Computational Fluid Dynamics – Spray modelling makes a splash

Having been in the vanguard of utilizing computational simulations to identify the effects of fluid dynamics on wet scrubbers, Austrian Energy & Environment was in a good position to meet the challenge of commissioning a wet limestone flue gas desulphurization unit at RWE’s modern, super-efficient Niederaussem power plant in Germany.

Dr. Klaus Baernthaler & Dr. Harald Reissner, Austrian Energy & Environment, Austria

The combustion of sulphurous fuels leads to high sulphur dioxide (SO2) emissions. Modern lignite and hard coal fired power stations require a flue gas desulphurization plant (FGD) to meet the current emission regulations. The most commonly used FGD plant in large-scale power plants is the wet scrubber where limestone slurry is used as washing liquid.


Power generation by source. (TWh). Baseline Scenario versus Role of Electricity Scenario
Click here to enlarge image

In many cases the FGD plant consists of a cylindrical absorber and a radial raw gas duct. Several spray banks are installed in the cylindrical part of the absorber.

The limestone/gypsum slurry is atomized by an array of nozzles which are fed by the spraybank piping. The main aim of the atomization process is to gain a large mass-transfer area in the spray.

SO2 from the flue gas is absorbed into these spray droplets where it is bonded as calcium sulphite. Further, it is oxidized to gypsum.

It is well known that the efficiency of wet scrubbers is affected by the fluid dynamics (beside chemical reactions). Uneven flow profiles caused by poorly optimized raw gas ducts, the arrangement of the spraybanks and the distribution of the nozzles may result in so-called bypass-effects which reduce the efficiency of the scrubber. Austrian Energy & Environment AG (AE&E) detected these effects during SO2-profile measurements in the FGD plant Sostanj 4 in Slovenia.

AE&E has used computational fluid dynamics (CFD) simulations to identify these effects since 1995. CFD simulations are used first to analyze flow phenomena and then to optimize the flow field. CFD is used to describe the local hydro- and aerodynamic effects because it is a complex and costly task to measure local flow phenomena or flue gas concentrations in the absorber.

Approximately 15 large-scale flue gas scrubbers have been optimized by CFD. Many FGD plant optimizations have been supported by measurements for evaluation of the optimization process.

Austrian approach

The approach of AE&E is based on calculating the two-phase flow, which is very distinct in the scrubber, by means of a Euler-Lagrange method, while considering the nozzle positions and the essential nozzle characteristics as well as important flow phenomena, such as the interaction of the slurry droplets with the flue gas and the scrubber wall.

For the turbulent effects, a standard k-epsilon model is sufficient. Additionally, the composition of the slurry and/or of the watery phase (the latter is treated as the absorbent of the absorbed) is of relevance to the dispersal phase.

Due to these circumstances, there is the necessity to consider the mechanisms of the mass and heat transfer and of the interaction with the wall. These mechanisms are packed into “user routines” for wall interaction, evaporation and absorption of SO2 that have been developed in recent years.

The user routines have been tested and verified individually. For the condensation routine the listing is as follows: multi-droplets experiments in a single control volume cell and detailed simulation of an injection condenser. For the SO2 user routine it is: single-droplets experiments as well as the calculation of a great number of FGD plants (e.g. a pilot plant of a diameter of 1.4 m, large-scale scrubbers of diameters between 5 m and 20 m). The benefit of AE&E’s clients is to have an optimized FGD system with lowest operational costs (especially regarding recirculation pump energy).

Niederaussem gas scrubber

When AE&E got the order for delivery of a wet limestone FGD plant at the 950 MW lignite fired boiler (unit K) at RWE’s Niederaussem power station in North Rhine-Westphalia, two new flue gas scrubbers, each taking care of 50 per cent of the total flue gas volume, had to be integrated into the existing flue gas cleaning systems (9 absorbers with a total capacity of 2850 MW).

During commissioning, several measurements campaigns, including SO2 profile measurements, were done to check the performance of the FGD plant. These measurements were well suited for a verification of CFD modelling results.

The cylindrical scrubber with a diameter of 15.3 m has a radial inlet and is equipped with four spraying levels. The fifth spraying level is intended as a spare level only and so, at present, it is not equipped with nozzles.

The simulated area consists of the raw gas duct that ends radially in the cylindrical scrubber, the absorber with the spraying levels, the demisters and the curved outflow area. The real absorber geometry was integrated almost completely into an unstructured calculation grid, which consists of almost one million cells.

