Exhaust Emissions from Euro 6 Vehicles in WLTC and RDE—Part 2: Verification by Experimental Measurement

Abstract

The subject of assessing exhaust emissions in real driving conditions has been relevant for a long time. Its introduction into approval tests focused attention on the comparative possibilities of tests performed on a chassis dynamometer and in road conditions. The article is a continuation of research on the possibilities of estimating emissions in the Real Driving Emission test based on emission data from Worldwide harmonized Light Vehicles Test Cycles. The first part discussed the possibility of comparing dynamic parameters in these tests, and the second part discussed the possibility of estimating road exhaust emissions. The work was done in two stages: the first stage involved the use of distance-specific emissions in individual parts of the WLTC test, and the second stage involved the use of exhaust emission rates as datasets divided into intervals defined by vehicle speed and acceleration. Comparative tests were performed for conventional vehicles (gasoline, diesel) and hybrid vehicles. A chassis dynamometer was used to carry out WLTC tests and PEMS equipment was used for the RDE tests. The exhaust gas components that had to be measured in road tests, namely: carbon monoxide, carbon dioxide, nitrogen oxides, and the number of particulate matter, were analyzed. Based on the data collected, parameters such as road emissions and the exhaust emissions rate were determined for each phase of the dynamometer test as well as the road test. Because of this, it was possible to compare the distance-specific exhaust emissions of each vehicle in the two emission tests. The comparison resulted in establishing that it is possible to estimate distance-specific exhaust emissions of conventional and hybrid vehicles in road test conditions, using only the results obtained in the approval test (for selected test phases). The research concluded that it is possible to estimate selected RDE test parameters based on the results obtained in the WLTC test for the tested vehicles.

1. Introduction

Data from the European Automobile Manufacturers Association (ACEA—Association des Constructeurs Européens d’Automobile) [1] and the Polish Automotive Industry Association (PZPM—Polski Związek Przemysłu Motoryzacyjnego) [2] confirmed that in 2021, the most popular power sources for passenger cars were conventional fuels, i.e., gasoline and diesel. Moreover, the total share of passenger cars equipped with an internal combustion engine was approximately 66% (Poland) and 59.6% (Europe). There was also a significant difference (especially in Poland) compared to data from 2020, when the above statistics were 81.5% and 66.4%, respectively. The cited data indicate that hybrid and electric vehicles are rapidly gaining popularity. According to PZPM data [2], in 2021 about 57% more hybrid electric vehicles (HEV) and about 150% more electric or plug-in vehicles were registered in Poland, compared to the year prior. The vehicle electrification trend can be seen throughout Europe, so over the coming years the proportion of vehicles powered by alternative sources may become equal to those with conventional systems.

Analyzing the data [3] obtained over the last 30 years indicates that most sectors of the economy have recorded a significant reduction in the emissions of pollutants. Unfortunately, road transport turned out to be an exception. In 2018, road transport alone was responsible for almost 95% of the emissions from the entire transport sector. Moreover, according to the same source [3], the share of road transport in total exhaust emissions increased from less than 13% (in 1990) to almost 21% (in 2018).

Reducing the negative environmental impact of economic activity has become a priority for the European Commission (EC). Undoubtedly, the fuel quality standards introduced so far, the tightening of exhaust emission limits for passenger cars and trucks, and the introduction of advanced technologies have significantly contributed to reducing air pollution generated by the transport sector. Nevertheless, the resulting values are still not satisfactory. According to data [4] recorded in Poland in 2010–2020, the greatest challenge is reduction of the atmospheric nitrogen oxides, non-methane volatile organic compounds (NMVOC), and particulate matter. The mentioned solutions that were introduced in the meantime resulted in a reduction of these values, but this reduction was not sufficient to reach a satisfactory result.

The current tasks set for the European Commission are largely focused on achieving climate neutrality related to greenhouse gas emissions by the year 2050. In the case of road transport, the need to increase public awareness of the exhaust emissions generated by their daily use of vehicles is also emphasized. Therefore, the priority has become the promotion of low-emission fuels and increasing the market share of low- or zero-emission vehicles. Vehicles powered by combustion engines still play the primary role in the structure of the European Union’s road vehicle fleet. This means that an equally important task is to improve methods for monitoring and controlling exhaust emissions from these vehicles. Focusing solely on low or zero-emission vehicles will not be enough.

2. Research Problem

As various technologies used in the automotive industry continue to advance and be developed, so too increase the requirements for measuring exhaust emissions and taking into account specific phases of engine operation. Eventually, another amendment to the regulations on vehicle exhaust emission control resulted in two independent tests being required to measure the exhaust emissions of a vehicle. The test is carried out both in laboratory conditions (WLTC—Worldwide Harmonized Light Vehicles Test Cycle) [5,6,7,8] and in real vehicle operation on the road (RDE—Real Driving Emissions) [9,10,11]. Thus it became easier to more realistically reflect the actual exhaust emissions that are released from the automotive sector into the atmosphere. However, the current legal status of approval tests is still not without flaws. Improvement of the current framework should include for RDE more dynamic test conditions, a wider temperature range, and limits for different unregulated pollutants.

