Combined NO x and NH 3 Slip Reduction in a Stoker Boiler Equipped with the Hybrid SNCR + SCR System FJBS+

: The application of secondary NO x control methods in medium to low-capacity furnaces is a relatively new topic on the energy market and thus requires further research. In this paper, the results of full-scale research of SNCR and hybrid SNCR + SCR methods applied into a 29 MW th solid fuel ﬁred stoker boiler is presented. The tests were performed for a full range of boiler loads, from 33% (12 MW th ) to 103% (30 MW th ) of nominal load. A novel SNCR + SCR hybrid process was demonstrated based on an enhanced in-furnace SNCR installation coupled with TiO 2 -WO 3 -V 2 O 5 catalyst, which provides extra NO x reduction and works as an excess NH 3 “catcher” as well. The performance of a brand-new catalyst was evaluated in comparison to a recovered one. The emission of NO x was reduced below 180 mg NO x /Nm 3 at 6% O 2 , with ammonia slip in ﬂue gas below 10 mg/Nm 3 . Special attention was paid to the analysis of ammonia slip in combustion products: ﬂue gas and ﬂy ash. An innovative and cost-effective method of ammonia removal from ﬂy ash was presented and tested. The main idea of this method is ﬂy ash recirculation onto the grate. As a result, ammonia content in ﬂy ash was reduced to a level below 6.1 mg/kg.


Introduction
Nitrogen oxides NO and NO 2 (known together as NO x ) emitted from power units are claimed to significantly contribute to the formation of numerous environmental pollutants, such as acid rain, haze, or smog. The technologies of selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR) are commonly implemented to reduce NO x emissions in the power sector [1,2]. Their effectiveness is based on the reaction of nitrogen oxides with ammonia or urea compounds. In general, the reduction of nitrogen oxides takes place according to the following chemical reactions [3,4]: In fact, the reaction mechanism is very complex: 31 compounds and 173 reactions that occurred during the NO x reduction process were identified [5]. The most efficient NO x conversion in the non-catalytic process can be reached in a temperature range of 870-1100 • C [6,7] while in catalytic conversion in 180-450 • C, for most catalysts 300-350 • C [8,9]. SCR technology is possible to be applied in waste-to-energy units, including combustion of medical and hazardous waste [10,11]. It can be potentially used to reduce the NO x emission not only from power plants and waste combustion units but other stationary sources as well (e.g., chemical and metallurgical industry) [12][13][14].
While large combustion plants were successfully confronted with the challenge of reducing pollutants, the small and medium-sized plants (<50 MW th ) became the subject of emission regulations quite recently. However, scaling the emission control technologies down from large to smaller plants has not been an easy task. Satisfactory NO x reduction efficiency is particularly challenging in small furnaces of low-to-medium load power units, such as stoker boilers.
Stoker boilers are commonly used in heating plants all over Europe. They are easy to operate, can consume low-quality fuels such as RDF (Refuse Derived Fuel) or waste, and do not require advanced maintenance. In the past, there was not enough focus on developing flue gas cleaning methods for stoker boilers and frequently they have not been equipped with efficient systems to reduce nitrogen oxides emissions. Hence the application of secondary NO x control methods is a relatively new topic in this energy sector. Therefore, selective catalytic and non-catalytic reduction methods and their hybrid installations still require research and improvement.
Compared to pulverized fuel boilers and fluidized bed boilers, the combustion process in a stoker furnace is characterized by different temperature and flue gas velocity distribution inside the combustion chamber. The following differences should be emphasized:

•
Lower maximum flue gas temperature • Significant fluctuations of flue gas temperature in the combustion chamber • Greater flame diffusion in the combustion chamber • Different flue gas flow profile.
These characteristics are vital for the design of flue gas cleaning systems, especially methods for nitrogen oxides reduction. Thus, common denitrification solutions may not be applicable in stoker boilers.
Usually, nitrogen oxides concentration in stoker boilers flue gas is in the range of 300-400 mg/Nm 3 at 6% O 2 . Based on the authors' experience and previous research, NO x concentration strongly depends on the stoker operating conditions: the operation of primary air zones, speed of grate movement and thickness of the fuel layer on the grate. There is very limited research concerning the application of denitrification systems in stoker boilers and usually SNCR is the only one investigated [15][16][17][18]. For most European technologies effectiveness of NO x reduction is in the range of 50-65%. Most of those technologies are based on ammonium reagent dissolved in demineralized water, in contrast with the FJBS (Furnace Jet Boiler System) and FJBS+ technologies proposed in this article.
