Developing On-Road NOx Emission Factors for Euro 6b Light-Duty Diesel Trucks in Korean Driving Conditions

: This study aimed to develop on-road NOx emission factors for Euro 6b light-duty diesel trucks (LDDTs) in Korea. On-road NOx emissions were measured using portable emissions measurement systems and compared with those measured using the Korean Driving Cycle (KDC), the conventional laboratory test used to develop emission factors. To ensure the representativeness of the LDDTs emission factors, ﬁve vehicles of three models were driven along two real driving routes for total traveled mileage of 2280 km. On-road NOx levels were 2.1 to 6.9 times higher on average than those measured using the KDC because the latter does not cover the wide variability in vehicle speed and relative positive acceleration, common in real driving conditions. The lean-NOx trap was found to have disappointingly low NOx reduction efﬁciency in on-road driving. The on-road NOx emission factors by vehicle speeds developed in this study were comparable to the COPERT 4 factors. The results of this study will be suggested for use in the Korean CAPSS inventory. The conclusion of this study will contribute to estimate realistic NOx emission amount from Euro 6b LDDTs in Korea and the developed NOx emission factors can be used in other countries having applied Euro emission regulations because the results of this study were validated with COPERT 4. On the other hand, COPERT 4 emission factors based on the on-road measurement could be backed up with this study. Furthermore, it can provide a justiﬁcation of the emission control measures for light-duty trucks such as the early introduction of electric trucks with subsidies by the Korean Ministry of Environment.


Introduction
The Korean government established the Clean Air Policy Support System (CAPSS) to estimate the emission inventories of air pollutants and implement a range of emission control measures for different sources of pollution. Nitrogen oxides (NOx), which are a class of major atmospheric pollutants, act as precursors in the photochemical formation of tropospheric ozone [1] and contribute to the secondary formation of PM 2.5 .
Road transport is known as the major source of NOx emissions; it accounts for 53.7% of the total NOx emissions in the Seoul metropolitan area according to CAPSS statistics for 2016 [2]. Among the emission sources of road transport, diesel trucks account for 53.0% of the NOx emissions; the light-duty diesel trucks (LDDTs) that carry a one-ton load, in particular, account for 69.5% of the registered diesel trucks [3] and 30.2% of the NOx emissions and are thus particularly significant sources that need to be controlled.
The government has been tightening vehicle emissions standards to reduce the amount of air pollution generated by road transport and has adopted the Euro regulations set by the European Union on LDDTs. LDDTs running in Korea are defined as N1-class vehicles according to the categories in the European regulation [4], and vehicles certified after September 2015 are to meet the Euro 6b emission standards [5].
Several studies that measured gaseous emissions in real driving conditions with portable emissions measurement system (PEMS) found that there were substantial differences between NOx emissions measured with diesel vehicles driven in the laboratory and on-road conditions. In on-road driving, Euro 6b light-duty vehicles were shown to have NOx levels considerably higher than the laboratory emission limit (0.08 g/km) based on the New European Driving Cycle (NEDC). For European diesel vehicles, Degraeuwe and Weiss [6] showed the median of on-road NOx for seven Euro 4-6 diesel vehicles exceeded by 2.7 times those measured on the NEDC and the analysis of PEMS data from 39 Euro 6b diesel passenger cars by O'Driscoll et al. [7] exhibited wide variability in NOx emissions Images and the key specifications of five LDDTs used as test vehicles in this study represent three vehicle models and are listed in Table 1 and Figure 1. factors that determine which test method is effective for reflecting Korean driving conditions, and the driving parameters were analyzed to find what factors cause differences in NOx emissions measured in the laboratory and real driving conditions. The on-road NOx emission factors were developed with 1 km-section data analysis and validated with EU COPERT 4 factors to suggest for use in the Korean CAPSS inventories.