The first part of the performed research took place on a chassis dynamometer. The work done concerned the measurement of ecological parameters of three vehicles: two vehicles with conventional drives—with a gasoline engine and a diesel engine; the third vehicle was a hybrid vehicle (gasoline engine and electric motor). The ecological assessment of these vehicles on a chassis dynamometer was carried out as outlined by the WLTP procedure.

The second part of the research concerned the measurement of exhaust emissions from the same vehicles in real road driving conditions. For this stage, an appropriate test route was selected that met the RDE test requirements, along with having to meet rigorous guidelines regarding both static and dynamic conditions at the same time.

In the third analytical part of the research, based on the data obtained in the first two parts, the discrepancies between various parameters of the WLTC and RDE tests of the tested vehicles were verified. At this stage of the work, the main thesis of the dissertation was also answered: whether it is possible to use the values of road emissions of selected harmful exhaust compounds obtained in the WLTC test, done on a chassis dynamometer, to determine the distance-specific emissions (referred henceforth as road emissions) of a vehicle in real driving conditions.

3. Literature Review

Initial work on the possibilities of comparing exhaust emissions measurements done on a chassis dynamometer to those in real driving conditions suggested quite large discrepancies between these measurement methods. This was mainly due to the specialized measuring devices used on the chassis dynamometer setup, which boasted high measurement accuracy for the concentration of harmful exhaust gas compounds. At the same time, the measurement of exhaust gas flow rate was quite well reflected by the number of revolutions of the displacement pump in CVS (constant volume sample) systems, which resulted in very small inaccuracies in the assessment of road emissions on the dynamometer station. Measurements of exhaust emissions in real driving conditions using a portable emission measurement system (PEMS) did not correspond to such high accuracy of exhaust components concentration measurements due to the limited size and weight of the measuring device as well as its costs [12,13,14]. The following literature analysis of changes in measurement accuracy over the last 5 years period provided insight into the development of exhaust emission measurement systems in real driving conditions.

The first comparative studies of the New European Driving Cycle (NEDC) and WLTC tests performed by Zachiotis et al. [15] (2020) on vehicles equipped with diesel engines led to the conclusion that in dynamic tests (WLTC) road emissions of particulate matter are 55% higher and road emissions of nitrogen oxides are 11% higher compared to the NEDC test. Research by DiPierro et al. [16] (2019) confirmed an increase in road emissions of CO₂ by 30% for hybrid (gasoline) vehicles, which was the result of a lower share of operating time spent driving using the electric engine. The results of subsequent research (Szalek et al. [17] (2021)) indicated that it was possible to control hybrid vehicles in such a way that they used the electric drive in approximately 70% of driving in urban conditions, and in the entire driving test the use of the electric engine was approximately 40%.

Research results published by Selleri et al. [18,19] (2022), have shown the importance of the State of Charge indicator in performing road tests. Not taking this indicator into account in the test results can lead to a difference of approximately 50% in relation to road emissions of NO_x and 30% for CO. However, for CO₂ the value tripled between SOC = 100% and SoC = 0% for the dynamometer and RDE tests. More recent comparative studies of hybrid electric vehicles (HEV) and plug-in hybrid vehicles (PHEV) presented by Pielecha et al. [20] (2022) indicate small differences in road emissions of harmful compounds between the WLTC and RDE tests. For HEVs, the differences were: CO₂—1%, CO—15%, NO_x—400%, and PN—8%, while for PHEVs the difference was greater only in the case of CO₂ and amounted to 20%.

Varela et al. [21] (2018) compared the exhaust emissions measuring equipment used for tests on a chassis dynamometer with PEMS devices for measuring exhaust emissions in real driving conditions. They showed that the difference in concentration measurement between these devices was 2% for CO₂ and 5% for NO_x. The result of road emissions of these pollutants was 10% for CO₂ and 15% for NO_x. This was largely due to the error in exhaust gas flow measurement, which reached 10% for urban conditions and 5% for high speeds. Giechaskiel et al. [22] (2018) described research carried out in eight laboratories using two emission measurement systems (AVL MOVE and Horiba OBS-ONE) on the same test vehicle (Euro 6 compliant gasoline port-fuel injection 1.4 L passenger car). The differences in the emission concentration measurement results were: for CO₂—2.9–5.5%, for CO the range was 20–25%, for NO_x the range was 23–31% and for PN the range was 29–39%. The average error for road emissions was: ±1% for CO₂, ±5% for CO, ±11% for NO_x, and ±40% for PN.

In comparative studies between WLTC and RDE Suarez-Bertoa et al. [23] (2019) found that emissions for vehicles in RDE tests were 15–20% greater than in the WLTC test. He stated that the emissions in RDE tests for a Euro 6d (gasoline) vehicle were 15–25 mg/km for NO_x, 133–189 mg/km for CO, and 1.3–2 × 10¹⁰ 1/km for PN. The change in dynamic conditions of operation caused the PN emissions to double along with a 20-fold increase in NO_x emissions. Also, similar results for the GDI vehicle were published by Baek et al. [24].

