Institute for Energy Process Engineering and Fuel Technology > Research > Development of fuel specic subprograms for CFD-Forecasts of industrial prozesses

Development of fuel specic subprograms for CFD-Forecasts of industrial prozesses

MILD / Flameless Combustion

Mathematical modeling of industrial processes has always been one of the areas of expertise of the IEVB. Currently the emphasis has been placed on both the development and validation of Computational Fluid Dynamics (CFD) computer codes that can handle simultaneously fluid flow, heat transfer and combustion chemistry. In this context MILD combustion (see Project 1 for a definition of MILD combustion) has been given a priority since the availability of mathematical models that can confidently predict this novel combustion technology is scarce. The ultimate goal of this project is the development and validation of CFD based software for designing industrial furnaces utilizing MILD combustion. The software is to predict the fluid flow, heat transfer rates as well as combustion chemistry including NOx and CO emissions. There is hardly any institution or R\&D group that can reliably predict properties of MILD combustion or Flameless Oxidation and there exists intense competition among the various research groups developing such a capability.

The CFD code FLUENT is used as a basic software. The code has been extended by a series of application-specific subroutines developed at the IEVB. Using this code, stationary simulations of a 580 kW burner in the MILD-operation have been concluded. The different models of combustion and NO\textsubscript{x} were then further explored and compared.
Detailed calculations on combustion chemistry are performed using CHEMKIN program. The overall mathematical modeling approach of the IEVB is to couple both computer codes in order combine their complementary capabilities. The FLUENT code is very good at calculating flow and temperature fields. Despite steady advances in computer technology, this is only possible on modern high performance computers using combustion models based on some crude simplifications of the reaction mechanisms.

The CHEMKIN approach follows a different philosophy - one devides furnances into distinct zones  and present the system as a network of reactors (see Figure 2.6.1). When calculating the properties of a furnace using such a reactor network, one considers extensive mechanisms with several hundred individual reactions. In this way, valuable information can be obtained about the primary reactions or paths leading to the creation of NO\textsubscript{x} and CO. The results of both modeling approaches are compared with each other and with taken measurements.

 

 

 

Figure 2.6.1: Coupling of the CHEMKIN-code (grey) with the FLUENT-code (colorful
part, top right) for simulations of MILD-Combustion

 

It has been demonstrated that with the exception of the small region located within the natural gas jet, the computations have resulted in predictions of good quality. The heat transfer characteristics have been very well predicted. The overall fluid flow pattern and chemistry field have also been well predicted. The predicted NOx emissions have been in very good agreement with the measured values for a number of combustion-model and NOx post processor combinations.

The mathematical modeling has demonstrated that one of the most striking characteristics of MILD combustion is the lack of intensive temperature fluctuations. To obtain good quality NOx predictions it was enough to predict correctly the time-mean temperature and oxygen concentrations in the flamelet region where NOx was formed.


Currently the work concentrates on:

  • The calculation of the performance of MILD combustion technology applied to
    industrial reheating furnaces.
  • The assessment of the potential of MILD combustion for other industrial applications.
  • The development of models for blast furnace gas combustion.
  • To develop the capabilities needed for the modelling of unsteady combustion processes.

 

 

Figure 2.6.2: Flow paths of gas and air during MILD combustion (gas lights, left) and
temperature distribution in a boiler specically designed for MILD combustion
(coal, right).

 

Thermodynamics, Heat Transfer, and Fluid Mechanics of Heat Recovery in a Coke Oven (HR)

Due to changing market demands concerning the environment, older varieties of coke ovens were built on an industrial scale in the last two decades in Brazil, India, the USA, and especially in China. In such coke-plants the raw gas generated during coal pyrolysis is not used for  production of chemicals. It is burned in the combustion chambers to produce the required process heat for the pyrolysis process. Such furnances are called Heat Recovery Ovens (HR).

In the realm of coking specialists, discussions in context of HR technology increasingly focus on determining the coking time. With this comes the completion of the pyrolysis process and the coke can then be removed from the furnace. When the correct operating temperature is reached, the amount of volatile matter in the coke is generally less then one percent and, accordingly, the temperature of the coke is over 950~°C. The reliable assessment of the coking time for a given coke production amount, and in conjunction with the usable oven volume, is an essential design criterion for the number of ovens required for a coking plant.

If one attempts from the literature, company brochures, and other data to derive an estimate of the coking time, one will be confronted with an enormous spread of data.
Although the mathematical description of unsteady heat transfer operations in horizontal ovens was already a subject of theoretical and practical investigation in recent decades, there has been so far no theoretical work presented to formulate the heat balance of the HR furnace.  The warming processes inside such ovens can only be calculated with the currently available models for a steady state system without internal heat sources.

Supported by a partner in industry, the IEVB is working to expand the thermodynamic and fluid dynamic principles of HR coking technology with an in-depth theoretical understanding and to present general models for calculating important process variables. This work has a comparable and, in parts, greater complexity than the mathematical models of conventional coke oven technology, which is already documented in the literature. In particular, the complex relationships between the furnace roof and the coupling of energy demand and supply of energy source present a major challenge to modeling.

For industrial applications, the development of a transient simulation model is desired that considers the main heat transfer processes in HR ovens and that can calculate the operating time. In this regard, the applicability of various one-and two-dimensional models is investigated.