Test Vehicles
Images and the key specifications of five LDDTs used as test vehicles in this study represent three vehicle models and are listed in Table 1 and Figure 1.  In Korea, LDDTs are commonly used to transport freight of one ton or less in large cities. The three models accounted for 88% of the light-duty truck market in Korea in 2018 [20]. Each has an engine volume of 2.2 L or 2.5 L. Both the manual and automatic transmission variants of the models having the largest and second-largest market share were tested. All the vehicles were certified to the Euro 6b emissions requirements, which has an acceptable limit of 0.125 g/km for NOx emissions when measured using the NEDC.
Test vehicles were equipped with a NOx emission control system using EGR and LNT technologies. EGR, which has long been used in diesel engines, works by recirculating a portion of the exhaust gas from an engine back to the combustion chamber, thereby lowering temperatures therein and consequently reducing the quantity of NOx generated in the combustion chamber. LNT is an after-treatment technology based on the concept of trapping NOx emissions as nitrates during lean operating conditions and then, reducing the trapped nitrates to N2 and releasing it by regenerating with rich fuel control. LNT technology has seen general widespread use in Euro 6b diesel vehicles. In Korea, LDDTs are commonly used to transport freight of one ton or less in large cities. The three models accounted for 88% of the light-duty truck market in Korea in 2018 [20]. Each has an engine volume of 2.2 L or 2.5 L. Both the manual and automatic transmission variants of the models having the largest and second-largest market share were tested. All the vehicles were certified to the Euro 6b emissions requirements, which has an acceptable limit of 0.125 g/km for NOx emissions when measured using the NEDC.
Test vehicles were equipped with a NOx emission control system using EGR and LNT technologies. EGR, which has long been used in diesel engines, works by recirculating a portion of the exhaust gas from an engine back to the combustion chamber, thereby lowering temperatures therein and consequently reducing the quantity of NOx generated in the combustion chamber. LNT is an after-treatment technology based on the concept of trapping NOx emissions as nitrates during lean operating conditions and then, reducing the trapped nitrates to N 2 and releasing it by regenerating with rich fuel control. LNT technology has seen general widespread use in Euro 6b diesel vehicles.

Laboratory Tests
For this study, the laboratory measurements of emissions were conducted in the chassis dynamometer lab at the National Institute of Environment Research (NIER). To examine the validity of the laboratory test and develop the emission factors with the conventional method, tests were run using the NEDC and the WLTC, which are emission certification test cycles and the KDC, as shown in Figure 2.

Laboratory Tests
For this study, the laboratory measurements of emissions were conducted in the chassis dynamometer lab at the National Institute of Environment Research (NIER). To examine the validity of the laboratory test and develop the emission factors with the conventional method, tests were run using the NEDC and the WLTC, which are emission certification test cycles and the KDC, as shown in Figure 2. The NEDC used to test emissions while certifying Euro 6b vehicles consists of four low-speed urban driving cycles and one high-speed extra-urban driving cycle specialized for emission tests and characterized by the constant acceleration. In Korea, vehicle certification tests have applied the WLTC procedure instead of the NEDC since September 2018. The WLTC is an international test cycle developed by the World Forum for the harmonization of vehicle regulations based on on-road driving data collected in Europe, the United States, Japan, Korea, and India; it is known for having driving dynamics that are closer to real-world driving patterns than those of the NEDC [21]. Test cycles used to develop emission factors for vehicles in Korea have been established by the NIER and consists of 15 driving cycles based on averaged vehicle speed [22]. The KDC is a combination of eight sub-driving cycles from a 15-driving cycle set and is assessed as a test mode that is highly representative of the driving patterns for Korean light-duty vehicles [18]. The NEDC used to test emissions while certifying Euro 6b vehicles consists of four lowspeed urban driving cycles and one high-speed extra-urban driving cycle specialized for emission tests and characterized by the constant acceleration. In Korea, vehicle certification tests have applied the WLTC procedure instead of the NEDC since September 2018. The WLTC is an international test cycle developed by the World Forum for the harmonization of vehicle regulations based on on-road driving data collected in Europe, the United States, Japan, Korea, and India; it is known for having driving dynamics that are closer to realworld driving patterns than those of the NEDC [21]. Test cycles used to develop emission factors for vehicles in Korea have been established by the NIER and consists of 15 driving cycles based on averaged vehicle speed [22]. The KDC is a combination of eight sub-driving cycles from a 15-driving cycle set and is assessed as a test mode that is highly representative of the driving patterns for Korean light-duty vehicles [18].
The road-load coefficients that are required to be input to a chassis dynamometer were established for NEDC pursuant to United Nations Regulation No. 83 [23]. It is beyond the scope of this study to determine whether the test vehicles are compliant with the emission requirements; the same road-load coefficients were used in all the laboratory tests in which the NOx emissions characteristics were evaluated according to the driving profiles described. The equipment used to measure emissions in the laboratory tests conducted as part of this study meets the requirements specified in UN Regulation No. 83 [23]; the AVL (Graz, Austria) 48" Chassis Dynamometer roller was used and exhaust gases were sampled with a constant volume sampler and measured using the Horiba (Kyoto, Japan) MEXA-7200H and 7100D systems.

