Dynamic Flowsheet Simulation of Solids Processes

Dynamic particulate process models are to be derived within the framework of dynamic flowsheet modeling and simulation (FSS). In the contribution FSS is developed for the separation processes of solid particles from a fluid resp. gas. Potential applications of the dynamic simulation environment of particle separation are for instance in the analysis and design of the dynamic process behavior in the purification of exhaust gases (for example to reduce emissions from combustion processes) but also in the production of particle layers with defined properties, as occur in the powder coating of surfaces.

The novel open-source flowsheet simulation software DYSSOL was used to simulate effects inside a system of interconnected fluidized bed reactors. The development of the needed models was done exemplary for the Chemical Looping Combustion process for solid and gaseous fuels. In CLC a solid oxygen carrier material is circulated between interconnected fluidized bed reactors. In the simulation, the focus is laid on the prediction of the dynamics of the whole system, especially the process start-up, shut-down and fuel load change. A dynamic model, which can be applied for bubbling beds and circulating fluidized beds was derived. Additionally, a cyclone was introduced for gas-solid separation. Loop seals ensure gas sealing between the reactors and were included into the modeling. Fluid mechanics inside the systems are modeled with empirical and semi-empirical, one-dimensional correlations, to enable fast calculations. These considerations allow real-time simulations of long-term effects in the system. The chemical reactions for gaseous and solid fuel combustions are included in the simulation. This has an effect on the solid oxygen carrier and so the oxidation and reduction of the carrier are regarded. The simulations were validated with experiments on a 25 kWth Chemical Looping Combustion facility at TUHH. The flowsheet models are able to predict the movement of the bed material between the units after operation changes as well as the time frames in which these changes occur. Besides, the gas and solids conversions in the fluidized bed reactors were simulated accurately.

This chapter presents new findings on process and product design in continuous spray layering granulation in horizontal fluidized beds. The results are achieved by a multi-scale approach, combining single-particle characterization of the influence of drying conditions on layer formation and properties with meso-scale information of particle flow and recirculation between process chambers separated by weirs. On the macro-scale, population balance modeling is used to describe the overall process dynamics taking into account apparatus design and process conditions. New process regime maps are presented along with control concepts to guarantee stable and safe operation as well as desired particle properties.

For a broad range of applications sieving/screening is well suited to separate bulk materials according to particle sizes. In the treated bulk materials particles frequently prevail in broad size distributions, with non-spherical shape and sometimes even under moist conditions, complicating the separation process. Therefore, it is inevitable to gain a deeper understanding of the subprocesses of screening (size based stratification, particle passage through the screen surface and possible transport along the screen) under the aforementioned conditions. To gain this knowledge, detailed particle-based simulation approaches like the discrete element method (DEM) are available. Based on the latter method, discontinuous and continuous screening as well as its subprocesses are investigated. Therein, different screen geometries and characteristics are considered along with various mechanical excitations applying model and real particle shapes first under dry conditions and later under the influence of various liquid amounts. In order to perform reliable DEM screening simulations, the exact determination of particle properties like size, shape, material and contact parameters is essential, which is required in advance of the simulations. Besides the DEM, the integral outcome of screening can be represented by various phenomenological process models. Usually, the material-, operating-, and apparatus-specific parameters of the latter process models are empirically determined by experiments, whereas, here, the parameters for screening process models are directly obtained from DEM simulations, which allows their benchmarking under defined conditions. Additionally, suitable process models are successfully extended to represent screening processes under the presence of moisture.

Fine grinding and dispersing, such as grinding in stirred media mills, gains importance in several industrial processes. Solid materials processing is frequently subjected to dynamic changes, effecting the performance of milling. To accurately model milling processes, dynamic flowsheet simulation turns out as a promising approach to gain quick and reliable solutions, describing the milling process over time. The connection of different process units is even closer to the industrial setup. Therefore, the focus of the study is the introduction of a dynamic model for stirred media mills that can be implemented into flowsheet simulation. The modelling approach aims at separating grinding and transport phenomena in the mill. Starting with an investigation of a batch grinding process in a “calibration mill”, the dependency of the breakage rate on machine and material parameters is shown. The stressing conditions in this calibration mill are determined theoretically and via simulations using coupled CFD-DEM simulations. In the study, the prediction of influences such as varying grinding media, stirrer speed and solids concentration on the breakage rate worked out well. In continuous processes, the particle transport and axial grinding media distribution, effecting the dynamics, are simulated as a series of instantly mixed cells, connected by mixing streams. With the dynamic flowsheet simulator Dyssol, the dynamic response of the product to changes in the feed was compared to experimental investigations with limestone in a laboratory stirred media mill. Material parameters for the model were tested in a newly designed breakage tester.