Given that the physical and chemical parameters for this coking process were, until now, not recorded in literature, the first step is to establish the data set for compact furnace charges. This is the foundation of the experimental undertaking. Both the thermodynamic boundary conditions (such as raw gas volumetric flow rate, excess air ratio, emission levels) and physical-chemical properties (such as temperature conductivity coefficients and specific heat capacity) must be, to a large extent, determined by practical methods of measurement. In order to achieve numerical convergence, the furnace design was ascribed a simple geometry. HR furnaces create a plug flow temperature profile along the length of the chamber.  To make sure that the stirred tank model conforms with the model assumptions, it is necessary to develop design solutions and principles that provide a uniform distribution of combustion air as well as even heat transfer to the coal along the base of the furnace.

This work is beyond the task of deriving design principles for the optimization of the mixing behavior in the combustion chamber. Even using three dimensional complex flow and combustion models, this is not achievable. The focus of the study is primarily to design and arrange the primary air openings and ducts of the downcomer.

To increase its numerical efficiency, it is advantages to divide the furnace into appropriate sections and half sections. Helping with this task is the FLUENT 6.2 software, whose system is based on a finite volume comprised of many segments, as considered above. A sub-model for calculating the pressure profile inside the furnace completes the fluid dynamic simulation package for the HR-oven.

The approach has been divided into several phases and is based on the distribution of the furnace geometry into several sub-models:

  • One-dimensional, transient coking model for the compacted coal / coke barrier in the chamber of the upper furnace based on the plate geometry to determine the operating status "done cooking".
  • Two-dimensional, transient coking and combustion model for the upper chamber of the furnace to determine the operating status "done cooking" and to display characteristic temperature proles. The calculation model has no geometric coupling between the upper and lower furnaces. The procedural coupling to the lower furnace is adapted to be time-dependent.
  • Two-dimensional, transient coking and combustion model to determine the operating status "done cooking" and to display characteristic temperature proles in the upper and lower furnace, while taking into account a comprehensive procedural and geometric coupling of the upper and lower furnaces.
  • Three-dimensional steady-ow models, in which the optimization of layout and design of the primary and secondary air inlets is the main focus.
  • Validation of the models using reference cases from bulk and compacting operations.

 

On the left side of Figure 2.6.3, a stationary temperature profile for a HR - furnace is drawn. The right side shows for a selected furnace operating condition the temperature curves for a two-dimensional, unsteady computation, that at the end of the investigated operating paths passes in a two-dimensional, arranged on the right side end profile.

 

Figure 2.6.3: Stationary and transient temperature proles of a HR furnace

 

Using the two-dimensional model in the next step, the application limits of this coking technique should be developed with process control in mind. The peculiarity of this coking technique comes from the low influence on the energy supplied by the release of raw gas in connection with the combustion control. Thus, we will only investigate energy and material-specific limits.From previous simulation results, it can already be deducted, that in order to achieve optimum operating times, the HR - oven must be operated at the highest temperature feasible. With regard to the temperature regime in the furnace, a defined and timely calibration of the primary and secondary air systems is of high importance. In addition, it is expected that there is a lower limit for the volatile components of the feed mixture above which the coking process is unacceptable or completely hindered. This is confirmed by practical experience. A high percentage of volatile matter leads to, however, extreme furnace temperatures in excess of the permissible limits of the refractory material. We already know from current knowledge that in the current processing state, there is only a limited amount of coal in the world available that provides the appropriate heat balance in a "high performance business".

 

Gasication of Biomass Sludge (Slurry)

The use of biomass as a CO2-neutral, renewable fuel has gained importance in recent years because of the limited availability of fossil fuels and their supposed impact on global warming. Therefore, it is expected that biomass will gradually replace some usage of fossil fuels.

In cooperation with the Research Center Karlsruhe (FZK), the IEVB is currently working on the design of a biomass gasifier. At the FZK, a new procedure - BIOLIQ(R) is being developed. This method converts straw and other agricultural byproducts consisting of lignocellulosic biomass by pyrolysis and subsequent gasification into a synthesis gas. This tar-free synthesis gas can then be converted into synthetic fuels or methanol.

The efficiency of the conversion of biomass into synthesis gas is the key factor for the success of this new process. In order to design the 500 MW gasifier, one must first understand the behavior of the biomass (a flammable liquid with embedded coke particles) gasification process. To help design the layout of the gasifier, CFD simulations are being used to calculate industrially important aspects that are directly related to one another, such as mixing, chemical reactions, and heat transfer.

Several models are simulated in order to meet the required accuracy of industrial applications, due to the high temperature and pressure requirements of the technology.

Unfortunately, most of the currently known models from literature are not suitable for the present operating conditions. Therefore, the following models have been developed or adapted at the IEVB:

  • Gas and liquid properties under non-ideal conditions;
  • Radiative properties of the substances used under real operating conditions (Temperature, pressure);
  • Bio-oil evaporation at high pressure and high temperaturer;
  • Coal and biomass gasication

An exemplary result is shown in Figure 2.6.4. One can see on the right side, the gas flow lines and on the left side, an excerpt of the particle trajectories. An important aspect of the design is the position of the gasifying zone. In the depicted design, it is expected that the gasifying zone of the solid coke is in the middle of reactor. This allows for the gasification process to be completed and provides the possibility that the water-gas shift reaction is moved towards chemical equilibrium. This shift results in more combustible gases being formed from water and carbon dioxide. This requires that the recirculation paths and forces be determined within the reactor. In this configuration, the recirculation zone of the reactor is positioned at the bottom. This interpretation ensures that coke particles, which have not yet reacted, are transported back to the middle of the reactor where the reactions take place more quickly and more effectively.

 


 

Figure 2.6.4: Particle trajectories in the entrained ow carburetor and ow lines of the
gas in the gas recirculation zone.

Contact:Dr.-Ing. M. Mancini
 

Search  Sitemap  Data Privacy  Imprint
© TU Clausthal 2019