                       UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                      NATIONAL VEHICLE AND FUEL EMISSIONS LABORATORY
                                   2000 TRAVERWOOD ROAD
                                   ANN ARBOR, MI  48105




                                                                                                                 	      			                     
May 4, 2012
                                                         OFFICE OF 
                                                      AIR AND RADIATION

MEMORANDUM

SUBJECT:	NOX Emissions from DPF Regeneration

FROM:	Lauren Steele, Office of Transportation and Air Quality

--------------------------------------------------------------------------------
TO:		Docket EPA-HQ-OAR-2011-1032

 Estimates of Emissions of Oxides of Nitrogen from Regulated Emergency Vehicles 
      The Heavy-Duty Highway Rule was projected to reduce oxides of nitrogen (NOX) exhaust emissions from the heavy-duty on-road fleet by 1.76 million tons in 2020, and by 2.48 million tons in 2030.  In that rule's final Regulatory Impact Analysis, EPA projected that in 2010 the controlled heavy-duty fleet would emit approximately 2 million tons of NOX emissions, decreasing to about 600,000 tons in 2020 and about 300,000 tons in 2030, as a result of that rule.  For purposes of this memo, we are considering that rule's control case as the current fleet baseline scenario.
      As noted in Section VII.B of the proposed rule preamble, class 3 through 8 fire apparatus and ambulances represent about 0.8 percent of annual new heavy-duty truck registrations.  If we assume these vehicles have the same emissions profile over time as other vehicles, we could estimate that their annual emissions collectively may also represent a comparable fraction of the fleet emissions in a given year.  For the years projected in the RIA for the Heavy-Duty Highway Rule, this would indicate emergency vehicles may have emitted on the order of 16,000 tons of NOX in 2010, may emit about 5,000 tons in 2020, and may emit about 2,000 tons in 2030.
      However, we have information that these vehicles have longer than average useful lives, leading to a slower fleet turnover rate.  We also have information that these vehicles have below average vehicle miles traveled, with a large fraction of their engine hours at low load or idle.  While these two factors would affect the emissions inventory in opposite directions, they are unlikely to cancel each other out.  Because we do not have enough data to quantify how these factors affect the emission estimates, we will use the values presented above for purposes of the current analysis.
      On a per-truck basis, we can estimate what the NOX exhaust emissions from a certified engine would be if it emitted NOX at the level of the standard, 0.2 gram per brake-horsepower-hour (g/bhp-hr). As an illustration, we have estimated what that might represent in terms of emissions from a truck operating at 200 horsepower over a typical year with 1,200 operating hours. This would be representative of a 500-hp pumper truck with an average engine operating load of 40 percent of capacity.  This illustration would yield a hypothetical fire truck that would emit about 100 pounds of NOX exhaust emissions in a year, or 0.05 ton per year.
      These estimates are provided as a basis for comparison, as the following sections explore some different scenarios.
 Estimates of NOX Emissions from Regeneration 
      The estimates of NOX emissions for certified engines generally include emissions during periods when the diesel particulate filter (DPF) is regenerating.  During both active automatic and parked manual regenerations, emission rates may increase for some pollutants, especially NOX when NOX-reducing aftertreatment devices are not present downstream of the DPF. The higher than normal temperatures during active regeneration with high rates of oxidation occurring across the catalyst create conditions conducive to NOX formation.
      Because of the potential effect on emissions, engine manufacturers are allowed to provide adjustment factors that account for the excess emissions during DPF regeneration, as part of the certification process. During a series of certification tests, the High NOX Emission Factor (EFHNOX) represents the NOX emission rate during an active regeneration, and the DPF regeneration frequency represents the fraction of tests during which an automatic active regeneration occurs. Passive regenerations would not be expected to increase NOX emissions, thus they are not counted in the DPF regeneration frequency for testing purposes.  Figure 1 illustrates the relationship between the EFHNOX and DPF regeneration frequency, based on EPA certification data for MY 2008 engines over the transient test cycle.
                                       