Real Driving Emissions Tests
On-road emission tests based on the European Union (EU) "Real Driving Emissions of Light-Duty Vehicles (RDE-LDV)" [15] were conducted in Seoul and on its outskirts, on test routes R1 and R2, as shown in Figure 3.
The road-load coefficients that are required to be input to a chassis dynamometer were established for NEDC pursuant to United Nations Regulation No. 83 [23]. It is beyond the scope of this study to determine whether the test vehicles are compliant with the emission requirements; the same road-load coefficients were used in all the laboratory tests in which the NOx emissions characteristics were evaluated according to the driving profiles described.
The equipment used to measure emissions in the laboratory tests conducted as part of this study meets the requirements specified in UN Regulation No. 83 [23]; the AVL (Graz, Austria) 48" Chassis Dynamometer roller was used and exhaust gases were sampled with a constant volume sampler and measured using the Horiba (Kyoto, Japan) MEXA-7200H and 7100D systems.

Real Driving Emissions Tests
On-road emission tests based on the European Union (EU) "Real Driving Emissions of Light-Duty Vehicles (RDE-LDV)" [15] were conducted in Seoul and on its outskirts, on test routes R1 and R2, as shown in Figure 3.  The type of road was divided by the vehicle speed, as prescribed by the EU RDE-LDV-less than 60 km/h in the urban section, 60 km/h to 90 km/h in the rural section, and 90 km/h or more on highways. This classification is in line with the speed limits based on road types in Korea. The tests were run a total of 30 times and 2280 km mileages. Test data with DPF regeneration was included in developing emission factors. The driving characteristics of the on-road test routes were compared with those of the driving cycles in laboratory tests ( Table 2).  The type of road was divided by the vehicle speed, as prescribed by the EU RDE-LDV-less than 60 km/h in the urban section, 60 km/h to 90 km/h in the rural section, and 90 km/h or more on highways. This classification is in line with the speed limits based on road types in Korea. The tests were run a total of 30 times and 2280 km mileages. Test data with DPF regeneration was included in developing emission factors. The driving characteristics of the on-road test routes were compared with those of the driving cycles in laboratory tests (Table 2). On-road emissions were measured by the second using PEMS. As shown in Figure 4, the PEMS (SEMTECH-LDV, SENSORS, Saline, MI, US) consists of a gas analyzer, flow meter, power supply, control unit, monitoring device, GPS, thermometer, and hygrometer. The exhaust flow meter with a pitot tube is used to translate concentration into mass. The vehicle speed is measured by the GPS, and the thermometer and hygrometer are used to measure ambient temperature and calculate correction factors, respectively. Electric power is supplied by the four 12 V batteries connected to the vehicles in parallel. The On-road emissions were measured by the second using PEMS. As shown in Figure 4, the PEMS (SEMTECH-LDV, SENSORS, Saline, MI, US) consists of a gas analyzer, flow meter, power supply, control unit, monitoring device, GPS, thermometer, and hygrometer. The exhaust flow meter with a pitot tube is used to translate concentration into mass. The vehicle speed is measured by the GPS, and the thermometer and hygrometer are used to measure ambient temperature and calculate correction factors, respectively. Electric power is supplied by the four 12 V batteries connected to the vehicles in parallel. The performance of the PEMS was verified with laboratory equipment in driving WLTC before the on-road tests.