Giechaskiel et al. [25] (2019) researched the cold start of a gasoline engine and its impact on road emission results in driving tests, they indicated differences of the order of 13% depending on whether wet exhaust gas correction factors were applied.

Valverde et al. [26] (2019) concluded that CO₂ road emission results were approximately 33% greater in the RDE test than in the WLTC test for a gasoline-powered vehicle and approximately 41% greater for a diesel vehicle. Giechaskiel et al. [27] (2021) stated, in relation to tests of diesel vehicles, that the accuracy of exhaust emission assessment in WLTC tests was 2.1–3.8% for CO₂, 41–45% for CO, 16–17% for NO_x, and 16–27% for PN. However, during road tests, the variability of results was greater and amounted to 2.8–4.7% for CO₂, 39–57% for CO, 14–17% for NO_x, and 36–38% for PN.

In their research, Simonen et al. [28] (2019) found that active regeneration of the particulate filter had a significant impact on the measured exhaust emissions. In tests conducted using several vehicles, it was found that this phenomenon had an impact on increasing emissions both in WLTC tests and also in RDE tests. Unfortunately, the increase in exhaust emissions was not uniform: in the WLTC test, filter regeneration increased CO₂ emissions by 22–64%, NO_x by 55–774%, and PN by 95–106 times; in the RDE test the increase was CO₂—4–18%, and NO_x—81–673%.

Research conducted by Andrych-Zalewska et al. [29] (2022) has also shown that for Euro 6d-Temp (gasoline) vehicles, the observed exhaust emissions of all flue gas components were increased in the RDE test compared to the WLTC test: CO₂ by 12%, CO by 300%, NO_x by 15% and PN by 100%.

Jaworski et al. [30] (2022) showed that road conditions influence exhaust emissions and fuel consumption in road tests of hybrid vehicles. It was found that the various driving tests did not produce results similar to those obtained during the RDE test. The uniqueness of the road tests resulted in CO₂ emissions in the range of 103–135 g/km, CO in the range of 1.8–15.3 mg/km, and NO_x in the range of 43–196 mg/km.

Mamala et al. [31] (2023) showed that fuel consumption and energy demand of vehicles depend largely on atmospheric factors and vehicle driving conditions. The impact of vehicles operating in different climatic zones resulted in significant variations of energy consumption (not correlated with the data from the WLTC test)—just maintaining a constant speed of 140 km/h increased the energy demand by 70% compared to the average value from the WLTC test.

Research by Giechaskiel et al. [32] (2021) showed that changing the conditions in which RDE tests were performed (within the permissible parameter limits as outlined by the regulations) has led to significant differences (over 50%) in the results of measured harmful exhaust emissions. It has also been shown that ambient temperature had a significant impact on exhaust emission results, as the cold start was seen to cause over 75% of CO and HC emissions and 45–90% of NO_x (emitted within the first 5 min of the test).

Grzelak et al. [33] (2021) indicated that the use of various calculation methods may result in approximately 7% underestimation of road CO₂ emission results when testing LPG-powered vehicles. This was also confirmed by Gis et al. [34] (2021), indicating that the interchangeable treatment of indicators (dilution ratio and dilution factor) in mathematical calculations of exhaust emissions affected the measurement error of CO₂ concentration by 0.4%, CO by a maximum of 4%, and NO_x by 1%.

Song et al. [35] (2022) compared the dynamic conditions of ULEV vehicles in the RDE test in their research. They determined the motion resistances present in the test and compared them with the dynamic test parameters: vehicle speed per positive acceleration (v⋅a_pos) and relative positive acceleration (RPA). Through tests, they obtained a correlation of 0.52–0.78 between carbon dioxide emission rate and vehicle speed per acceleration. The discrepancies were mainly due to the vehicle losing energy (mainly through dynamometers) amounting to 57%, 18%, and 7% in the urban, rural, and motorway sections, respectively. Thomas et al. [36] (2022) compared the traffic conditions of the WLTC and RDE tests (for hybrid vehicles). However, they did not seek to correlate them, but only to demonstrate the differences between them. The comparison concerned, among others: the number of starts of the hybrid vehicle’s combustion engine in individual parts of the WLTC and RDE tests.

Paper [37] (2023) proved that it was possible to use fractal theory to describe driving tests and, on that basis, determine their similarity indicators. Such indicators could be used to assess exhaust emissions of both conventional and hybrid vehicles. At the same time, the authors emphasized that in order to achieve appropriate accuracy this step must be preceded by quite detailed research.

Developing a new exhaust emissions measurement method presents a significant challenge. Not only the main assumptions of the test must be met, but, as the publications presented above have shown, an additional and detailed analysis of each parameter is necessary. This means that such methods must not only be flexible but even to some extent anticipate and lead the decisions of legislators. An example of such a situation was for authors of [15] (2020), who drew attention to a gap in the approval procedure, which was the lack of taking into account wind speed and direction. Their proposal was to use the Monte Carlo method, which allowed the assessment of the impact of wind on vehicle performance and exhaust emissions. The authors determined the correlations between the change in fuel consumption and exhaust emissions with the wind speed and direction in individual phases of the WLTC cycle. The obtained results showed that in the case of the high and very high-speed phases, the increase in exhaust emissions was much greater, amounting to 28% for CO₂, 22% for NO, and 13% for PN in phase 4 of the WLTC test (the very high-speed phase).

