Method estimates NOx from combustion equipment

June 21, 2004
A new estimating technique calculates unit-specific NOx emissions estimates for petroleum-industry combustion equipment that are more accurate than older methods.

A new estimating technique calculates unit-specific NOx emissions estimates for petroleum-industry combustion equipment that are more accurate than older methods.

This simple method accounts for a number of variables that are not covered in the estimating method available in the US Environmental Protection Agency's Emission Estimating Handbook (AP-42). These variables include fuel composition, air preheat temperature, combustion air moisture content, burner intensity, percent of design load, and NOx control technologies.

This improved calculation tool allows users more accurately and cost-effectively to predict NOx emissions. In doing so, they may avoid costly monitoring methods.

Background

Engineers typically estimate NOx emissions from fired heaters and boilers in the petroleum industry by extrapolating the simple emission factors provided in the EPA AP-42 section for industrial boilers.

AP-42 provides NOx emission factors for industrial boilers that burn natural gas, distillates, and residual fuel oil; it recommends extrapolating these boiler emission factors to estimate emissions from other types of combustion equipment.

Petroleum-industry combustion equipment, such as refinery process heaters, however, generally burns refinery fuel gases that contain significant quantities of hydrogen. This can produce NOx emissions that are quite different from the amount generated from a natural-gas-fired industrial boiler.

The difference in emissions can result specifically from the flame temperature and time in the peak temperature zone in a fired heater. The factors specific to industry equipment that can be corrected for different fuels and operating conditions are therefore preferable to the extrapolated factors used for industrial boilers.

Knowing how NOx forms helps in understanding which factors are important. There are three types of NOx: thermal, fuel, and prompt.

Thermal NOx is the thermal fixation of molecular nitrogen. Its formation is generally a function of flame temperature, residence time, and oxygen level in the flame zone. Factors affecting these variables, especially flame temperature (fuel content, air preheat, humidity, etc.), will impact NOx formation.

Fuel NOx is the direct oxidation of nitrogen-containing species in the fuel and is directly related to the fuel's nitrogen content.

Prompt NOx forms from the reaction with N2, O2, and hydrocarbon fragments. The mechanism depends on flame conditions, especially the formation of fuel-rich zones. Prompt NOx, however, especially at concentrations greater than 20-30 ppm (vol), is a relatively minor contribution to total NOx.

The effect of NOx-reduction technologies on NOx emissions can be complicated. Some technologies are applied at the early stages of the combustion process and have an immediate impact on reducing flame temperature or local oxygen concentrations, thus reducing the thermal NOx emissions.

Other NOx-reduction technologies are applied downstream of the combustion process, after combustion is complete. These technologies reduce equally well the thermal and fuel nitrogen NOx emissions.

The various NOx-reduction technologies, therefore, should be taken into account at different places in the calculation procedure.

There have been some indications that control technologies for the combustion process, such as low-NOx burners, can reduce fuel nitrogen conversion; but there are currently insufficient data to substantiate a significant reduction. Our new method, therefore, cannot account for this potential effect.

Our emission-estimating method considers the variables that influence thermal and fuel NOx formation, including fuel content, air preheat, humidity, and equipment. The method also accounts for NOx post-combustion control technologies, such as selective catalytic reduction (SCR), selective non-catalytic reduction (SNCR), etc.

By comparison, AP-42 provides factors based on fuel burned (natural gas, fuel oil, LNG, etc.) and size of combustion unit. AP-42, for example, provides a NOx factor for boilers using natural gas, with little correction for any of these variables.

Lower factors are given for low-NOx burners and flue-gas recirculation. AP-42, however, does not account for variations in fuel composition, especially nitrogen and H2 content, which can affect NOx.

AP-42 has no methodology to adjust for other factors that can affect NOx, such as humidity and air preheat. Also, whereas AP-42 recognizes that several control technologies exist for NOx control, such as SCR or SNCR, there is no recommended method for adjustments due to these technologies.