Data Analysis
The NOx results in laboratory tests were calculated by applying the average concentration of the gas sampled in the dilution bag through the constant volume sampler system in each phase of the driving cycles. The emission factors for Korean vehicles were developed by analyzing the pollutant emissions by the average vehicle speed for each sub-cycle of the KDC.
In the on-road tests, the emission results were calculated based on the concentration and flow rate of the exhaust gas measured in seconds. The average values of the on-road test in this study were calculated with the moving averaging window (MAW) method prescribed in the EU RDE-LDV [15]. When developing emission factors using the on-road test results, the data manager for emission rates and activities of road transport (DEAR)a 1 km-section analysis method suggested by Lee et al. [18]-was applied. DEAR segments the pollutant emissions measured per second in 1 km-sections and analyzes the average of the vehicle speeds and emissions for each section. The averaged values of the individual 1 km-sections were binned to the unit of average speed at 10 km/h to analyze the emissions characteristics according to the average vehicle speed. Figure 5 shows the average NOx emissions for each test vehicle in the laboratory and RDE tests. The NEDC result for the V2 vehicle was not valid due to the setting of laboratory equipment. All five test vehicles showed higher NOx emissions in the latter than in the former. In the NEDC, NOx emissions were found to meet or slightly exceed the allowable limit of 0.125 g/km under Euro 6b standard.

Data Analysis
The NOx results in laboratory tests were calculated by applying the average concentration of the gas sampled in the dilution bag through the constant volume sampler system in each phase of the driving cycles. The emission factors for Korean vehicles were developed by analyzing the pollutant emissions by the average vehicle speed for each sub-cycle of the KDC.
In the on-road tests, the emission results were calculated based on the concentration and flow rate of the exhaust gas measured in seconds. The average values of the on-road test in this study were calculated with the moving averaging window (MAW) method prescribed in the EU RDE-LDV [15]. When developing emission factors using the on-road test results, the data manager for emission rates and activities of road transport (DEAR)-a 1 km-section analysis method suggested by Lee et al. [18]-was applied. DEAR segments the pollutant emissions measured per second in 1 km-sections and analyzes the average of the vehicle speeds and emissions for each section. The averaged values of the individual 1 km-sections were binned to the unit of average speed at 10 km/h to analyze the emissions characteristics according to the average vehicle speed. Figure 5 shows the average NOx emissions for each test vehicle in the laboratory and RDE tests. The NEDC result for the V2 vehicle was not valid due to the setting of laboratory equipment. All five test vehicles showed higher NOx emissions in the latter than in the former. In the NEDC, NOx emissions were found to meet or slightly exceed the allowable limit of 0.125 g/km under Euro 6b standard.

Comparisons of NOx Emissions from Real Driving Emissions (RDE) and Laboratory Tests
The average NOx emissions from the test vehicles were 0.17-0.55 g/km under the KDC and 0.13-1.0 g/km under the WLTC. In the on-road measurements, the average NOx emissions were 0.55-1.83 g/km, which are 4.4-14.6 times higher than the allowable limits under the NEDC, with substantial variance among the test vehicles.
The difference in the laboratory and RDE NOx emissions results in this study is similar to those from other studies of on-road NOx emissions of Euro 6b-certified lightduty diesel vehicles. These previous studies explained the excessive NOx emissions in RDEs by noting that emission control technologies applied to Euro 6b LDDTs, including EGR and LNT, are not sufficient to offset NOx increases caused by external factors such as ambient temperature, steep roads, and acceleration. The same reasons seem to apply to the The average NOx emissions from the test vehicles were 0.17-0.55 g/km under the KDC and 0.13-1.0 g/km under the WLTC. In the on-road measurements, the average NOx emissions were 0.55-1.83 g/km, which are 4.4-14.6 times higher than the allowable limits under the NEDC, with substantial variance among the test vehicles.
The difference in the laboratory and RDE NOx emissions results in this study is similar to those from other studies of on-road NOx emissions of Euro 6b-certified light-duty diesel vehicles. These previous studies explained the excessive NOx emissions in RDEs by noting that emission control technologies applied to Euro 6b LDDTs, including EGR and LNT, are not sufficient to offset NOx increases caused by external factors such as ambient temperature, steep roads, and acceleration. The same reasons seem to apply to the results of this study, where test vehicles showed a significant increase in NOx emissions in real driving conditions compared to other laboratory test conditions.
Given that on-road NOx emissions are 2.1-6.9 times higher than those measured under the KDC, it is necessary to develop NOx emission factors based on the RDE results to reflect realistic NOx emissions in the CAPSS inventory.