Presented FJBS+ technology for NO x reduction combines SNCR and SCR methods. For typical SNCR installations, a stoichiometric excess of the reagent does not exceed 2.5 while SCR usually operates with an almost stoichiometric ratio of reducing compound what finally guarantees low NH 3 slip. The characteristics of both methods allow them to be combined into a so-called hybrid SNCR + SCR system. In this case, ammonia slip generated by the SNCR is used as a reducing compound at the catalyst surface. As a result, the effectiveness of NO x reduction is enhanced together with the elimination of remaining NH 3 in the flue gas. Such installations may be attractive for facilities that already have been equipped with non-catalytic systems and for which emission limits became more restrictive so that independent operation of SNCR does not permit the fulfillment of current standards. Moreover, a catalyst may operate as NH 3 "catcher" to reduce excess ammonia slip in flue gas.
Increased ammonia concentration in flue gas rich in SO 3 and H 2 SO 4 favor the formation of ammonium salts such as ammonium bisulfate NH 4 HSO 4 (ABS) according to Equations (3) and (4). Other salts (e.g., ammonium chloride) can be also found but usually in minor amounts.
Ammonium bisulfate is a sticky, corrosive substance prone to the deposition in the flue gas ducts. Forming ABS deposits is more likely in lower temperature zones (<340 • C) [19,20]. ABS deposition is likely to occur on the SCR injection components, in the cold end areas of the rotary air heaters and on catalyst surfaces reducing catalyst reactivity [21]. The degradation of catalysts due to ammonium bisulfate deposition is an crucial issue that decreases the catalyst efficiency. ABS formation takes place in the gas phase by the nucleation of NH 3 , SO 3 and H 2 O. The substance then builds up onto the catalyst and blocks the access to its surface.
In order to minimize the risk of ABS formation, it is reasonable to pre-clean the flue gas of sulfur compounds before it enters the catalyst. Many effective technologies of wet, semi-dry and dry flue gas desulphurization have been developed for industrial implementations in power engineering and other sectors [22][23][24][25]. The main sorbents that enable high efficiency of SO 2 and other acidic pollutants removal are based on calcium (Ca(OH) 2 , CaCO 3 ) and sodium (NaHCO 3 ) [24].
In addition to ABS the main factors influencing catalysts deactivation are alkali metals and arsenic [26,27]. The compounds of alkali metals tend to agglomerate on the interface of a catalyst. They affect the catalytic process by preventing NO and NH 3 from accessing the reaction zone [28]. Furthermore, a loss of activity is likely to occur by the reactions of alkali metals with acid sites [29,30]. The presence of arsenic may reduce the strength of acid sites, the catalytic surface area and the number density of active sites [26,31,32]. For these reasons the lifespan of the most commercial SCR is usually no longer than 24-36 months. As a result, the problem of safe disposal of catalysts after their depletion is growing. SCR catalysts contain various hazardous compounds, hence their proper utilization is a great challenge.
To minimize the production of hazardous waste and make the SCR technology more cost-effective some methods of catalyst regeneration have been investigated. Up to date, only limited research deals with the regeneration of industrial catalysts from full-scale power units. The procedures of regeneration are usually investigated under simulated, laboratory conditions, that may be inconsistent with field conditions [27,33,34]. The investigated regeneration methods may differ, however, some of them can be characterized by high efficiency. Deactivated SCR catalyst may reclaim 92% of its initial efficiency after being treated with 0.5 M H 2 SO 4 solution [35] or 91.4% after in situ high-temperature regeneration with water vapor [36]. However, some methods, such as washing with sulfuric acid solution, may result in the corrosion of power equipment or destruction of active components [37].