The emission-estimating method that we developed for refinery combustion units is similar in approach to AP-42 but includes additional parameters.

It involves separately determining the thermal and fuel nitrogen NOx contributions and then adding these together to obtain the overall emission estimate in the absence of any NOx-reduction technologies. The effect of NOx-reduction technologies on the emission estimate is then calculated to determine total NOx.

Click here to enlarge image

Fig. 1 compares predicted NOx emission factors using our new method with actual NOx emission factors based on measurements in several fired-heater stacks. Fig. 1 also compares emission factors provided in AP-42 with the measurement-based emission factors.

The heaters in this case were all fitted with a similar type of gas-only fired burner—a natural draft, conventional, raw-gas burner. Multiple data points for each heater at different operating conditions are included.

Fig. 1 shows that the additional factors accounted for in our NOx prediction method improve the emission factor predictions relative to AP-42.

Emission method

The first step in calculating NOx emissions is to determine a base factor.

Differences in the design and configuration of fired heaters and boilers may result in slightly different emission rates. Because there is insufficient information available, however, on which to base separate emission factors, we recommend the same emission base factor for fired heaters and boilers in the petroleum industry.

The thermal and fuel nitrogen components are both particularly important when estimating emissions for fired heaters and boilers in refineries and petrochemical plants.

At these plants, the fired heaters and boilers can burn a variety of fuels. These fuels include natural gas, refinery fuel gas, and liquid fuels. In addition, preheated combustion air is sometimes used to increase the facility's energy utilization.

The different fuel compositions and the air preheat change the flame temperature, the most sensitive thermal NOx determinant. Refinery fuels also often contain ammonia or other nitrogen-containing constituents that contribute to fuel nitrogen NOx.

One can estimate satisfactorily the thermal NOx contribution based on flame temperature. The fuel composition, air preheat, relative humidity, and heat-removal rate determine the flame temperature.

A flame temperature calculation would consider each of these contributing factors and would be acceptable for relating NOx emissions. Due to the difficulty in calculating the effect of heat removal rate, however, the calculated adiabatic flame temperature is usually used as a surrogate.

There are programs to calculate the adiabatic flame temperature; such a calculation is too complex for a simple estimation method, but one should consider it for complex refinery fuel-gas mixtures.

Despite the contribution from the oxygen level available to the flame, fired heaters and boilers usually operate with 15% excess air (3% excess O2) to maximize energy efficiency. The oxygen concentration correction is insignificant at these oxygen levels. The new method does not consider oxygen levels and residence time in the flame zone in the estimating procedures.

Calculation method

The first, and often the most time consuming, step in determining emission factors is to collect the information needed to perform the calculations.

Click here to enlarge image

Table 1 shows the sources of information needed to estimate the emissions. This information is readily available for most units in a well-operated petroleum plant and is generally included in typical frequently generated reports.

One can estimate fuel and thermal NOx with this information. The sum of these is then adjusted for any post-combustion NOx control technologies. This results in a total NOx estimate.

Click here to enlarge image

Fig. 2 shows the estimation procedure.

Calculating thermal NOx

One calculates thermal NOx by adjusting a base emission factor. This base factor (Table 2) accounts for differences in flame temperature due to fuel composition, excluding the effect of hydrogen. It is based on plant operating data that we collected from ExxonMobil heaters.

Click here to enlarge image

To account for a variety of factors that affect NOx, one should adjust this factor according to Equation 1 (see equation box).

Molecular hydrogen has a high adiabatic flame temperature. Hydrogen-rich fuels will therefore have higher flame temperatures and will generate more NOx. Many refinery and petrochemical plant fuels include waste gases, some of which contain hydrogen in large quantities.

For the factors in Table 2, we assumed an H2 content of:

23% for refinery fuel gas.

14.7% for low-btu gas.

0% for methane and fuel oil.

If the H2 content is equal to or less than this assumption, there is no need to adjust the base factor (i.e., fH2 = 1). If the H2 content is higher, however, NOx emissions will increase and must be adjusted. Table 3a shows this factor that is an approximation for the NOx increase resulting from the hydrogen in the fuel.

