespoke techniques, i.e. DISCO and Petroscope, provide 3D characterization (measuring length, width, and thickness). Comparing the measured multidimensional PSSDs reveal small differences between the two bespoke devices. The outcomes from this study will inform practitioners on the strengths and limitations of PS(S)D characterization techniques and in turn their selection for specific applications.
- 9 June 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessIs There a "Right" Particle Size and Shape Measurement Tool?: A Comparative Study of Seven Online and Offline Devices on Nine Different Particle Populations
The accurate characterization of the particle size and shape distribution (PSSD) of ensembles of crystals is essential for effective design, modeling, and control of crystallization processes. Various techniques for measuring PSSDs have been introduced at both laboratory and pilot scales. Despite the wide range of techniques available for the measurement of PSSDs, there is a noted gap in the literature concerning a comparative analysis of these methods when characterizing one, two, and three-dimensional particle size (and shape) distributions (PS(S)D) of population of crystals. To this end, five commercial and two bespoke techniques were used to characterize nine particle ensembles, with some obtained with supplier-specified properties, and others specifically prepared for this study. The evaluated instruments include the Mettler Toledo's FBRM, BlazeMetrics' probe, Mettler Toledo's EasyViewer, Malvern's Laser Diffraction, and Malvern's Morphologi; in addition to a stereoscopic imaging device (DISCO) and a confocal microscopy device (Petroscope). Test samples include small, medium, and large spherical glass beads with known size ranges; small and large sodium chloride cubes obtained through sieving; two populations of needle-like D-mannitol crystals; plate-like adipic acid crystals; and agglomerated L-Glutamic acid crystals. The results show that online devices generally disagree with each other, with offline devices, and with independent size references such as sieve fractions. For equant particles, offline devices exhibit good agreement with each other and independent size references. However, discrepancies arise for non-equant crystals (i.e. needles and plates). Only the Morphologi, DISCO, and Petroscope offer 2D characterization (measuring length and width), while only the non-commercial bespoke techniques, i.e. DISCO and Petroscope, provide 3D characterization (measuring length, width, and thickness). Comparing the measured multidimensional PSSDs reveal small differences between the two bespoke devices. The outcomes from this study will inform practitioners on the strengths and limitations of PS(S)D characterization techniques and in turn their selection for specific applications.