The presence of NH 3 remaining after the denitrification process in flue gas and fly ash is unfavorable not only because it leads to ABS formation. It is also important in terms of its further use in the concrete industry. Fly ash may be used as a concrete additive if it meets the appropriate criteria of chemical and physical properties. In the process of concrete production, the pozzolanic properties of ash are initialized by the generation of highly alkaline free lime. In such alkaline conditions, ammonia is liberated from ash into the gas phase. The equilibrium of ammonium ion in solution is shifted to molecular ammonia in accordance with the equation: The released molecular ammonia is characterized by a strong smell. The physical and mechanical properties of the concrete are not affected by the presence of ammonia, however, the smell is intolerable to the users and manufacturers [38,39]. Hence the ammonia content of 100-200 mg NH 3 /kg is claimed to be the limit for concrete production [40]. Nevertheless, many industrial manufacturers accept only fly ash with no more than 50-100 mg NH 3 /kg.
As a remedy, numerous methods to eliminate ammonia from fly ash have been investigated last years. These methods can be divided into thermal and chemical ones. Thermal methods base on the high-temperature ammonia desorption process and usually take place at a temperature range of 300-450 • C. They are usually less complex than chemical methods, which makes them more cost-effective [41]. In the Energy Research Center Method fly ash is processed in a fluidized bed reactor where air is supplied as a fluidizing medium. This method allows removing up to 90% of ammonia, however, it has been tested for ash with 500-1000 mg NH 3 /kg and its efficiency for low NH 3 concentrations has not been investigated [42]. The Carbon Burnout method requires a fluidized bed reactor as well to decompose ammonium compounds together with burning out an unburnt carbon. The process lasts for 45 min at a temperature of 700 • C. Ammonia content below 5 mg NH 3 /kg can be reached with a simultaneous reduction of total organic carbon, which has a positive impact on the ash characteristics in terms of the use in civil engineering [43].
The basis of most chemical methods is the release of molecular ammonia from ammonium salts solutions with an alkaline pH. Such as in the STI method, where a mixture of water and alkaline compounds is introduced into the ash [44]. Similarly, the ASM Technology assumes the introduction of calcium hypochlorite. The efficiency of these processes is up to 95%. [45]. Simple wet methods are based on the water solubility of ammonium compounds only. According to the studies of Wang et al. [46] up to 85% of ammonia can be removed after 10 min of water extraction.
Another approach is ozone oxidation. This method assumes the use of ozone admixed to humid air, which is supplied to the fly ash as an oxidizing medium. The ozone concentration of 2% was found to be the most effective in ammonia removal at a temperature of 150 • C [47].
Moreover, catalytic methods have been investigated. Their principle is the reaction of selective catalytic oxidation of ammonia according to (6): The best reported efficiency of ammonia conversion is 60% with nitrogen selectivity over 90% [48,49].
All those methods require high investment costs. In thermal methods, a dedicated reactor is needed, while in chemical methods ash becomes wet and requires drying. The undoubted disadvantage of catalytic methods is the necessity of assembling the catalyst and the danger of its degradation by SO 2 and heavy metals. Therefore, there is a strong need to develop a simple and cost-effective method for ammonia removal from fly ash.
The objectives of presented paper are: (1) to examine the parameters of a novel SNCR installation called FJBS (Furnace Jet Boiler System) applied to a stoker boiler for the full range of boiler load: 12-30 MW th , which corresponds to 33-103% of the nominal load. (2) to investigate the effectiveness of a novel, hybrid FJBS+ installation (a combination of FJBS with a catalyst) applied to a stoker boiler, including a comparison of brand new and regenerated plate type TiO 2 -WO 3 -V 2 O 5 catalyst, (3) to determine the fate of NH 3 in combustion by-products (flue gas and fly ash) during the hybrid FJBS+ process, (4) to test an innovative method of ammonium removal from fly ash by high-temperature desorption on boilers' grate.

Materials and Methods
The investigated object is a 29 MW th solid-fuel-fired unit with a mechanically driven grate producing hot water for district heating purposes. The combustion chamber and the second pass are built with membrane wall technology. The flue gas composition at the boiler outlet is as follows: 6-11% O 2 , 0-100 mg/Nm 3 CO, 12-15% CO 2 , 1200-1500 mg/Nm 3 SO 2 and fly ash concentration below 50 mg/Nm 3 . The flue gas flow on maximal boiler load is 59,000 Nm 3 /h. Apart from the primary air under-the-grate distributing zones, the boiler is equipped with secondary air nozzles, that are assembled above the ignition zone. Combustion conditions specific to stoker boilers make the base concentration of NO x to be relatively low. The right grate movement rate and fuel particles size make the flame temperature lower compared to PC boilers. Average NO x emission without FJBS was determined to be 340 mg NO x /Nm 3 at 6% O 2 with an upper value of approximately 380 mg NO x /Nm 3 .