Comparisons of NOx Emissions from RDE and Laboratory Tests Based on Average Vehicle Speed
The Korean CAPSS system has estimated the emission inventories of vehicles using emission factors according to the average vehicle speed based on the results of KDC emission tests. Figure 6 shows the results of the RDE tests as a box plot after the average NOx emissions of a 1 km-section were binned by average vehicle speeds. The results of the KDC test are represented by the average NOx emissions from the eight sub-cycles by average vehicle speed. In all sections, the amount of on-road NOx emissions are significantly higher than those from the KDC test.
Compared to results of KDC, the average NOx emissions in the RDE test were 2.8 times higher in the urban section at an average speed of 50 km/h or less, 4.9 times higher in the rural section at an average speed of 60-80 km/h, and 3.4 times higher on motorways at an average speed of 90 km/h. The NOx emissions per 1 km unit tended to increase at low and high speeds under the KDC, while the results of on-road tests were high in the overall average vehicle speed. Although the KDC is considered to reflect the typical driving patterns in Korea, the results of this study show that it has limitations in terms of how it represents the NOx emissions of Euro 6b LDDTs in real driving conditions. Given that on-road NOx emissions are 2.1-6.9 times higher than those measured under the KDC, it is necessary to develop NOx emission factors based on the RDE results to reflect realistic NOx emissions in the CAPSS inventory.

Comparisons of NOx Emissions from RDE and Laboratory Tests Based on Average Vehicle Speed
The Korean CAPSS system has estimated the emission inventories of vehicles using emission factors according to the average vehicle speed based on the results of KDC emission tests. Figure 6 shows the results of the RDE tests as a box plot after the average NOx emissions of a 1 km-section were binned by average vehicle speeds. The results of the KDC test are represented by the average NOx emissions from the eight sub-cycles by average vehicle speed. In all sections, the amount of on-road NOx emissions are significantly higher than those from the KDC test. Compared to results of KDC, the average NOx emissions in the RDE test were 2.8 times higher in the urban section at an average speed of 50 km/h or less, 4.9 times higher in the rural section at an average speed of 60-80 km/h, and 3.4 times higher on motorways at an average speed of 90 km/h. The NOx emissions per 1 km unit tended to increase at low and high speeds under the KDC, while the results of on-road tests were high in the overall average vehicle speed. Although the KDC is considered to reflect the typical driving patterns in Korea, the results of this study show that it has limitations in terms of how it represents the NOx emissions of Euro 6b LDDTs in real driving conditions.

Influence of Driving Dynamics
To examine the difference in NOx emissions generated from the RDE and KDC, we used the test results of vehicle V4. Figures 7 and 8 show the NOx emissions based on vehicle speed and relative positive acceleration (RPA) for the RDE and KDC, respectively. Figure 7 shows that the higher the RPA, the higher the average NOx emissions are when vehicle speeds are similar. The NOx emissions vary according to the average vehicle speed as reported in previous studies [7,9]. A high RPA increases the load on the engine during acceleration. According to the Netherlands Vehicle Authority [24], the EGR technology employed in diesel engines decreases the rate of operation to protect the engine as a greater load is placed on it. A comparison of the average vehicle speed-RPA distributions shows that RDE tests have higher RPAs than those of the KDC. It seems clear the KDC has limitations in terms of representing real driving dynamics although it is considered to reflect average driving patterns in Korea.   Figure 8, however, shows that the NOx emissions under the KDC are lower than those of the RDE tests at the similar average vehicle speed and RPA. This can be explained via the NOx emissions of Euro 6b LDDTs being influenced by more than just driving dynamics. Degraeuwe et al. [6] showed that the NOx emissions of the RDE test were still higher than the NEDC even when RDE data filtered with vehicle speed, acceleration, CO2 emissions, and ambient temperature was compared with the NEDC data having similar driving dynamics. This suggests that the difference in NOx emissions is due to the differ-   [7,9]. A high RPA increases the load on the engine during acceleration. According to the Netherlands Vehicle Authority [24], the EGR technology employed in diesel engines decreases the rate of operation to protect the engine as a greater load is placed on it. A comparison of the average vehicle speed-RPA distributions shows that RDE tests have higher RPAs than those of the KDC. It seems clear the KDC has limitations in terms of representing real driving dynamics although it is considered to reflect average driving patterns in Korea. Figure 8, however, shows that the NOx emissions under the KDC are lower than those of the RDE tests at the similar average vehicle speed and RPA. This can be explained via the NOx emissions of Euro 6b LDDTs being influenced by more than just driving dynamics. Degraeuwe et al. [6] showed that the NOx emissions of the RDE test were still higher than the NEDC even when RDE data filtered with vehicle speed, acceleration, CO 2 emissions, and ambient temperature was compared with the NEDC data having similar driving dynamics. This suggests that the difference in NOx emissions is due to the different driving dynamics and different control strategies adopted by manufacturers for laboratory and real driving conditions. The results of this study for the NOx emissions of Euro 6b LDDTs can also be explained by the fact that the performance of the NOx emission control technologies would not be maintained in real driving conditions.

