REDUCTION OF NITRIC OXIDE FROM DOMESTIC HEATERS
E. Baraldi, G. Berthold*
University of Modena, via G. Campi, 213/B 41100 Modena Italy
*Worgas Bruciatori, Via Coppi 17, 41043 Formigine, Italy
ABSTRACT
This paper reports combustion measurements on a laboratory prototype of an unvented space heater. In particular, the influence of the air / fuel ratio, the exhaust gas temperature, the gas type and gas sampling position on the formation of NO and NO2 is studied. The Butterfly wing shaped flame technology having an extended bladed combustion front, characterised by low average temperatures is applied to the unvented heater technology. Emission measurements carried out on a prototype heater show that the NO2 / NOx ratios are up to 90%. Measurements according the ANSI Z21.11.2 1996 standard (American National Gas Institute, 1996) resulted in NO2 =6.6 ppm for natural gas and in 9.8 ppm in the case of propane gas.
Keywords: unvented heaters, hyperstoichiometric flame, reduction NO thermal, NO2 reduction
INTRODUCTION
Indoor air pollution is receiving substantial attention as a public health concern because of the high percentage of time that people spend indoors. In many cases, the concentrations of pollutants in indoor air are higher than those found outdoors due to the plethora of sources associated with human activities and building operations. The anthropical sources of nitrogen oxides are represented by all combustion reactions for fossil fuels. Two of these principle oxides are generated in a gas flame during high temperature combustion and are usually referred to as NOx, with x = 1 or 2. They are mostly formed as nitrous oxide (x= 1, NO 95 %) and the rest nitric oxide NO2. Nitric oxide is an important air pollutant due to its potential effect on human health and its role in atmospheric chemical processes leading to ozone formation and the production of atmospheric acids. Undesirable effects are never described when normal NO concentrations are found in the environment, so from a health point of view, the most important compound is NO2, because it is about four times more toxic than NO. Various studies (Beard, 1982; Gilli et al., 1988; Saltzman, 1990; Admur, 1991 Uccelli e Mauro, 1992 and 1997) suggest that exposure to NO2 could lead to adverse health effects in humans impairing respiratory defence mechanisms and increases susceptibility to infection.
Therefore studies have been carried out to establish a correlation between toxic effects and the levels of the pollutants which have induced them. On the basis of these considerations the World Health Organisation (1987) has elaborated the air quality guidelines to provide an information base to defend public health from harmful effects of pollution, or to reduce those pollutants which are probably dangerous for health and the well being of human beings.
For a number of years several countries, like for instance the USA, Japan and Australia have been concerned about the adverse health effects of low level NO2 exposure to human being (CPSC, 1983; Porter and Rogers, 1988; Babich, 1990) caused by direct contact with combustion gases which are a significant source of indoor air pollutants such as nitric oxide, including emission from home heaters, unvented gas space heaters and kerosene heaters; gas stoves and gas ranges used in catering appliances (Stolwijk et al., 1983; Leaderer et al., 1986; Young et al., 1987; Koontz and Nadga, 1988; Spicer and Billick, 1995). To prevent any negative health effects caused by NO2 it is recommended not to exceed levels over 0.3 ppm per 8 hours, which correspond to an NO2 emission factor of 20 ppm air free in the exhaust gas (ANSI Z21.11.2, 1996).
The NO formed in natural gas and propane flame comes from two sources (Reuther, 1995) «prompt» NO, which originates «promptly» in the combustion processes as a result of the reaction of hydrocarbon radicals with atmospheric nitrogen, and «thermal» NO formed by the high temperature reaction of nitrogen and oxygen. Also it is generally accepted that prompt is weakly dependent upon temperature and dominates at primary aeration levels <100% while thermal NO has a strong positive temperature dependence (the hotter the flame, the more thermal NO) dominates at primary aeration levels > 100 %. However, the rate of the NO formation depends on the flame temperature, oxygen concentration and the residence time of the mixture in the combustion zone.
Whereas NO is formed early in a flame it is well known that under certain conditions a high proportion of NO is converted to NO2 in exhaust gases from combustion systems such as gas turbines (Johnson and Smith, 1978) and unvented space heaters (Bromly et al., 1984; Traynor et al., 1985).