During every test presented in this study the continuous measurement of flue gas composition (O 2 , NO, NH 3 , CO) was performed with infrared FTIR Gasmet DX-4000 (measuring accuracy at 0,1 mg/Nm 3 ) and Siemens Ultramat 23 (measuring accuracy at 1 ppm for NO x , CO and 0,1% for O 2 sensor) online gas analyzers.
Presented Furnace Jet Boiler System (FJBS) is a non-catalytic technology that combines primary and secondary methods for reducing nitrogen oxides emissions. The NO x formation in the combustion chamber is reduced by realizing the objectives of the secondary air nozzles. At the same time, the injection of 40% urea water solution is realized to meet the objectives of selective non-catalytic reduction. The combustion chamber is equipped with jet blowers (Figure 1) whose location was determined based on temperature measurements, numerical calculations [50] and previous experience [51]. The number of jet blowers and their arrangement on the boiler walls can vary depending on the boiler specification. In the investigated case, the combustion chamber is equipped with 8 jet blowers, installed on 4 levels. The FJBS blowers are distributed in two columns on the opposite sidewalls of the boiler. The system includes a vertical inclining mechanism that allows adjusting of the reagent injection. The FJBS technology favors flue gas mixing, which leads to the stabilization of temperature in the combustion zone and the homogenization of flue gas composition. A typical motive medium (MM) of a jet blower is air, but water vapor or recirculated flue gas may be used as well. The high-pressure MM works as OFA (Over Fire Air) nozzles and is simultaneously used to inject the chemical reagent for NO x reduction. In the presented study, a 40% urea water solution was used as a reagent.
The hybrid system FJBS+ was created by the combination of SNCR and SCR methods. Ammonia slip, which is not desirable if the SNCR method is used as a stand-alone method, is the source of reducing compound for the catalytic method. A brand new and regenerated plate type TiO 2 -WO 3 -V 2 O 5 catalysts were installed in the investigated object.
The catalyst was regenerated by two-stage flushing with demineralized water. The catalyst was placed in a stainless-steel tank filled with demineralized water at 50 • C. The tanks were aerated through air nozzles placed in the bottom. The initial stage of the cleaning process took 20 min and was followed by the second stage with the same duration time. After that, the baskets with catalysts were placed on the platform to dry and cool down to ambient temperature. Finally, the catalyst was dried with hot air.
A catalyst may be typically situated in two different locations in the technological structure of a boiler. In the High Dust system, it is placed before the de-dusting device, usually in the second or third boiler pass with the proper thermal conditions. An unquestionable disadvantage of this solution is the direct catalyst exposure to dust (fly ash) contamination resulting in its fast deactivation. The other solution is in the Tail End, where dust-free flue gas enters the catalyst. However, for this solution flue gas has needs reheating to reach the optimum SCR temperature. In presented research catalysts were placed in the space between the upper and lower coil of the second pass water heater (High Dust system) without interfering with the pressure section of the boiler. Both brand-new and regenerated catalysts are equipped with a measuring set which measures emissions directly before and after the reduction process. The overall layout of the FJBS+ technology is presented in Figure 2. Fly ash collected after the processes presented in this study may contain a considerable amount of ammonium compounds, reaching up to 3800 mg/kg. Moreover, denitrification processes may lead to an increased presence of unburned carbon (UBC) in the fly ash. To solve this problem an innovative study of fly ash cleaning method was carried out based on the fact, that high-temperature treatment leads to the decomposition of ammonium compounds. A simple and cost-effective idea of fly ash recirculation back to the furnace was proposed for that purpose. The presented process does not require the assembly of any dedicated reactor or the usage of any chemical compounds nor water.