Click here to enlarge image

The effect from base H2 content is already accounted for in the base thermal NOx emission factors. So only the incremental H2 should be used to calculate the H2 adjustment factor. For example, a refinery fuel gas with 33% H2 has an fH2 of 1.04.

Many control technologies seek to lower the combustion-zone flame temperature. These include low-NOx burners, ultralow-NOx burners, and flue gas recirculation. One should use fctrl (Tables 3b and 3c) to account for these technologies.

As air preheat increases, the flame temperature and NOx emissions increase. This can be significant when the preheat approaches 300-500° F. Table 3d shows the factor used for adjustment for preheat, fpht.

Click here to enlarge image

Water in the combustion zone will cool a flame. Steam injection is commonly used as a control technology for gas turbines. Humid air will also cool a flame.

Table 3e shows adjustment factors for the humidity of combustion air (fH20) as a function of moisture content (lb H2O/lb dry air). Moisture contents are shown on psychometric charts as a function of dry bulb temperature and relative humidity.

The load adjustment factor (Table 3f) reflects the change in intensity in the flame zone when the furnace or boiler load changes. For example, the change in intensity that accompanies a reduced load reflects a smaller heat release under low-load conditions while the heat-removal surface in the heater or boiler is the same.

Load reduction is an easy way to quickly reduce NOx emissions for a short period of time, for example, during an ozone alert. It does, however, come at a high cost: unit underutilization.

Burner intensity, a measure of the heat release per volume of burner space or flame space also affects NOx emissions. High intensity results in a zone of high heat release with corresponding high temperatures and increased NOx production. High heat-intensity burners are usually in boilers where uniform temperature profiles are less important.

Table 3g shows that high-intensity burners can nearly double NOx emissions.

There is no precise measure for burner intensity; however, a quick examination of a unit's design and operating parameters allows a reasonable determination based on some guidelines.

Low-intensity burners are usually natural-draft burners with a burner intensity of 0.150-0.25 MMbtu/hr-cu ft flame volume. High-intensity burners are usually forced-draft burners with a pressure drop of 4-15 in H2O and a burner intensity of 0.3-0.5 MMbtu/hr-cu ft flame volume.

On an enclosure volume basis (firebox volume), high-intensity burners can have heat release rates greater than 70,000 btu/hr-cu ft vs. rates of about 3,000 btu/hr-cu ft for low-intensity burners. The major difference between the two burner types is that the high-intensity burner achieves much more intense mixing of fuel and air.

Lacking other information, the high burner-intensity factor should be used for packaged boilers and pyrolysis furnaces. The normal burner-intensity factor should be used for other fired heaters and boilers.

Calculating fuel NOx

One calculates fuel NOx by first determining the stoichiometric amount of NOx formed, assuming that all the fuel nitrogen forms NOx. The fuel NOx is then adjusted because not all the fuel nitrogen forms NOx.

Some of the fuel nitrogen reacts with hydrocarbon fragments in the flame or with nitrogen oxides to form nitrogen gas, N2. Low levels of fuel nitrogen are completely converted to NOx. Higher nitrogen-containing fuels level off at a conversion of about 30%.

To calculate fuel NOx, the fuel nitrogen content is needed. Equation 2 shows the stoichiometric calculation.

Table 3h gives the fraction of fuel nitrogen converted to NOx for different initial levels of fuel nitrogen content. This data is for staged air burners with liquid fuels. Equation 3 adjusts for fuel NOx.

Calculation of total NOx

One calculates total NOx by summing the fuel and thermal NOx and, if needed, reducing the total to account for any post-combustion control technologies, such as SCR or SNCR (Equation 4).

Table 3i shows values of fpcc for post-combustion control technologies.

Sample calculation

Determine NOx emissions for a 10 MMbtu/hr furnace operating with a fuel gas composition of:

Methane = 25 vol %.