Solid bowl centrifuges are used in a wide range of applications in the process industry. The aim is to separate the individual phases of a liquid/liquid, liquid/solid or liquid/liquid/solid system. The design of solid bowl centrifuges is based on the Σ-theory, which does not describe the separation process with a sufficiently high accuracy. This process results in numbers of experiments with high time and cost expenditure. In addition, Σ-theory only describes the stationary state and therefore do not allow the calculation of start-up processes and load changes. This chapter shows a new real-time capable numerical algorithm, which ensures a high computational efficiency and is therefore suitable for dynamic simulations of the process behavior of solid bowl centrifuges. The introduction deals with the state of the art and the existing problems concerning of the design of solid bowl centrifuges. Subsequently, material functions representing the separation properties in solid bowl centrifuges are expounded. The developed material functions are the basis for the dynamic simulation of the process behavior in solid bowl centrifuges described below. The residence time and flow conditions of the apparatus significantly influence the process behavior for semi-batch and continuous processes. The last two sections present the dynamic modeling of continuously operating decanter and semi-batch tubular centrifuges. Example simulations and comparisons to experiments validate the developed dynamic models and demonstrate the applicability for dynamic simulations.

This work presents the fundamentals and exemplary applications of a generalized model for precipitation, aggregation and ripening processes including the formation of solid phases with two dimensions. The particle formation is governed by a widely applicable population balance approach. Solid formation processes are described via the numerically efficient Direct Quadrature Method of Moments (DQMOM), which can calculate the evolution of multiple solid phases simultaneously. The particle size distribution (PSD) is approximated by a summation of delta functions while the moment source term is approximated by a two-point quadrature. The moments to calculate the multivariate distributions are chosen carefully to represent the second order moments. Solid formation is based on the model of Haderlein et al. (2017) and is extended by a multidimensional aggregation model. Now, the influences of mixing, complex hydrochemistry and particle formation dynamics including nucleation, growth and aggregation on multiphase precipitation processes are modelled and simulated along independent dimensions with high efficiency.

Modelling the comminution in jet mills with respect to the complex two-phase flow and the dynamic process behaviour is still a challenging task. The processed solids pass through several stages in the mill: The comminution process in the lower part, the pneumatic transport towards the classifier in the middle section, and the classification step at the top. In this contribution, the grinding kinetics and process behaviour during quasi-batch and fed-batch operational mode for different holdups, classifier speeds, and particle sizes are examined in detail. A previously developed method using well-characterized aluminium particle probes to access the stressing conditions is adapted for application in the investigated jet mill: The relative particle impact velocity is linked to the geometric changes of the particles upon impact. A high number of impact events happen in the mill, while at the same time, the average particle velocity is comparatively low. Besides the stressing conditions, breakage probabilities for the used glass beads are determined by single particle impact experiments and described by the model of Vogel and Peukert. Solids concentration measurements and high-speed imaging reveal the formation of particle clusters at the classifier and its periphery. These clusters have a massive impact on the classification step itself: Fine particles are trapped inside the clusters and are not discharged. Based on an adaption of the breakage model, and using the mean relative particle impact velocity determined by the particle probes, a model for the product mass flow is introduced.

In spite of the broad range of applications of flow and sieve classification, the physical phenomena for higher particle loadings are not completely understood. As a starting point, common models such as the one of Molerus may be used and optimized to include particle-particle and particle-wall collisions. In this contribution, it is investigated to which extent single particle models may be employed to describe the performance of a deflector wheel classifier and a circular vibratory screening machine at higher loadings. For the sieving process, the Molerus model was modified with a selectivity parameter, while for the deflector wheel, a differentiation of particles with low and high Stokes numbers was made. For high Stokes numbers, in a first approximation, the particularities of the airflow can be neglected, but the impaction behavior on the wheel blades needs to be taken into account. With the detailed knowledge of the mean airflow, a much better prediction of the separation curve can be obtained. In particular, the dynamic aspects of flow and sieving classification have been studied.

In most industrial solid processing operations, the classification of particles is important and designed based on the terminal settling velocity as the main control parameter. This settling velocity is dependent on characteristic particle properties like size, density, and shape. Turbulent particle diffusion is the other key property controlling the efficiency of the separation. In this project, multi-stage separation experiments of a variety of materials have been performed using different flow velocities, mass loadings of the air, number of stages. Separation has been investigated separately concerning particle size, particle density, and particle shape. Continuous operation in terms of solid material and airflow has been mostly considered. However, variations in mass loading and pulsating operation of the fan have been investigated as well. The performance has been analyzed and discussed with respect to the separation functions, for instance regarding separation sharpness. Several modelling approaches have been checked and/or developed to describe theoretically the corresponding observations. After fitting the free model parameters, a very good agreement has been obtained compared to experimental measurements. Finally, the reduced model has been implemented into the central software DYSSOL.