The principle of the process is presented schematically in Figure 3. The main idea of this process is recirculation of the fly ash into the combustion chamber of the boiler (1) where thermal decomposition of ammonium compounds takes place on the grate (2). Fly ash separated from flue gas in the de-dusting system (4) or/and ash hoppers (3) is transported by the pneumatic conveyor (5) or by the mechanical feeder to the fuel feeder (6) or directly to the fuel tank (7) instead of being removed by the slag hopper (8) or slag removing system (9). After that, fly ash is placed onto the grate together with the fuel. The temperature of above 800 • C leads not only to the decomposition of ammonium compounds but the reduction of unburned carbon as well. At the end of this process, fly ash with slag is removed from the combustion chamber into the slag hopper (8). 1-combustion chamber, 2-grate, 3-ash hoppers, 4-de-dusting system, 5-pneumatic conveyor, 6-fuel feeder, 7-fuel tank, 8-slag hopper, 9-slag removing system. NH 3 content in fly ash samples before and after every desorption test was determined according to VGB-B 401:1998 Blatt 4.4.2 standard. First, to assess the potential of hightemperature ammonia removal, a laboratory test was conducted. Ash samples with a mass of 50 g were placed in an electric furnace heated up to a temperature of 1050 • C for 25 min. Such parameters are a reflection of the typical conditions on a stoker grate. In presented research, the ammonia content in every sample was decreased below the detection limit.
After successful determination of process efficiency in a small scale a field test was carried out. The ash samples were placed in closed boxes made of fine-meshed heatresistant steel mesh to maintain sample homogeneity and allow free flame access and flue gas outflow. Then samples were put into the furnace of the boiler through the inspection door and placed at the beginning of the moving grate. After approx. 25 min of residence time mesh boxes with ash samples were removed from the boiler at the end of the moving grate. NH 3 content in the samples was determined before and after the process. Presented method represents all the advantages of existing thermal methods, but investment costs are significantly reduced. Presented technology does not require the assembly of any external reactor hence the fly ash treatment takes place in the combustion chamber.

Selective Non-Catalytic Reduction FJBS
The first step of the research was to determine the effectiveness of non-catalytic FJBS technology in the wide range of boiler load.
The continuous flue gas composition measurement for 90 h is shown in Figure 4. For boiler load below 14.5 MW th (50% of nominal capacity) the optimal temperature zone is located relatively close to the grate hence the lowest level of jet blowers was used to inject the urea solution. At a moderate boiler capacity of 18-27 MW th (60-90% of nominal), the operation of second and third levels of jet blowers satisfied the emission limits. The most important condition is to provide a satisfactory NO x reduction at nominal boiler load and during overload. In this case, the urea injection was carried out at the highest possible jet blowers level. The average NO x emission without denitrification system depends on the boiler load and for investigated stoker is as follows: 320 mg NO x /Nm 3 for low load, 340 mg NO x /Nm 3 for moderate load and 375 mg NO x /Nm 3 for maximum load, all expressed at 6% O 2 . For the load over 22 MW th (75% of nominal), a significant effect of secondary air was found to support the reduction of NO x . The air delivered from above the ignition arch lowers the flame temperature and moves it away from the boiler front wall, which is desired in terms of stable system operation. During the test, the NO x emission below 180 mg NO x /Nm 3 was continuously observed with an average emission of 169 mg NO x /Nm 3 .
The efficiency of NO x reduction can be calculated as: where NO x is the average NO x emission without FJBS, and NO x FJBS is the average NO x emission with FJBS expressed in mg NO x /Nm 3 at 6% O 2 . For every boiler load, NO x emission was reduced below 180 mg NO x /Nm 3 with NO x reduction of 48-53% as shown in Table 1.

Hybrid SNCR + SCR System FJBS+
After successful determination of FJBS efficiency, the hybrid FJBS+ installation was tested. A single-layer catalyst was placed in the space between the upper and lower water heaters in the second flue gas pass. This location was selected based on the boiler design and on-site temperature measurements taken in the flue gas duct section, the results of which can be seen in Figure 5. The tests were divided into two parts: Part I: low to medium boiler load and Part II: high boiler load.

