|SNCR could overtake SCR as the deNOx system of choice in large coal-fired power plants|
In the eighties and nineties when Western Europe’s large coal-fired boilers were retrofitted with nitrogen oxides (NOx) control systems, selective catalytic reduction (SCR) process was considered the best available technology (BAT). But subsequently, when many Central and Eastern European countries joined the European Union and had to accept its emission limits, interest in SNCR grew because it offered advantages such as lower investment costs.
Especially in recent years, the SNCR process has been steadily improved for small and medium-sized boilers like waste incineration plants, for which it is widely considered the BAT. But power plant owners are now also investigating whether SNCR is feasible for their large coal-fired boilers too, on both performance and cost considerations.
SNCR and SCR are post-combustion NOx control technologies and both work to reduce NOx to nitrogen and water by using reagents based on either ammonia or urea. The main difference between the two systems is the ‘temperature window’, i.e. without a catalyst the reaction takes place at 900-1050 à‚°C, whereas with the catalyst the range drops to 160-350 à‚°C.
In the SNCR process, reagents in aqueous solution (ammonia water, urea) or in gaseous form (ammonia) are injected into hot flue gases. For an optimum NOx-reduction with a minimum ammonia slip (NH3 slip) it is necessary to evenly distribute and thoroughly mix the reagent in the flue gases within the appropriate temperature window in which NOx-reduction is possible. The optimum temperature range to achieve high NOx-reduction combined with a minimum consumption of reagent and a low NH3 slip is rather narrow and depends to a great extent on the flue gas composition. For coal-fired boilers the optimum temperature lies between about 960 à‚°C and 1020 à‚°C.
Above this temperature range ammonia is oxidized to an increasing extent, i. e. NOx are formed, while at lower temperatures the reaction rate is slowed down, causing an NH3 slip, which may result in the formation of ammonia salts and lead to secondary problems.
However, because temperatures over the cross-section in the furnace are rarely uniform and considerable imbalances are often found, special measures need to be taken to identify the right positions for the injectors to distribute the reagent properly into the flue gas under all operating conditions.
To determine whether the SNCR process would suit an existing coal-fired boiler, it is recommended to perform simple tests with a portable test installation. Such tests can provide valuable information, not only on what efforts need to be made with regard to the design and the equipment of a commercial SNCR plant, but also on the performance that can be expected and guaranteed under varying operating conditions.
Regardless of whether ammonia water or urea is to be used in the subsequent commercial plant, tests are generally performed with urea solution because it is easy to handle. In addition, from a performance point of view, both reagents are comparable in most applications. To-date tests have been conducted on several boilers with capacities up to 225 MWe in Germany, the Czech Republic and Poland.
In one particular example, a German utility decided to go with the SNCR process for a 200 MWe coal-fired boiler, after successful testing and taking into consideration relevant aspects, such as the level of NOx reduction, the cost-benefit ratio and overall plant availability,
Effective combustion chamber diagnosis
However, temperature measurements with suction pyrometers and the readings from permanently installed thermocouples only permit a rough estimate regarding the temperature profiles in the individual potential injection levels during the respective boiler loads. Furthermore, the temperature distribution and imbalances resulting from the boiler load, the ignition behavior and the burner configuration, for example, may vary strongly.
To ensure that the reagent is always injected in the upper range of the temperature window under any operating condition, i.e. in the range where NOx reduction is highest and NH3 slip is lowest, acoustic gas temperature measurement systems (agam) (Figure 1) should be installed. Agam measures the real gas temperature and determines profiles across the entire combustion chamber cross-section.
|Figure 1: Temperature profiles measured by agam ensure the SNCR plant is optimally operated|
The system consists of transmitter and receiver units that have an identical mechanical and electrical design mounted to the walls of the combustion chamber and an external control unit. During the measurement the solenoid valve in the compressed air line on the transmitter side is opened, generating acoustic signals. The signals are recorded simultaneously on both the transmitter and the receiver sides. The digitalised signals are then used to measure the transmission time.
Since the distance is known, the velocity of sound can be determined, which is then converted into a temperature, i.e. the path temperature. With several combined transmitter/receiver units acting on one level, multiple path configurations can be obtained to calculate the two-dimensional temperature distribution in one level immediately.