Claßen et al. [38] (2021) made a prediction that future emission standards, that will replace the EU6d standard, will force a change in the RDE tests. This may be done by extending the permissible dynamic parameters of the RDE test, including: by increasing the weighed value of the urban part of the RDE test. The introduction of limit values for nitrogen dioxide, nitrous oxide, ammonia, and formaldehyde is also under consideration. Additionally, a reduction in the permitted number of particles (for diameters > 10 nm: PN10), methane emissions, and emissions of non-methane organic gases was predicted. To simplify vehicle approval, there are discussions about abandoning dynamometer tests in favor of tests only in real driving conditions.

Valverde [39] (2023) finds no statistically significant differences in NO_x, CO, and PN emissions measured over the complete WLTC and RDE tests, both for gasoline and diesel vehicles.

Other authors draw attention to the dependence of nitrogen oxide emissions on altitude. The team of authors [40,41,42] examined the emission of nitrogen oxides from diesel engines and found a significant depletion of the fuel-air mixture at high altitudes.

Literature analysis indicated that work has already been carried out on individual parts of the new measurement method. However, a final, comprehensive, and usable procedure has not yet been achieved. This means that it is necessary to fill this gap by proposing a method for estimating exhaust emissions in the RDE test based on the results obtained in laboratory conditions.

4. Research Method

4.1. Test Objects

Three vehicles with different drive systems were used in the research (detailed characteristics were given in the first part of the article):

• a vehicle equipped with a gasoline engine;
• a vehicle equipped with a diesel engine;
• a hybrid vehicle.

Detailed characteristics of the tested vehicles were assessed (

Table 1. Characteristics of the drive units of the tested objects.

Parameter	Unit	Gasoline	Diesel	Hybrid
Engine displacement	dm3	1.6	2.0	1.8
Number of cylinders/valves	–	4/16	4/8	4/16
Maximum power	kW/rpm	81/5500	75/2750	73/5200 100 (electric)
Torque	Nm/rpm	152/4500	280/1500	142/4000
Volumetric power indicator	kW/dm3	50.6	37.5	55.5
Vehicle curb weight	kg	1349	1584	1415
Exhaust emission standard	–	Euro 6d	Euro 6d	Euro 6d
Drive type	–	front	front	front
Designation used at work	–

). The common feature of the selected vehicles was the vehicle class (M1, passenger cars), with similar curb weight (1350–1584 kg) and a similar maximum power of the internal combustion engine (73–81 kW) and emission class (Euro 6d).

4.2. Equipment

The tests were performed on a chassis dynamometer. All exhaust emission measurements from the test vehicles were carried out in accordance with the WLTC test requirements enforced in Europe. An AVL 4WD chassis dynamometer was used in the tests. It was designed for testing passenger cars with both front and rear drives. The laboratory was equipped with a set of devices for exhaust gas sampling and analysis manufactured by HORIBA. It consisted of an exhaust gas sampling system (CVS—constant volume sample) with a dilution tunnel and a set of HORIBA MEXA exhaust gas analyzers.

Measurement Equipment for Road Tests

The Semtech DS mobile analyzer from Sensors was used to measure the concentration of harmful compounds in exhaust gases and the exhaust gas mass flow rate (CO, HC, NO_x, and CO₂).

Table 2. The specifications of the Semtech DS measuring apparatus along with the data from the vehicle’s internal data transmission system.

Parameter	Measurement Method	Accuracy
CO	NDIR, 0–10%	±3% range
HC	FID, 0–10,000 ppm	±2.5% range
NOx (NO + NO2)	NDUV, 0–3000 ppm	±3% range
CO2	NDIR, 0–20%	±3% range
O2	paramagnetic, 0–20%	±1% range

shows its characteristics along with the data from the vehicle’s on-board system.

The Semtech DS. analyzer consists of several autonomous measurement modules:

• A flame ionization detector (FID) is used to determine the total concentration of hydrocarbons in exhaust gases, referred to as HC or THC (total hydrocarbons);
• Non-dispersive ultraviolet analyzer (NDUV), measuring the concentration of nitric oxide and nitrogen dioxide;
• Non-dispersive infrared analyzer (NDIR), designed to measure the concentration of carbon monoxide and carbon dioxide;
• A paramagnetic detector (PMD) to determine the oxygen concentration in the exhaust gases.

Exhaust gases were delivered into the analyzer using a sampling probe kept at a temperature of 191 °C, where particulate matter was filtered. In the next step, the hydrocarbon concentration was measured using the flame ionization detector. The exhaust gases were then cooled to a temperature of 4 °C, after which the concentration of nitrogen oxides was measured (by the non-dispersive method using ultraviolet radiation, enabling simultaneous measurement of the concentration of nitrogen oxide, and nitrogen dioxide), carbon monoxide, carbon dioxide (by the non-dispersive method using infrared radiation) and oxygen (electrochemical analyzer). The exhaust gas flow rate was measured using 2″ and 2.5″ sampling probes. Due to road conditions, it was necessary to mount the probe in such a way as to ensure it was airtight with the vehicle’s exhaust system.