FJBS+ Part I: Low/Medium Boiler Load
The first part of the hybrid FJBS+ research was conducted in the summer. For most of the time the boiler load was 12.5 MW th and therefore the temperature in the catalyst area was below 300 • C. At a low load, it may be necessary to transfer part of the heat output from the 2nd pass water heater to the one located in the 3rd pass. As a result, the temperature within the catalyst could be more favorable (300-400 • C). To avoid it, the study was conducted during the morning hours when the water consumption by the network user was high and a temporary increase in the boiler output was observed. The efficiency of the catalysts was measured for two boiler loads: 12.5 MW th and 18.5 MW th , which corresponds to 43% and 63% of nominal capacity, respectively. According to in-furnace temperature measurement, both lower and middle nozzles levels were used and the urea solution was injected into a temperature zone of approximately 750 • C. To retain the NH 3 slip needed for the SCR catalyst, various overstoichiometric ratios of urea were used: NSR 1 = 1.35 and NSR 2 = 1.8. The stoichiometric ratio NSR (8) is defined as a ratio of urea actually introduced to its stoichiometric demand: NSR = . m CO(NH2)2 actual . m CO(NH2)2 stoichiometric (8) As a result, the high efficiency of SNCR was observed, together with extra catalytic reduction. This resulted in a low final NH 3 slip determined at the outlet of the catalyst. Figure 6 presents a collective graph from on-stack continuous emission measurement system. The NO x emission was reduced by 50 to 70%, from an average level of 380 mg/Nm 3 down to 156 mg/Nm 3 . The transition stage is a time needed for stabilization of the FJBS+ parameters after changing the stoichiometric ratio from NSR = 1.35 to NSR = 1.8. An increase in CO emissions is observed as NSR is increased, with a direct correlation to the urea decomposition process. Figures 7 and 8 present the effectiveness of a brand new (CN) and regenerated (CR) catalyst at the boiler load of 18.5 MW th . The concentrations of NO x and NH 3 were determined at the inlet and outlet of both catalysts. The brand new catalyst is characterized by an average NO x conversion of 80% while for regenerated one conversion reached 57% ( Table 2). The NH 3 slip at the catalysts outlets was reduced to relatively low levels: below 10 mg/Nm 3 for the new catalyst and below 20 mg/Nm 3 for the regenerated one. It should be emphasized, that the measurement was performed within a boiler operation that does not provide optimal temperature conditions for the SCR process. Thus, results may be considered unexpectedly satisfactory. The effectiveness of the catalytic layer is strongly affected by the distribution of the reducing compound, which may not be optimal in this case. A catalytic reduction can be attractive for units where ammonia slip is a major problem. A single or multi-layer catalyst can work as an effective NH 3 slip catcher.    Part II of the research took place in the winter season with a high boiler output of 28 MW th . In this case, a suitable temperature for effective NO x reduction was provided. However, because of the high temperature in the furnace, it was not easy to produce sufficient ammonia slip. I this case, the FJBS nozzles located on the highest, 4th level were used and a high stoichiometric excess of urea NSR = 3 was used to ensure the presence of NH 3 in the flue gas exiting the combustion chamber. The overall emission measurement is shown in Figure 9. The concentration of nitrogen oxide at the catalyst inlet ranged from 160 to 220 mg/Nm 3 . The use of catalyst allowed local reduction of NO x emission to 30 mg/Nm 3 with stack emission ranging from 150 to 180 mg/Nm 3 in most points. The FJBS+ was switched off after 3 h of operation, which can be seen in Figure 9 as an instant peak in NO x emission.
The direct effect of the catalysts can be seen in Figures 10 and 11. The efficiency of the new and regenerated catalyst was 63% and 50%, respectively (Table 3). This indicates the lower SCR efficiency compared to Part I. The temperature range was considered optimal; however, it should be emphasized that the concentration of NH 3 in the flue gas at the catalyst inlet was much lower than expected, ranging from 5 mg/Nm 3 for CN to 6 mg/Nm 3 for CR. Catalyst outlet emission for NH 3 resulted in at least halving the inlet value.    The measurement methodology did not allow the simultaneous emission measurement on the inlet and outlet of the catalyst. The measurement was conducted firstly as the catalyst inlet and then at the outlet. The characteristics of the FTIR measurements, the length of the measurement pipe and the response time affected the offset final NH 3 values. This can be observed at Figure 11, showing the efficiency of the regenerated catalyst, Figure 11.