Influence of NOx Emission Control System
EGR and LNT technologies used in the test vehicles are generally applied to Euro 6b diesel engines. However, some studies [25,26] have confirmed that EGR systems do not function well in driving conditions apart from the NEDC. In this study, we found that there are also significant differences in NOx reduction performance of LNT under real driving conditions and NEDC. Figure 9 shows the LNT operational characteristics for V4, with modal NOx and CO emissions under both the RDE and NEDC tests. LNT technology reduces NOx emissions via a process referred to as "de-NOx regeneration," in which NOx is stored under lean conditions, general operating conditions of diesel engines, and then reduced via control of a temporary stoichiometric or rich air-fuel ratio. According to Matsumoto et al. [27], CO concentrations peak during the regeneration period due to the incomplete fuel combustion, whereas NOx concentrations reach the peak value due to the NOx slip. Under the NEDC, a low NOx emission level is achieved when both the storage of NOx and de-NOx regeneration function properly. Under real driving conditions, peaks in CO concentration are observed several times, implying that the storage and de-NOx regeneration occur with LNT operation. The time between the moments when the CO levels peak is when the LNT catalysts store NOx but the NOx emission rate is substantially higher in real driving conditions than they are during the NEDC. Even during the de-NOx generation of LNT in real driving, the NOx emissions increase because its reduction performance is lower than that of NEDC.
Vehicle manufacturers control NOx emissions using this technology by modeling the amount of NOx that the LNT can store and the amount of fuel required to reduce the stored NOx. LNT emission control technology works quite well under simple driving conditions, such as those of the NEDC; however, its function is limited under real driving conditions when a variety of driving parameters affect NOx emissions. Better NOx emissions can be achieved if some peaks of NOx are mitigated with an improved calibration strategy during transient maneuvers. Vehicle manufacturers control NOx emissions using this technology by modeling the amount of NOx that the LNT can store and the amount of fuel required to reduce the stored NOx. LNT emission control technology works quite well under simple driving conditions, such as those of the NEDC; however, its function is limited under real driving conditions when a variety of driving parameters affect NOx emissions. Better NOx emissions can be achieved if some peaks of NOx are mitigated with an improved calibration strategy during transient maneuvers.

Validation of on-Road NOx Emission Factors
As confirmed by the results of the tests carried out in this study, the NOx emissions of Euro 6b LDDTs are influenced by various driving parameters and EGR and LNT technologies have a limited ability to control NOx emissions under a wide range of on-road driving conditions. Therefore, it is necessary to consider a number of environmental factors that affect NOx emissions in real driving conditions-various driving patterns, road load, ambient temperature, and road steepness-to develop realistic NOx emissions for Euro 6b LDDTs. Because it is difficult to reflect the various driving parameters in averaged or normalized laboratory driving cycles, it would be more effective to develop emission factors using data measured in RDE tests. Figure 10 compares the Euro 6b LDDT emission factors developed in this study, based on a DEAR analysis of the on-road NOx emissions, with those from European studies [28]. The NOx emission factors developed for LDDTs based on RDE results show a remarkable difference compared to those with laboratory KDC tests. The on-road NOx emission factors for Euro 6b LDDTs developed in this study were similar to those in COPERT 4 [29]. While NOx emission factors in COPERT 4 show higher levels at both low and high speeds than those developed in this study, the latter displays a smaller difference in NOx emissions in the overall average vehicle speed.