The EEPS 3090 (Engine Exhaust Particle Sizer™ Spectrometer, TSI Incorporated, MN, USA) analyzer was used to measure particle diameters. It allowed for the measurement of particles in the diameter range from 5.6 nm to 560 nm (in this study, particles > 23 nm were counted). Note that EEPS uncertainty is higher than particle counters and that volatiles are also measured [43]. Technical data of the TSI 3090 EEPS analyzer are provided below (

Table 3. The EEPS particle sizer parameters [44].

Operating Features	Value
Particle size range	5.6 to 560 nm
Particle size resolution	16 channels per decade (32 total)
Electrometer channels	22
Time resolution	10 size distributions/s
Sample flow	10 L/min
Sheath air	40 L/min
Operating temperature	0 to 40 °C

).

4.3. Test Route—Road Tests

The test route was selected in accordance with the requirements as outlined in regulations [9,10,11]. The vehicles—in terms of emissions—were tested in RDE road tests on the route shown in Figure 1.

It was possible to compare the data from the tested vehicles due to the fact that the road carbon dioxide emissions were on the characteristic curve for each vehicle between the values of ±50% (red line in Figure 2) from the value determined by points P₁, P₂, and P₃. These points correspond to the CO₂ road emission values in individual phases of the WLTC test. Due to the fact that almost all data recorded was between the values of ±25% (green lines), it can be assumed that the analysis of the values in the measurement windows and from the two-dimensional characteristics will be the same.

4.4. Data Comparison Methods Used

Comparing the results of road exhaust emissions in road tests can only be valid if the journeys during road tests are similar. However, road tests are characterized by low repeatability of road conditions. The definition of correctly performed road tests, as described by dynamic parameters (relative positive acceleration or the product of speed and positive acceleration) may be insufficient to confirm the comparability of the results obtained in road tests.

Repeating the tests several times on the same test route was characterized by different vehicle operating conditions, as well as different engine operating conditions. A comparison of vehicle (or engine) operating conditions can be made based on two-dimensional vehicle operating characteristics. For this purpose, the individual ranges of vehicle speed and acceleration were numbered on the two-dimensional characteristics comparison (Figure 3). The numbered individual ranges of vehicle speed and acceleration for a given test journey were compared with the same ranges (field numbers) from the compared journey. This resulted in two columns of data that would be used to determine the regression equation (y = ax + b), in which the coefficient of determination (R²) was a measure by which different trips can be compared. The compared data was similar if the slope of the line (a) was close to 1 and the value of the intercept (b) was close to 0.

Four calculation procedures were used to analyze and compare emission results in the WLTC and RDE tests:

• Procedure 1 (labeled as WLTC)—where the measurement of road exhaust emissions was done according to the standard WLTP procedure;
• Procedure 2 (labeled as RDE)—where the road test was divided into phases and exhaust emissions were measured according to the standard RDE procedure;
• Procedure 3 (labeled as WLTC₁₊₂)—where the test was divided into phases: 1 + 2, 3, and 4; the combination of phases 1 and 2 corresponded to the urban phase of the RDE test;
• Procedure 4 (labeled as WLTC_RDE)—where the test was divided into phases and exhaust emissions were measured according to the RDE procedure, which divides test phases based on the vehicle speed.

5. Estimation of Road Emissions in Road Tests Based on Data from Homologation Tests

5.1. Comparison of Road Emissions for All Tested Cases

A comparison of the final results obtained in the laboratory and road tests, i.e., road exhaust emissions in individual test phases according to the previously established test procedures (WLTC, WLTC₁₊₂, WLTC_RDE, and RDE) was made (Figure 4). It should be noted that the final results in the WLTC test and WLTC₁₊₂ were the same—because they are the same emission measurement test. When narrowing down this result, only the other measurement procedures were considered.

The obtained results of the tests performed, divided into urban, rural, and motorway phases, showed the convergence of the results obtained. The road exhaust emissions results obtained for a vehicle equipped with a gasoline engine were most similar to those for road emissions of carbon dioxide. The differences, in this case, were small, and the range of results obtained was: for the urban phase 183–191 g/km, for the rural phase 147–154 g/km, for the motorway phase 161–189 g/km and for the entire test 164–173 g/km. There were significant differences in road emissions of carbon monoxide: the highest value was in the motorway phase (327–430 mg/km), and the lowest in the rural phase (147–242 mg/km). Road emissions of nitrogen oxides were the highest in phases 1 and 2 of the WLTC test and were three times higher than in the urban phase of the RDE test as well as for the WLTC_RDE procedure. The road emission of particle number was most similar to the RDE test during the urban phase and throughout the entire test when determined according to the WLTC_RDE procedure.