The dustiness of a disperse solid can be understood as a property, which when handled in a gaseous environment, behaves similar to an aerosol, releasing the respective particle fraction of given quantity and size distribution. In general, this release of dust is undesirable because it might result in material loss and often is associated with an exposure of personnel involved or represents a risk of environmental pollution. The dustiness is therefore a product property, which might change along the process path, for example through comminution, agglomeration, classification or mixing of solids involved. Property functions which describe time variable dustiness integrated in dynamic processes as a function of the distribution of particle size, particle shape, and particle interaction during a certain handling, were determined as part of this project. For this purpose, experiments with laboratory equipment such as “free fall in still air”, “moving in a rotating drum”, “dispersion, pressure surges method”, or “airflow dispersion” were performed at very well defined boundary conditions and physically based models were established. The prediction functions were successfully implemented in the flow sheet simulation DYSSOL. These models will be further used through the introduction of the so called “Fractionated grade of release”. Together with the description of time-dependent changes of the related strain-functions (apparatus properties) and rigidity-functions (material properties) this approach will help to better predict transient processes of dustiness in future.

The shape of crystals is an important property that has a great impact on their physical behavior. Examples are flowability, dissolution, and growth kinetics. Still, crystals are often described by a single size parameter. One reason is, that today shape information is still hard to measure. Additionally, only few modeling techniques exist that are able to describe the shape of crystals. In this chapter, these issues are addressed by accurately describing crystals with mathematical models, making the full morphological structure of crystals and their agglomerates accessible by stereoscopic and three-dimensional (3D) imaging techniques and using these methods to model crystallization while considering the complex shape of the crystals. In addition, artificial neural networks (ANN) are used to classify whether projections of crystals show single crystals or agglomerates. As a final step, a case study of a model of a mixed suspension mixed product removal (MSMPR) crystallizer and a hydrocyclone are integrated into the software platform Dyssol and used to dynamically simulate a crystallization process with recycling stream.

Uni- and bi-variate crystallization processes are considered that are modeled with population balance systems (PBSs). Experimental results for uni-variate processes in a helically coiled flow tube crystallizer are presented. A survey on numerical methods for the simulation of uni-variate PBSs is provided with the emphasis on a coupled stochastic-deterministic method. In this method, the equations of the PBS from computational fluid dynamics are solved deterministically and the population balance equation is solved with a stochastic algorithm. With this method, simulations of a crystallization process in a fluidized bed crystallizer are performed that identify appropriate values for two parameters of the model such that considerably improved results are obtained than reported so far in the literature. For bi-variate processes, the identification of agglomeration kernels from experimental data is briefly discussed. Even for multi-variate processes, an efficient algorithm for evaluating the agglomeration term is presented that is based on the fast Fourier transform (FFT). The complexity of this algorithm is discussed as well as the number of moments that can be conserved.

Stochastic simulation techniques for the solution of a network of population balance equations (PBE) are discussed in this chapter. The application of weighted Monte Carlo (MC) particles for the solution of compartmental PBE systems is summarized and its computational efficacy in form of a parallel GPU implementation is pointed out. Solution strategies for coagulation, nucleation, breakage, growth and evaporation are thereby presented. An application example treats the simultaneous coagulation, nucleation, evaporation and growth encountered during particle production through the aerosol route. Furthermore, the simulation of a compartmental network is discussed and parallel simulation techniques for the transport of weighted MC particles are presented. The proposed methodology is benchmarked by comparison with a pivot method for a variety of test cases with an increasing degree of complexity. Simulation conditions are identified, for which conventional, non-weighted MC simulation techniques are not applicable. It is found, that the specific combination of a screen unit with tear-streams cannot be simulated by conventional methods, termed ‘random removal’, and make thus other techniques—like the here introduced merging techniques necessary.

Processes in the field of chemical engineering do not consist of one single step, but typically a high number of strongly interconnected unit operations linked with recycling streams. This inherent complexity further exacerbates when distributed particle properties, i.e., dispersity, must be considered, noteworthy being the case whenever particulate products are in focus. Out of all five possible dimensions of dispersity (size, shape, composition, surface and structure) particle size most often determines the efficiency of particulate products. Thus, its optimization is key to reach tailored handling and end product properties. In this work, a model-based optimization tool for particle synthesis was elaborated which is often the first step of a process chain. It is described by population balance equations relying on the method of characteristics for the numerical simulation and on the usage of gradient information to enhance the performance of the optimization. The presented scheme to optimize time-dependent process conditions in a time efficient manner is applicable for a wide range of particle syntheses.

The application of flowsheet models to dynamic solids processes pose significant challenges, especially regarding the handling of the inherent multidimensionality of granular material properties, like particle size, shape and porosity distributions. The novel open-source flowsheet simulation framework Dyssol deals with this by applying an approach based on transformation matrices, which allows for the tracking of temporal changes in the multi-dimensional distributed parameters of the granular materials. The modelling system utilizes the sequential-modular approach in combination with partitioning and tearing methods as well as the waveform relaxation method for increased modelling flexibility while offering high computational performance. Dyssol includes an extensive and expandable model library for various unit operations in process engineering, that in turn may be calculated by user-defined solver units from a distinct library. To enhance the computational performance, the user may choose from different convergence and extrapolation methods. Material properties are defined in an extendable material database. Various case studies show robust stability and high convergence rates. The application of a global optimization algorithm shows promising results for the operational parameter adjustment in case of transient system behaviour. A concept of applying artificial neural networks to extend the scope of dynamic flowsheet simulation systems is proposed.