A temperature profile is divided into sections and can be assigned to individual lances or groups of lances to switch them to another level depending on the flue gas temperature measured. This ensures the reagent gets to the most effective locations for the reaction, even with rapidly varying flue gas temperatures, and that the SNCR plant is always operated in the optimum temperature range.
After the German utility decided in favour of SNCR, a preliminary agam was installed to obtain detailed information and help inform the design of the commercial SNCR plant, in particular its injection levels and the number and positions of the injectors.
The temperature measurements were performed at the end of the combustion chamber (at 39 metres) with different loads and configurations of pulverisers. Four symmetric zone temperatures were determined from the temperature matrix and the surface average value was used to calculate deviations for the zones. It showed the average temperature at the end of the combustion chamber varied between 750 à‚°C at low load (45 MWe, burner level 1) and 1155 à‚°C at full load (185 MWe, with all burners in operation).
The final engineering concept for the SNCR plant for the the 200 MWe coal-fired plant was based on the analyses of the temperature measurements and the tests with the SNCR demonstration plant
A simplified process flow chart in Figure 2 shows the function and the scope of supply of the commercial SNCR plant as designed, installed and commissioned in the power plant. Because of the significant temperature differences between low load and full load, as well as the extreme temperature imbalances, five injection levels were installed from 26-51.8 metres. The injectors were arranged so that the right and the left sides of the boiler could be controlled independently, with each injection lance individually activated or deactivated.
|Figure 2: A flow diagram of the commercial SNCR plant, featuring five injection levels and agam|
The commercial SNCR plant entered operation in March 2010. The guaranteed NOx and NH3 clean gas values were attained in most cases, with boiler loads ranging from 20-100 per cent.
The subsequent optimisation phase, however, was time consuming because at each of the five injection levels the temperature profile had to be measured at various loads with suction pyrometers to calculate the difference from the temperatures measured with the agam at the 39-metre level. This was necessary to determine which lances should be operated at various average temperatures in the zones and at which temperatures the switching should be effected at given loads.
SNCR demo in a 225 MWe plant
In a Polish power station with five 225 MW coal-fired boilers, an SNCR demonstration was carried out to validate that a NOx reduction of at least 25 per cent can be achieved safely at any boiler load between 40-100 per cent.
Temperature measurements, which could only be performed at two openings at 47.4 metres, found imbalances of more than 120 K between the measuring points. It was not possible to make further measurements because there were no other openings large enough to accommodate a pyrometer lance. During the tests, the urea was injected through openings at levels 37.9 metres and 47.4 metres from the front wall, as well as from the side walls at 47.4 metres.
Despite these challenges the results were very positive, with NOx reduction far above the 25 per cent target at all loads and at almost 60 per cent with 75 per cent load.
In a commercial plant, a third level for injecting the reagent would improve performance, especially regarding efficiency and NH3 slip. To minimise this, a small catalyst could be introduced at the end of the boiler. But with an agam like the one installed in the German boiler the reagent could be injected more precisely at the optimum temperatures. As a result the slip could be maintained low enough to keep the ammonia concentration in the fly ash below an acceptable limit so that an additional catalyst slice would not be needed.
Overall plant availability is essentially unaffected by SNCR systems. Components critical for plant operation such as pumps are provided with redundancy. Although, the injection lances in contact with the flue gas must be regularly checked and serviced, they can be checked during operation and replaced relatively quickly if required.
The SNCR system in the German power plant, for example, is equipped with an automatic data acquisition system to facilitate fault diagnosis and settings via remote data connection. The higher investment costs of such a system can be paid off within a short period of time since the expense of costly visits of service engineers can be avoided.
The TWiN-NOx process
Once the decision to use a SNCR system has been made it is crucial to select the best reagent. Urea offers advantages in availability, logistics and cost. Yet process considerations could make ammonia water the better option.
Coal-fired boilers essentially fall into two design concepts. The main boiler design features two flue gas passes and a contraction nose at the end of the furnace. The alternative is the tower boiler.