The road emission results obtained for a vehicle equipped with a diesel engine were similar (regardless of the test performed) with respect to road emissions of carbon dioxide and carbon monoxide. The differences in road carbon dioxide emissions were small, and the obtained results were in the range: for the urban phase 164–180 g/km, for the rural phase 143–180 g/km, the motorway phase 174–215 g/km and for the entire test 161–196 g/km. The differences in road carbon monoxide emissions were: for the urban phase 104–118 mg/km, for the rural phase 75–86 mg/km, for the motorway phase 82–101 mg/km, and for the entire test 90–99 mg/km. Road emissions of nitrogen oxides were the highest in the urban phase (in the range of 23–52 mg/km), in the remaining phases it did not exceed 14 mg/km, 4 mg/km, and 22 mg/km for the rural phase, the motorway phase and in the entire test, respectively. The highest values (outside the urban phase) were recorded in the standard RDE test, and the lowest—were found in the WLTC test done according to the RDE procedure. The road number of particulate matter emitted was most similar to the RDE test during the urban phase, while throughout the test it was most similar to the values of the WLTC_RDE procedure.

The road emission results obtained for a vehicle equipped with a hybrid drive were similar in terms of measured emissions of carbon dioxide, carbon monoxide, and the number of particles. The differences in the first case were small regardless of the test procedure, and the range of results obtained was as follows:

• for the urban phase 94–111 g/km;
• for the rural phase 94–109 g/km;
• for the motorway phase 122–154 g/km;
• for the whole test 107–118 g/km.

The differences in road carbon monoxide emissions were greater (especially in the rural phase), and the road emission results were as follows:

• for the urban phase 80–93 mg/km;
• for the rural phase 32–69 mg/km;
• for the motorway phase 34–39 mg/km;
• for the whole test 54–61 mg/km.

Road emissions of nitrogen oxides were the highest in phases 1 and 2 of the WLTC test and were 4 times higher than in the urban phase of the RDE test and when measured according to the WLTC_RDE procedure. The smallest differences in road emissions of nitrogen oxides were recorded for the motorway phase and in the entire test. The road particle number emission was similar in each test phase; the largest value was observed in the urban phase and was approximately 10–20 times greater than in the rural and motorway phases. The final values were similar to each other (range 3.5–4.0 × 10¹¹ 1/km).

5.2. Correlations of Road Exhaust Emission Results between Individual Tests

Correlations of road exhaust emission results between individual tests were made in three stages, in which comparisons were made not only of the final values but also for each of the individual phases (or their groups) in accordance with the nomenclature applicable in the article:

• step I: comparison of road emissions in the WLTC test according to the calculation procedure combining phases 1 and 2 with road emissions in the WLTC test according to the RDE calculation procedure;
• step II: comparison of road emissions in the WLTC test according to the calculation procedure combining phases 1 and 2 with road emissions in the RDE test;
• step III: comparison of road emissions in the WLTC test according to the RDE calculation procedure with road emissions in the RDE test.

A comparison of the obtained road emission values in the WLTC test using two calculation methods: the first—combining phases 1 and 2 of the test and the second—using the division of phases according to the RDE procedure as shown in Figure 5 (stage I). It shows that road carbon dioxide emissions determined using these two methods were similar to each other (R² > 0.85) irrespective of the type of vehicle tested. High values of the determination coefficient were also obtained when comparing the road carbon monoxide emission values for a vehicle with a diesel engine and for a vehicle with a hybrid drive (R² > 0.78). Road emissions of nitrogen oxides determined for the WLTC₁₊₂ procedure were very consistent with those from the WLTC_RDE procedure (R² > 0.96), with the exception of the hybrid vehicle, for which no such relationship was found. Satisfactory results were also obtained for the particle number, where similar results were obtained for a vehicle equipped with a diesel engine and for a vehicle with a hybrid drive.

In the second step, road exhaust emissions in the WLTC test were compared according to the calculation procedure combining phases 1 and 2 with road emissions in individual phases and the entire RDE test (Figure 6). The highest value of the coefficient of determination for road carbon dioxide emissions in the performed tests was found for a vehicle equipped with a diesel engine (0.853), while a lower value was obtained for other vehicles; 0.538—for a vehicle equipped with a gasoline engine and 0.68—for a vehicle with a hybrid drive. When comparing road emissions of carbon monoxide, the highest value of the determination coefficient was observed for a vehicle with a diesel engine (0.939), while the lowest for a vehicle with a hybrid drive (0.385). A similar situation occurred when comparing the determination coefficients for road exhaust emissions of nitrogen oxides, where the highest value (0.963) was also applied to a vehicle with a diesel engine, and the lowest for a vehicle with a gasoline engine. A different result was obtained when considering the particle number, where the results obtained between the tests were very consistent.

The considered results of road exhaust emissions in the WLTC₁₊₂ and RDE tests were most consistent with each other for a vehicle equipped with a diesel engine. However, the results of road exhaust emissions for other vehicles did not show such consistency.