The amount of NO2 found in combustion products depends on the complex interactions of the production and destruction mechanisms of all of the species involved (NO, NO2, HO2, O2, O, etc.) as well as on the flue temperature and residence time. Extensive studies of NO2 formation in combustion systems (Cernansky and Sawyer, 1975; Johnson et al., 1979, Hargreaves et al., 1981; Sano, 1985; Hori et al., 1992; Amano and Hase, 1994; Reuter and Billik, 1995) and in gas-sampling probes (Allen, 1975; Cernasky, 1977; Hori, 1980) regarding the temperature history, like cooling rate, residence time and high exhaust gas temperature have suggested that i) NO2 is formed during the cooling of hot exhaust gases and it increases with increasing cooling rate, ii) the conversion of NO to NO2 is promoted by thermal quenching and by trace hydrocarbons (HCs), iii) NO is converted to NO2 mainly through the oxidation reaction NO + HO2 NO2 + OH that occurs in the cooler regions downstream of the flame, denominated post-flame zone, where the temperature is below 1000°C.
This paper emphasises how the NO2 formation in unvented heaters is influenced by means of the air / fuel ratio, sampling position, temperature and gas type. The present situation of emission is also shown for products that are currently available on the market. The study is carried out by using a double slot Worgas burner whereas at the end of the article an unvented heater is presented with very low NO and NO2 emissions.
EPERIMENTAL SET-UP
According to the American Standard (Z21.11.2, 1996) to measure the exhaust gas products the heater is placed below a hood assembly as shown in Figure 1. The height of the heater coincidences with the bottom line of the hood. As shown in the insert on the left of Figure 1 the exhaust gases on their way downstream become highly diluted resulting in typical oxygen concentration of roughly 20.2 % at the probe position P4. Such a high and rapid dilution of the flue gases with cold ambient air is a typical characteristic for unvented heaters and serves mainly to cool down hot flue gases to prevent hazards caused by they. Assuming that in a typical heater the exhaust gas velocity is approximately 1 m/s an averaged cooling rate of roughly 10 K/ms results.
Figure 1: At the right schematically cross section of the experimental set-up is shown. The unvented heater is equipped with a rectangular twin slot burner. To collect the exhaust gases a hood is installed above the heater. Flue gas samples are taken at different measurement position indicated with P1 through P4. The insert at the left shows the temperature at the different probe positions as a function of the local oxygen content.
The burner used in the experiments is characterised by means of a rectangular twin slot flame pattern. Such a slot pattern allows the creation of stable flames over a wide range of the air/fuel ratio. The primary air p and gas flow rates are measured and controlled by means of mass flow meters.
The combustion gas measurements have been carried out at different positions P1, P2, P3 and P4, as indicated in Figure 1, using a chemiluminescent NO / NOx analyser (Rosemount Analytical Inc. – Model 955), a non-dispersive infrared CO / CO2 analyser (Rosemount Analytical Inc. – Model 880) and paramagnetic O2 analyser (SERVOMEX – Series 1100). The temperatures were measured by Cr / Al type K thermocouples. Thus the temperatures and the concentration of pollutants in the gas sample can be measured without substantially disturbing the combustion. For the sake of comparing the results under different conditions all the concentrations have been measured are conveniently been normalised to their air free values. All the measurements of the combustion products are either carried out for natural gas as well as for propane.
FLAMES SHAPES AND STABILITY AT DIFFERENT AIR / FUEL RATIOS
Figure 2 shows several flame shapes. Flame stability can be achieved depending on the equivalence ratio and mixture flow rate, four different main flame configurations can be generated: in fuel rich mixtures, depending on the flow rate, either yellow or blue flame Bunsen type or M-shaped (double Bunsen) flames are present, as shown in Figure 2.
Figure 2: Different flame shapes are shown as a function of the primary air p at constant input rate resulting from a burner characterised by means of a rectangular twin slot pattern. Changing the primary air concentration creates the different flames. At extreme rich conditions a yellow flame is formed, while at primary air levels of around 80 % of the stoichiometric value two Bunsen type flames are created. For lean fuel mixtures a V-shaped flame is formed.