In the two-pass boiler at full load the optimum temperature is mostly in the level of or within the super heaters. The use of ammonia water as a reagent is often limited by the temperatures, which are mostly too high, so that a lot of the ammonia will burn to NOx before it can reach the area with lower temperatures within the heat exchangers. Therefore, the overall NOx-reduction is not optimised.
With urea solution, the situation is easier to handle because by the time the water droplet surrounding the urea particle has evaporated, the NH2 of the decomposed urea will have reached the cooler area. However, there is serious concern that droplets containing urea would impinge on the boiler tubes causing corrosion and damage of the tubes. Therefore, special attention has to be paid to the positioning, maintenance and operation of the injectors.
The situation with tower boilers is no easier, although the reagent can be injected in most applications from all four sides of the boiler.Only the intermediate area lying between the colder boiler walls and the hot centre offers an optimum temperature range for the reactions. Special measures are needed to achieve sufficient distribution of the reagent in the flue gas. One alternative is to inject the reagent in several levels simultaneously with different penetration depths and to use lances of different lengths. But optimum distribution of the reagent is still difficult to achieve.
During the testing of the SNCR process in the 200 MWe coal-fired boiler, in Germany urea solution was used despite the subsequent commercial plant utilising ammonia water. However, the operating results of the commercial plant did not meet expectations, especially at full load.
Disappointingly results showed that automatic control was no better than the manually-controlled trial equipment. The only significant difference is that ammonia water is used in the commercial plant. It may perform less well than urea because it reacts too close to the boiler wall.
To investigate this, additional tests with urea were performed in the commercial plant. The results showed that immediately after injection of urea, NOx reduction rose and consumption of the reagent fell, but concern remained over urea’s impact on the boiler tubes.
Further tests showed that the low volatility reagents (urea solution – NOxAMID) are indeed released at the end of the droplet trajectories while the high volatility reagents (NH3) are released near the droplet source close to the boiler walls.
Subsequent tests showed that by changing the reagents according to operating conditions the performance of the SNCR could be improved considerably. From there, it was a small step to mix the two reagents and inject various mixtures into the furnace to combine the best features of both.
A commercial plant has now been built that can be operated alternately or simultaneously with urea solution and ammonia water (Figure 3). This process, called TWiN-NOx, gives a more effective and wider temperature and load range, higher efficiency, lower NH3 slip, less consumption of reagent and minimum risk of corrosion.
|Figure 3: A commercial plant utilising the TWiN-NOx SNCR process in now in operation|
The future for SNCR
SNCR has now been demonstrated to provide results that are comparable with those for catalytic NOx reduction but at a fraction of the cost. Even in large combustion plants, greater NOx reduction can now be achieved through temperature-controlled adjustment of individual lances. The temperature profile could be significantly improved and extreme NOx peaks prevented if temperatures measured by agam were used not only for regulating the SCNR plant but also for optimising the combustion process.
All feasible and commercially justified technological measures such as optimising combustion and flue gas recirculation should be taken. A small additional slice of catalyst at the tail end of the boiler could minimise the NH3 slip. The TWiN-NOx process is expected to open up further potential for improvement.
Over many years of continuous operation at various combustion plants, SNCR has proven to be reliable and economical for NOx reduction. In the power plants highlighted here expectations were always met and generally exceeded.
From the process point of view, it is almost irrelevant whether urea solution or ammonia water is used as long as plants are engineered, installed and operated in an appropriate manner. In Germany, Sweden and the Netherlands, SNCR has been operating for several years in waste incineration plants with designed NOx limits of <100 mg/Nmà‚³. These plants reliably comply with guaranteed values in continuous operation. Newer plants, equipped with agam and three injection levels, achieve low NH3 slip, low NOx clean gas values and high efficiency.
Although SCR can offer slightly higher NOx reduction levels, the cost-benefit ratio is generally lower, particularly as NOx values below 300 mg/Nmà‚³ are now generally obtained at large coal-fired boilers through combustion modifications alone.
Finally, promising test results have now been recorded for oil-fired, as well as coal-fired plants with capacities up to 225 MW. In heavily coal dependent countries like Poland and the Czech Republic, the SNCR deNOx process for large power plants are beginning to find favour.
Power Engineering International Archives
View Power Generation Articles on PennEnergy.com