In the third step of the analysis, road emissions obtained in the WLTC test were compared according to the RDE calculation procedure with road emissions in the RDE test (Figure 7). For road carbon dioxide emissions, regardless of the type of vehicle, the obtained values of the determination coefficient were high (0.499–0.923). Unfortunately, a clear relationship (the highest value of the determination coefficient) could be observed only for a vehicle with a hybrid drive. For road emissions of carbon monoxide, the highest values of the determination coefficient were obtained when comparing test results for vehicles with a gasoline engine (0.936) and with a hybrid drive (0.980). For the vehicle with a diesel engine, the obtained value was only half that (0.557), which was not a satisfactory result. For nitrogen oxide road emissions, the coefficient of determination value can be concluded, based on the WLTC test and the RDE procedure (WLTC_RDE marking) in the RDE test, only for the vehicle with a diesel engine, which in this case was 0.953. Considering the particle number, the best matching of emission values between the tests was found for a hybrid vehicle and a vehicle with a diesel engine.

To sum up, no clear grounds for replacing the final results of road exhaust emissions in road tests with the results obtained in individual phases (also determined according to the RDE procedure) in the WLTC homologation test could be established. The consistency of individual results would have to be confirmed for each specific research case. It should be noted that the high value of the coefficient of determination obtained for a specific test does not mean that the result from the WLTC test can be directly accepted as representative of the value obtained in the RDE test. The coefficient of determination can only be used to assert that the measured values were proportional, but the equation that will be used to convert the WLTC_RDE test values to the RDE test estimated values must always be specified.

Due to the fact that the general possibility of estimating road exhaust emissions in road conditions based on the approval test was not confirmed, it became necessary to undertake another analysis, which was to assess howe consistent are the test results based on the exhaust emission rate, which was done in the next subsection of this article.

5.3. Using the Exhaust Emission Rates Obtained in the Homologation Test to Estimate the Results in the Road Test

Another aspect of the research done was a detailed analysis of the exhaust emission rate of carbon dioxide, carbon monoxide, nitrogen oxides, and the particle number in relation to the dynamic vehicle parameters (i.e., speed and acceleration). This stage of the research relied on data on the emission rate of individual harmful exhaust components obtained in the WLTC_RDE and RDE procedures for the tested drive types. The obtained results were presented as emission rate matrices in vehicle speed-acceleration coordinates, divided according to the test phase.

Analyzing the carbon dioxide emission rate from the tested vehicles, a fairly large convergence of the obtained results can be found (Figure 8, Figure 9 and Figure 10). For a conventional vehicle equipped with a gasoline engine, the highest values of carbon dioxide emission rate (10–12 g/s) occurred in the speed range of 80–110 km/h and acceleration of 1.25–1.5 m/s² in the WLTC_RDE test, and in the RDE test, these occurred in the speed range of 70–90 km/h and acceleration 1.75–2 m/s² as well as the range v = 110–130 km/h and a = 1.5 m/s² (Figure 8). The range of carbon dioxide emission rate variability was found to be very similar for both tests (the maximum value was approximately 12 g/s). The values of carbon dioxide emission rate in the urban, rural, and motorway phases were also similar, which created the possibility that the mutual comparison of the values of this parameter in two different tests may be similar.

The analysis of carbon dioxide emissions for a vehicle with a diesel engine (Figure 9) has shown that the highest values occurred in the speed range of 130–160 km/h and acceleration in the range of 0.5–0.75 m/s² (in the WLTC_RDE test) and in the speed range of 100–140 km/h and acceleration of 1.0–1.5 m/s² (in the RDE test). The changes in the emission rate of this exhaust component ranged from 0 g/s to 11.5 g/s—regardless of the test used. In this case, there was also a similar trend of changes in both the exhaust emission tests.

The lowest values of carbon dioxide emission rate were obtained for the hybrid drive vehicle (Figure 10), and extreme values were obtained in the speed range of 110–130 km/h and acceleration of 1.0–1.25 m/s² (in the WLTC_RDE test) and v = 80–90 km/h and a = 1.0–1.5 m/s² (in the RDE test in the rural phase).

When analyzing the carbon monoxide emission rates from the tested vehicles, a large convergence of the obtained results was observed, but it was smaller than in the case of the previous relationship (Figure 11, Figure 12 and Figure 13). For a conventional vehicle equipped with a gasoline engine, the highest values of carbon monoxide emission rate (approx. 80 mg/s) occurred in the speed range of 100–110 km/h and acceleration of 1.25–1.5 m/s² in the WLTC_RDE test, and in speed ranges of 110–130 km/h and acceleration of 1.5–1.75 m/s² in the RDE test (Figure 11). The range of variability of the carbon monoxide emission rate was larger for the WLTC_RDE test—approximately 80 mg/s, while for the RDE test, it was approximately 60 mg/s.

The carbon monoxide emission rate for a vehicle with a diesel engine (Figure 12) showed that these values were much lower and the extreme values occurred in the speed range of 130–160 km/h and acceleration in the range of 0.75–1.0 m/s² (in the WLTC_RDE test), while in the RDE test, this took place in the speed range of 100–110 km/h and acceleration 1.25–1.5 m/s². The values of changes in the emission rate of this component ranged from 0 mg/s to 4.7 mg/s (for the WLTC_RDE test) and up to 5.7 mg/s (in the RDE test).