Figure 1: High NOX Emission Factor and DPF Regeneration Frequency
      As noted above, excess emissions from DPF regenerations are generally accounted for as part of the engine certification process. Thus, during DPF soot accumulation mode the NOX emissions may be lower than the standard for some engines, and we see a general inverse relationship between DPF regeneration frequency and the EFHNOX. We can also observe from Figure 1 that the NOX emission rate during a regeneration may be many times higher than the standard. 
      The amount of excess NOX emissions due to DPF regeneration would generally be expected to be small on a vehicle whose engine is equipped with downstream NOX emissions control, such as selective catalytic reduction. EPA's certification data for MY 2011 engines indicate the EFHNOX are generally below 1.0 g/bhp-hr, with some outliers. DPF regeneration frequencies for MY 2011 engines do not differ substantially from those of MY 2008 engines. For more information on DPF regeneration frequencies, see the docket memo entitled Fuel Use with Dosing for DPF Regeneration, submitted to the docket for this rulemaking.
 Estimates of Excess NOX Emissions from Emergency Vehicles with More Frequent Regenerations and Higher Emission Rates
      As noted in the memorandum to the docket on fuel dosing, it is possible that some manufacturers will submit applications for AECD's or field modifications for engines on emergency vehicles with liberalized parameters under which automatic active and/or parked manual regenerations may occur, potentially increasing the NOX emissions from those vehicles.  It is also possible that manufacturers may submit requests for engine recalibrations that enable more frequent passive regeneration and rely less on active regeneration, potentially changing the NOX emissions in more complex ways.  It is thus difficult to predict the effects this rule may have on the NOX emissions from emergency vehicles.  
      As described in the preamble to the proposed rule, EPA has proposed that engines with approved emergency vehicle AECD's would not need to include these excess emissions during certification. In other words, the engines would be tested with emergency vehicle AECD settings inactive. For purposes of this illustration, we assume the regeneration NOX in the base case is included in the certified emission level, and excess regeneration NOX due to the AECD would be above the certified emission level.
      Without speculating what fraction of the emergency vehicle fleet may receive AECD's or field modifications with NOX recalibrations, for this illustration we will focus on a per-truck emissions estimate.  For this hypothetical truck, we are assuming that the current certified configuration (the base case) includes a DPF regeneration frequency (FREQDPF) of 5 percent, a high NOX emission rate of 2.5 g/bhp/hr (EFHNOX during regeneration), and a low NOX emission rate of 0.2 g/bhp-hr (EFLNOX when not regenerating). This example is purely hypothetical  -  EPA is not aware of an engine with these specific emissions characteristics.  The alternate cases explore the emissions effects when the DPF regeneration frequency (FREQDPF) increases to 20 percent, and when the EFHNOX increases to 5.0 g/bhp-hr, while holding other parameters the same.  
      The values presented in Table 1 have been calculated using the following formula for the average NOX emission rate (EFA):
                EFA = (FREQDPF)(EFHNOX) + (1- FREQDPF)(EFLNOX)

      Where
      FREQDPF = Base 5% or Alternate 20%
      EFHNOX = Base 2.5 g/bhp-hr or Alternate 5.0 g/bhp-hr
      EFLNOX =  0.2 g/bhp-hr in all cases
      To extrapolate the average emission rate (EFA) to an annual value, we are continuing with the example of a truck with a 500-hp engine operating at an average load of 40%, for a working power of 200 hp and average of 1,200 engine operating hours in a year.
Table 1:  Per-Truck NOX Emission Estimates by Regeneration Percent
Scenario
                                      EFA
                                Pounds per Year
                                 Tons Per Year
Base Case
                                     0.32
                                      170
                                     0.08
Base FREQDPF
Increased EFHNOX to 5g/bhp-hr
                                     0.44
                                      230
                                     0.12
Increased FREQDPF to 20 Percent
Base EFHNOX
                                     0.66
                                      350
                                     0.17
Increased FREQDPF to 20 Percent
Increased EFHNOX to 5g/bhp-hr
                                     1.16
                                      610
                                     0.31
      As shown in the table, if the both regeneration frequency and the high NOX emission rate were to increase as described, the annual NOX emissions from this truck could increase fourfold.  We believe this is a highly conservative estimate. As shown in Figure 1, these parameters have an inverse relationship, so it is highly unlikely that both parameters would be maximized in the same engine configuration.
      We present below in Table 2 the emissions estimates using an alternate calculation method that is based on the regeneration interval. Similar to the method used in the docket memo entitled PM Emissions Impacts, we have estimated in the base case that a single regeneration event has a duration of 30 minutes, and the interval between regenerations is 25 engine hours.  In the alternate cases we explore the emissions effects if the regeneration interval is reduced to 12 or 8 engine hours, and each case uses the same EFHNOX and EFLNOX as in the previous illustration.
      
Table 2:  Per-Truck NOX Emission Estimates by Regeneration Interval
                             Regeneration Interval
                          Operating Time Regenerating
                               EFHNOX (g/bhp-hr)
                              NOX Emissions (TPY)
                                                                       25 hours
                                                                           2.0%
                                                                            2.5
                                                                           0.06
                                                                       25 hours
                                                                           2.0%
                                                                            5.0
                                                                           0.08
                                                                       12 hours
                                                                           4.0%
                                                                            2.5
                                                                           0.08
                                                                       12 hours
                                                                           4.0%
                                                                            5.0
                                                                           0.10
                                                                        8 hours
                                                                           5.9%
                                                                            2.5
                                                                           0.09
                                                                        8 hours
                                                                           5.9%
                                                                            5.0
                                                                           0.13
      The values in Table 2 indicate that if regenerations were to occur at 8 engine-hour intervals instead of 25 engine-hour intervals, with the EFHNOX emission rate doubling from 2.5 to 5 g/bhp-hr, then the annual NOX emissions from this hypothetical fire truck could double. 
      All of the above examples assume that no additional NOX control occurs downstream of the DPF.  However, the majority of the heavy-duty fleet of MY 2010 and later is equipped with selective catalytic reduction NOX aftertreatment systems. When active, this technology reduces NOX emissions dramatically, including any excess NOX emissions due to DPF regeneration.