Validation of On-Road NOx Emission Factors
As confirmed by the results of the tests carried out in this study, the NOx emissions of Euro 6b LDDTs are influenced by various driving parameters and EGR and LNT technologies have a limited ability to control NOx emissions under a wide range of on-road driving conditions. Therefore, it is necessary to consider a number of environmental factors that affect NOx emissions in real driving conditions-various driving patterns, road load, ambient temperature, and road steepness-to develop realistic NOx emissions for Euro 6b LDDTs. Because it is difficult to reflect the various driving parameters in averaged or normalized laboratory driving cycles, it would be more effective to develop emission factors using data measured in RDE tests. Figure 10 compares the Euro 6b LDDT emission factors developed in this study, based on a DEAR analysis of the on-road NOx emissions, with those from European studies [28]. The NOx emission factors developed for LDDTs based on RDE results show a remarkable difference compared to those with laboratory KDC tests. The on-road NOx emission factors for Euro 6b LDDTs developed in this study were similar to those in COPERT 4 [29]. While NOx emission factors in COPERT 4 show higher levels at both low and high speeds than those developed in this study, the latter displays a smaller difference in NOx emissions in the overall average vehicle speed.
Therefore, it appears to be more reasonable to use emission factors developed from the data measured in RDE tests, compared to laboratory measurements, if the NOx emission characteristics of Korean Euro 6b LDDTs are to be reflected realistically. Therefore, it appears to be more reasonable to use emission factors developed from the data measured in RDE tests, compared to laboratory measurements, if the NOx emission characteristics of Korean Euro 6b LDDTs are to be reflected realistically.

Conclusions
We measured NOx emissions for five test vehicles using both laboratory test procedures and real driving conditions to develop the NOx emission factors for Euro 6b LDDTs in Korea. In this study, although the test data in DPF regeneration was included to develop NOx emission factors, the increased NOx emissions by DPF regeneration were not accurately quantified. It is difficult to quantify the increased NOx emissions in this event because the control of DPF regeneration would be affected by the accumulated soot amount in the filter, additive fuel injection that is controlled by the engine control unit depending on various driving conditions, durations of DPF regeneration, and other driving parameters. Although it is known that NOx emissions from Euro 6b diesel vehicles are deviated by ambient temperature, in this study, the seasonal effects were not estimated to develop NOx emission factors. Despite these limitations, this study suggests more realistic NOx emission factors of Euro 6b LDDTs in the Korean driving conditions by selecting representative models as test vehicles and conducting on-road emission tests in Seoul, which has the biggest traffic volume in Korea. The conclusions of this study could be summarized as follows: •

Conclusions
We measured NOx emissions for five test vehicles using both laboratory test procedures and real driving conditions to develop the NOx emission factors for Euro 6b LDDTs in Korea. In this study, although the test data in DPF regeneration was included to develop NOx emission factors, the increased NOx emissions by DPF regeneration were not accurately quantified. It is difficult to quantify the increased NOx emissions in this event because the control of DPF regeneration would be affected by the accumulated soot amount in the filter, additive fuel injection that is controlled by the engine control unit depending on various driving conditions, durations of DPF regeneration, and other driving parameters. Although it is known that NOx emissions from Euro 6b diesel vehicles are deviated by ambient temperature, in this study, the seasonal effects were not estimated to develop NOx emission factors. Despite these limitations, this study suggests more realistic NOx emission factors of Euro 6b LDDTs in the Korean driving conditions by selecting representative models as test vehicles and conducting on-road emission tests in Seoul, which has the biggest traffic volume in Korea. The conclusions of this study could be summarized as follows:

•
The average on-road NOx emissions of the Euro 6b LDDTs were 4.4-14.6 times higher than those measured with the NEDC, an emission certification test cycle. The NOx levels were also 2. The results of this study will be suggested for use in the Korean CAPSS inventory. The conclusion of this study will contribute to estimate realistic NOx emission amount from Euro 6b LDDTs in Korea and the developed NOx emission factors can be used in other countries having applied Euro emission regulations because the results of this study were validated with COPERT 4. On the other hand, COPERT 4 emission factors based on the on-road measurement could be backed up with this study. Furthermore, it can provide a justification of the emission control measures for light-duty trucks such as the early introduction of electric trucks with subsidies by the Korean Ministry of Environment.