In lean fuel conditions either V-shaped or M-shaped flames are formed. The various flame typologies present great differences not only from the point of view of flame front morphology, but also concerning the main combustion features. As it has been observed in previous works (Coghe et al. 1995), under lean fuel conditions, the mechanism of flame stabilisation, transition between the two different flame shapes, and pollutant emission reduction are mainly connected to the fluid dynamic features of the flame, particularly to the value of air flow rate and to the formation of vortex structures outside the V-flame front. This wing-shaped flame is attached over the burner and is extremely stable, owing to fluid dynamic effects in a wide range of operating conditions and gives rise to low and uniform temperature all over its area.
In the following a systematic analysis of the combustion products and in particular the formation of NO2 is carried out for the case of an unvented heater. The input rate was fixed in all experiments to 4.5 kW; natural gas as well as propane gas was used respectively at injector pressures of 10 mbar and 20 mbar, respectively.
NO2 FORMATION IN TYPICAL UNVENTED HEATERS
The emissions of NOx, NO and NO2, respectively, for a typical unvented heater without the heat exchanger are shown in Figure 3 as a function of the primary air p. The flue gas samples are taken in position P4 downstream from the mixing chamber.
Figure 3: The NOx, NO and NO2 emissions are shown as a function of the primary air p. The data are taken for natural gas with an heater input rate of 4.5 kW at the position P4.
For levels of 0<p<0.5 the NOx and the NO decrease with increasing primary air. In this range a yellow flame is formed and changes to a blue flame with increasing primary air. At p = 0.5 a local minimum in the NOx and NO concentration is observed. A «soft» blue Bunsen type flame, where the inner and outer cones are not distinguishable and the NO formation is subjected mainly to the prompt mechanism, accompanies this minimum. Further increasing of the primary air in the burner up to p = 0.7 results in increasing NOx and NO, while at even higher values a strong decrease in the NOx and NO formation is observed. This behaviour reflects the well-known correlation with the flame temperature as a function of the primary air levels and the formation of NO is explained by means of the Zeldovich mechanism.
The NO2 concentration, also illustrated in Figure 3, shows a less pronounced dependency over the wide range of primary air. In the regime of the Bunsen type flames (0<p<1.2) the NO2 is almost constant, while above, in the regime of the inverted flames, a strong decrease down to 2.5 ppm with increasing primary air is observed. This behaviour is accompanied by a dramatic increase in the NO2 / NOx ratio up to values of about 90 %. Note that as a comparison in a typical boiler application the NO2 / NOx ratio is normally lower than 5 %.
EXHAUST GAS TEMPERATURE
To study the effect of the exhaust gas temperature on the NO2 / NOx ratio a heat exchanger was constructed to fit inside the unvented heater as shown in Figure 1. By changing the amount of the water flux in the heat exchanger the temperature measured above the heat exchanger in position P1 could be tuned from T1 = 150 °C up to T1 = 480 °C. The O2 concentration directly above the heat exchanger was measured at 12.9 % and did not significantly change with temperature.
By increasing the outlet flue gas temperature T1 a strong decrease in the NO2 / NOx ratio is observed as shown in Figure 4. The NO2 / NOx ratio decreases from 14 % to 3 % with increasing T1 from 150 °C to 450 °C in the flue gas outlet.
Figure 4: It is shown the NO2 / NOx ratio as a function of the exhaust gas temperature at position P1. The data are taken for propane gas at p = 0.8 with a heater input rate of 4.5 kW.