Each spraying level has 124 Lechler nozzles. For the inner area double eccentric hollow-cone nozzles with a spraying angle of 90à‚º were selected in order to achieve a high mean residence time of the droplets. These double eccentric hollow-cone nozzles spray 35 per cent of the volume flow downwards and 65 per cent upwards.

In the outer area 90à‚º eccentric full-cone nozzles (direction of spray: 100 per cent downwards) are used in order to counteract the high gas velocities in the edge zones of the scrubber and to minimize the share of the slurry flow sprayed onto the scrubber wall. The nozzle inlet pressure is approximately 1.0 bar and the scrubbing slurry volume flow sprayed by each nozzle level amounts to 7400 m3/h.

The inlet parameters of the raw gas for the CFD modelling were chosen in accordance with a real load case: the flue gas mass flow is 720 kg/s and the raw gas temperature 100à‚°C. At this load case only spraying levels 2, 3 and 4 are switched on; scrubbing slurry is not admitted to the lowermost spraying level.

The most important data are summarized in the table below.

Assessing the results

After the inflow into the cylindrical part of the scrubber the flue gas is distributed very homogeneously over the diameter of the absorber by means of a design that has been optimized several times. The optimized control of the flow, together with the combination of various nozzle types and properly adjusted spraying levels, enable the avoidance of so-called short-circuited flows and flows in the edge zones of the gas in the apparatus (the latter results in high-flow velocities in the edge zone).

Click here to enlarge image

Furthermore, the flow around the internals, where the flue gas is displaced, and the downstream influence upon the flow after such obstacles, can be seen as blue areas (low upward velocity) within green areas (high gas velocities).

The influence of the residence time of the flue gas in the active scrubbing zone of the scrubber starting at the upper edge of the inlet duct can also be recognized. The aforementioned optimizations of the nozzles and the scrubber meant increased concentrations of SO2 at the edge zones of the scrubber were avoided almost completely.

Zones with high concentrations of SO2 often appear where the vertical velocities are high. It can be shown that in the area of the fourth spraying level SO2 concentrations appear exactly where high vertical velocities had been determined before for the areas of the first and the second spraying levels. This can be attributed to the short local residence time of the flue gas in the areas with high vertical velocities.

The figure below shows the comparison of the measurement and the CFD simulation at the measuring level of 26.3 m above the sump. The profile measurement is located directly downstream of the conical part of the scrubber.

The position of the four measuring axes results from the arrangement of the four measuring lances in perpendicular orientation to each other. The flue gas was sampled via heated measuring probes, heated piping and a gas cooler, and then admitted to the analyzing devices.

Sulphur dioxide concentrations

The CFD simulation provided a good description of the increased SO2 concentrations in the outer areas and the low concentrations in the core area.


SO2 concentration profile in parts per million, comparison of measurements and the CFD calculation for the measuring level of 26.3 m above the sump
Click here to enlarge image

It was also shown that the absorption efficiency in the right half of the scrubber is better than in the left half. This was mainly due to minor obstacles that are constituted by the internals in the left half of the scrubber.

It should be noted that the differences appearing between the measurement and the calculation are due to the considerably different numbers of measuring points and calculation points. During the described load case the measured absorption efficiency was 93.9 per cent. The CFD simulation calculated an absorption efficiency of 95.2 per cent, just 1.3 per cent different from the measured value.

On the basis of a two-phase standard CFD code several additional functions were integrated by means of user routines. This allowed the calculation of the evaporation and/or condensation, the mass transfer of SO2, the acidification of the scrubbing slurry droplets during their flights and the effects of the impact of a droplet at a wall.

During comprehensive numerical simulations and measurements it was shown that the selection of the nozzles has a decisive influence on the flow conditions and therefore also on the absorption efficiency of SO2. The SO2 concentrations determined during complex profile measurements are in good agreement with the values of the CFD simulation.

This powerful engineering tool, developed as a result of AE&E’s experience of over 25 years in spray tower design, is used for the retrofit of old spray tower systems as well as the design of new scrubber systems .

These retrofitted spray tower systems are optimized via CFD modelling to fulfil the more stringent emission limits (e.g. EU-2001-80) by the lowest effort principle regarding invest costs and future operation costs.

No posts to display