The Fate of NH 3 in Combustion By-Products
For the FJBS+ process, an ammonia balance was determined taking into account mass fractions of urea and ammonia. As an input, a dose of 40% urea water solution was considered. The output was ammonia slip in fly ash, ammonia slip in flue gas and urea involved in stoichiometric NO x reduction. Balance incorrectness δ was defined as a variance between total urea input and output.
As a urea input, a dose of urea was calculated taking into account the mass flow and composition of urea solution used as a reagent (Table 4).
According to the reaction of urea decomposition Hence 1 kg of NH 3 is a result of the decomposition of 1.765 kg of CO(NH 2 ) 2 . Thus, ammonia stream flow can be recalculated out of the urea stream. Ammonia output as a sum of ammonia in ash, flue gas slip and urea stream for NO x reduction is presented in Table 5.  Balance incorrectness δ was determined as a variance between total ammonia input and output: Finally, the mass balance of the FJBS+ process is summarized in Table 6 and presented in Figure 12.
As can be seen, 64.35% of ammonium compounds used in the denitrification process take part in stoichiometric NO x reduction, 28.55% migrates into fly ash and 5.52% appears in flue gas as ammonia slip. The total value of ammonia output is almost equal to ammonia input. This difference is defined as balance incorrectness and for considered process parameters equals 1.58%. Some reasons for balance inaccuracy can be defined, such as the possible formation of ammonia-rich ash deposits inside the boiler ducts and migration of ammonia compounds into bottom ash.

Reduction of Ammonia in Fly Ash
Fly ash collected after the process was tested due to the potential of ammonia removal in both laboratory and full scale. The results are presented in Table 7. Three samples of fly ash were chosen to be tested in terms of diversified ammonia contents: 479, 1326 and 3796 mg NH 3 /kg. As a result of the laboratory test, the ammonia content in every sample was decreased below 6.1 mg NH 3 /kg (below detection limit). After successful determination of process efficiency in the laboratory scale a full-scale (field) test was carried out. Ash samples were placed onto a grate in steel-mesh boxes. Figure 13 presents samples before and after thermal decomposition on a grate. For the field test, ammonia content was significantly reduced as well to a level up to 32.3 mg/kg. Presented demonstration proves the potential for industrial use of this innovative and cost-effective method.

Conclusions
The full-scale investigation of non-catalytic (FJBS) and hybrid SNCR + SCR (FJBS+) denitrification processes was successfully performed in a stoker boiler for the full range of power output. In every case, a stable NO x reduction was achieved. The NO x emission was reduced down to a level below 180 mg NO x /Nm 3 at 6% O 2 with NO x reduction rate reaching 53%. The FJBS installation was combined with the TiO 2 -WO 3 -V 2 O 5 catalyst to create the hybrid SNCR + SCR system FJBS+. This solution reduces problems with proper injection and mixing of reducing compound and flue gas before entering the catalyst. The catalyst application in high dust area resulted in achieving the reduction efficiency of up to 80% for new catalyst and 57% for regenerated catalyst. The result can be considered very satisfactory, taking into account that only one catalyst layer was used and for the low/medium boiler load the temperature in the catalyst section cannot be considered optimal. The catalytic reduction can solve problems with elevated ammonia slip since the catalyst works as an effective NH 3 slip catcher. The use of regenerated catalyst allows reducing operating costs, hazardous waste disposal and environmental danger. Stokers are characterized by a small amount of fly ash compared to pulverized coal boilers, thus a long service life of the catalyst is expected. Presented technology can be applied in waste-to-energy units, including the combustion of medical and hazardous waste.
As follows from a balance 64.35% of ammonium compounds used in the denitrification process take part in stoichiometric NOx reduction, 28.55% migrates into fly ash and 5.52% appears in flue gas as ammonia slip. Ammonia content in fly ash is vital in terms of its possible use in civil engineering, thus an innovative and cost-effective method of ammonia removal from fly ash was presented and tested. Fly ash was recirculated into the combustion chamber, where ammonium compounds were thermally decomposed in a temperature of above 800 • C. The process was initially tested in the laboratory conditions and after that, the successful results were obtained in the full scale. In most cases the NH 3 content in fly ash was reduced below the detection limit of 6.1 mg/kg. In contrast with other methods, presented method does not require an external reactor nor any chemical compounds, which makes it competitive and prospective.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the fact, that presented technology is under commercial R&D process.