As shown by Amano et al. (1994) during the cooling process from flame temperature down to the flue gas outlet temperature the amount of NO transformed into NO2 is mainly determined through the balance of the following reaction:
HO2 + NO NO2 + OH (1)
NO2 + H NO + OH(2)
Whereas both reactions are in balance until the temperature reaches nearly 1000 °C, below this temperature an effective formation of NO2 is present, because is favoured the reaction (1); while at higher temperatures a decomposition of the NO2 occurs because the reaction (2) is favoured. However, the reaction chemistry in the case of combustion in an unvented heater is not under equilibrium conditions, therefore the time evolution of the combustion gases is of importance. Numerical simulation carried out by Amano et al. (1994) has shown that the cooling rate becomes important in explaining the NO2 formed by oxidation of NO. In the case of our heater, cooling rates are estimated as follows: the flame temperature as the upper limit is around (1200 1800) °C; while the lower temperature limit between 150 °C and 470 °C is given at position P1 above the heat exchanger, as shown in Figure 4. Therefore, the extreme temperature drop is between 1650 K to 730 K and a typical residence time of the flue gases before passing the heat exchanger is approximately 0.1 s. This results in cooling rates in the range of 7 to 16 K/ms and according theoretical prediction to a conversion ratio of NO into NO2 of 5 % to 10 %, which is in excellent agreement with the experimental data shown in Figure 4.
POST HEATER ZONE
In the following the NO2 / NOx ratio is studied in the zone between the exhaust gas outlet, position P1, and the probe position P4 in the hood. For the following experiments propane gas is used and the heater without the heat exchanger is operated at 4.5 kW with p=0.8.
Figure 5 shows the total NOx (air free) subdivided in its two components NO and NO2, at the position P1 to P4. Within the accuracy of the measurement the total NOx can be considered constant at all different positions while the ratio NO2 / NOx increases downstream. The increase most pronounced is in the area where the exhaust gases get in contact with the surrounding cold ambient air, as shown in Figure 5 for the measurements at position 1 with NO2 = 14 ppm and at position 2 with NO2 = 25 ppm. Further downstream the NO2 is still increasing to 32 ppm at position P3 which is at the top of the heater, while above no further changes in the ratio is observed.
Figure 5: The NO and NO2 components of the total NOx are shown at the different probe positions P1 to P4. The data are taken for propane gas at p = 0.8 with a heater input rate of 4.5 kW.
Due to the dilution of the exhaust gases with the surrounding cold air the flue gas temperature is decreasing downstream as it is shown in the insert of Figure 1. The largest gradients occur in the vicinity of the exhaust gas outlet of the heater at the position P1 to P3, while between position P3 and P4 no significant change in the oxygen content is noticed and therefore also no temperature change observed. However, the average temperature gradients from position P1 to P2 is estimated approximately to 1 K/ms, which at maximum results in a transformation of less than 1 %. Further, the distance between the positions P2 and the burner surface is roughly 400 mm downstream the flue gases, which is believed to have a low concentration of the radicals HO2, OH and H. Thus the change of NO2 concentration from 14 to 25 ppm in this area seems to be improbable by means of the model discussed above. However, at the outlet of the exhaust gas higher cooling rates up to (10 100) K/ms might locally occur, which might explain the observed conversion also at a decreased presence of the radical concentration.
Since it is of importance to determine the effective NO2 indoor room concentration it is important that the measured NO2 at position P4 can be considered as stable for at least the duration of the average air exchange time of a closed living area. The air exchange time is typically in the order of some hours. As shown in the data of Figure 5 the NO2 does not increase between position P3 and P4 and therefore can be considered as stable for at least the air exchange time.
V-SHAPED FLAME FOR UNVENTED HEATER
According to health concern it was mentioned in the introduction part, that NO2 should not be higher than 20ppm normalised to air free. As shown through the paper one effective way to reduce the NO2 emissions can be the use of a high primary air ratio above 1.2. Therefore, the V-shaped flame realised on a twin slot burner has been chosen, whereas in our case the primary air is supported only through natural draft caused by the gas jet. Such burners are often called atmospheric burners and are typically used for unvented heaters.
In Figure 6 and Figure 3 it is shown, respectively, that the NO2 emissions are clearly below the 20 ppm level for p above 1.2. Further, the low NO2 concentration is accompanied by a NOx concentration below 20 ppm, resulting in high NO2 / NOx above 50 % up to 90 %. Such high ratios at very low total NOx have been theoretically predicted by Amano et al. (1994). However, even if the remaining percentage of NOx converts to NO2, still the NO2 level remains below the recommended concentration causing negative health effects.