Ethane = 25 vol %.

Hydrogen = 50 vol %.

Ammonia = 0.05 wt %.

Like many real-world situations, there is no easy classification for the fuel used in this problem. The fuel is clearly not natural gas. It is a refinery gas, but rather unique due to its content.

The user must assume refinery fuel gas as a basis and correct for the incremental H2 content, or assume a natural gas basis and correct for added H2 content. The results will be similar; for this example, we used refinery fuel gas as the basis.

First, the given fuel properties and operating conditions are:

  • Fuel H2 = 50%.
  • Air preheat = none.
  • Air humidity = 40%.
  • Average air temperature = 60° F.
  • Moisture content from psychometric chart = 0.0042 lb H2O/lb dry air.
  • Unit load factor = 100%.
  • Firing rate = 10 MMbtu /hr.
  • Control technologies = none.
  • Higher heating value = 25,900 btu/lb.

Using refinery fuel gas as a fuel type, the adjustment factors used in this calculation are:

Fbase (Table 2) = 0.16 lb/MMbtu.

fH2 (interpolated from Table 3a for a 27% increase in fuel H2) = 1.13.

fctrl (Tables 3b and 3c) = 1.

fpht (Table 3d) = 1.

fH2O (interpolated from Table 3e) = 0.88.

fload (Table 3f) = 1.

fburn (Table 3g) = 1.

fN (Table 3h) = 0.87.

Using Equation 1, the thermal NOx emission factor is:

0.16 lb/MMbtu.3.1.13 3 0.88 = 0.16 lb/MMbtu

Next, Equation 2 determines the fuel NOx emission factor:

Stoichiometric NOx = (46 3 0.0005) /(14 3 0.0259) = 0.06 lb/MMbtu

This is adjusted using Equation 3:

Fuel NOx factor = 0.06 lb/MMbtu.3.0.87 = 0.05 lb/MMbtu

Next, the overall NOx emission factor is the sum of thermal and fuel NOx:

Overall NOx emission factor = 0.16 lb/MMbtu + 0.05 lb/MMbtu = 0.21 lb/MMbtu

Finally, determine the NOx mass emissions:

NOx emissions = 10,000 btu/hr 3 0.21 lb NOx/MMbtu = 2.1 lb/hr.

Stationary gas turbines, internal combustion engines

For stationary gas turbines or internal combustion engines, a similar technique can apply; however, using manufacturer or plant test data at typical operating conditions is preferable.

If data at typical operating conditions are lacking, one can use the previously discussed factors to adjust known emission factors for H2 content, humidity, load, etc.

For internal combustion engines, there are control technologies that can be used other than those listed in Table 3i. Table 3j lists these technologies and their associated adjustment factors.

The authors

Theresa J. Takacs (theresa.j.takacs @exxonmobil.com) is an engineering associate with ExxonMobil Research & Engineering Co., Fairfax, Va. She is currently working on CO2 and NOx emissions estimating and control technologies. Before joining ExxonMobil in 1998, she worked for 7 years as an environmental engineer in Oak Ridge, Tenn. At ExxonMobil, Takacs has specialized in wastewater and air-pollution control technologies. She holds a BS in chemical engineering from the University of Cincinnati.

Dennis L. Juedes is a senior engineering associate with ExxonMobil Research & Engineering Co., where he heads a group responsible for providing central engineering support on combustion equipment in refineries. He has more than 30 years' industrial experience related to the design and operation of refinery combustion equipment including burners, fired heaters, boilers, flares, and incinerators. Juedes holds a BS and an MS, both in mechanical engineering, from the University of Wisconsin-Madison.

I.D. Crane retired from ExxonMobil Research & Engineering Co. in 2000 after more than 35 years of developing, designing, and optimizing fired heat transfer and combustion equipment. He was head of the combustion emissions estimating and control group. Crane holds a BS and MS in mechanical engineering from Clarkson University, Potsdam, NY, and is a licensed professional engineer.