The lowest values of carbon monoxide emission rate were obtained for the hybrid vehicle (Figure 13), and the highest values were obtained in the speed range of 100–130 km/h and acceleration of 1.25–1.5 m/s² (in the WLTC_RDE test) and v = 100–110 km/h and a = 1.25–1.5 m/s² (in the RDE test in the urban phase). The operating characteristics of the hybrid drive system meant that there was no significant similarity in the rate of carbon monoxide emissions both in the entire research test and in its individual parts.

The nitrogen oxide emissions rate from the tested vehicles has shown a large convergence of the obtained results, especially the maximum values, and only for a vehicle with a gasoline engine. For a conventional vehicle equipped with a gasoline engine, the highest nitrogen oxide emission rate values (of approx. 0.7 mg/s) were observed in the speed range of 100–110 km/h and acceleration of 1.25–1.5 m/s² in the WLTC_RDE test, while in the RDE test, these were found in speed ranges of 90–100 km/h and acceleration of 1.25–1.5 m/s² (Figure 14). The range of variability of the carbon monoxide emission rate was larger for the RDE test—approximately 1.05 mg/s, than for the WLTC_RDE test—approximately 0.7 mg/s.

The nitrogen oxide emission rates for a vehicle with a diesel engine (Figure 15) had extreme values 6 times greater in the RDE test than in the WLTC_RDE test. The maximum emission rates of nitrogen oxides occurred in the speed range of 50–60 km/h and acceleration in the range of 0–1.75 m/s² (in the WLTC_RDE test), while in the RDE test, it was in the speed range of 100–110 km/h and acceleration of 1.25–1.5 m/s². The changes in the emission rate values for this component ranged from 0 mg/s to 0.9 mg/s (for the WLTC_RDE test) and up to 5.3 mg/s (in the RDE test).

Once again the lowest values of nitrogen oxide emission rate were obtained for the hybrid vehicle (Figure 16), and the highest values were obtained in the speed range of 120–130 km/h and acceleration of 0.25–0.5 m/s² (in the WLTC_RDE test) and v = 90–100 km/h and a = 1.25–1.5 m/s² (in the urban phase of the RDE test). The hybrid drive’s operating characteristics meant that no significant similarity in the nitrogen oxide emission rate was found, both in the entire emission test and in its individual parts.

The particle number emission rate from the tested vehicles was analyzed, and the convergence of the obtained results was found only for a vehicle with a gasoline engine. For a conventional vehicle equipped with a gasoline engine, the highest values of the particle number emission rate (approx. 1.3 × 10¹¹ 1/s) occurred in the speed range of 80–90 km/h and acceleration of 1.25–1.5 m/s² in the WLTC_RDE test, and in speed ranges of 70–90 km/h and acceleration 1.75–2.0 m/s² for the RDE test (Figure 17). The range of variability of the particle number emission rate was larger for the WLTC_RDE test at approximately 1.3 × 10¹¹ 1/s than for the RDE test (7.5 × 10¹⁰ 1/s).

The particle number emission rate for a vehicle with a diesel engine (Figure 18) showed that the top emission values were twice as high in the stationary dynamometer test than in real driving conditions. The maximum particle number emission rate occurred in the speed range of 80–100 km/h and acceleration in the range of 1.0–1.25 m/s² (in the WLTC_RDE test) and in the speed range of 110–120 km/h and acceleration 0.75 m/s² for the RDE test. The emission rate values varied within a range of up to 6.5 × 10⁹ 1/s (for the WLTC_RDE test) and up to 3.2 × 10⁹ 1/s (in the RDE test).

The hybrid vehicle produced similar values of the particle number emission rate to the vehicle with a gasoline engine (Figure 19), and the extreme values were obtained in the speed range of 50–70 km/h and acceleration up to 0.5 m/s² (in the WLTC_RDE test), and v = 50–70 km/h and a = 1.0–1.5 m/s² (in the urban phase of the RDE test, at high acceleration values).

In order to compare the above data, the relationships between the emission rates of the analyzed exhaust gas components (carbon monoxide and dioxide, nitrogen oxides, and the number of particles) were determined in both WLTC and RDE tests. When analyzing the emission rate, the test evaluation procedure was not important, because the emission rate was considered for each operating point of the vehicle anyway. The obtained results were divided according to the type of vehicle tested. The coefficient of determination was calculated for each correlation. Figure 20, Figure 21, Figure 22 and Figure 23 depict the dependencies determined for the exhaust emission rate of the considered harmful compounds, taking into account all operating points (for the entire emission test). For the rate of carbon dioxide emissions (Figure 20), the results of the determination coefficient varied in the range from 0.835 (vehicle with a hybrid drive) up to 0.876 (vehicle equipped with a diesel engine). For a vehicle with a gasoline engine, the value of the determination coefficient was also high at 0.867.

For the carbon monoxide emission rate (Figure 21), the values of the determination coefficient were significantly lower and varied in the range from 0.141 (vehicle equipped with a gasoline engine) to 0.333 (vehicle with a hybrid drive). For a vehicle with a diesel engine, the value of the determination coefficient was 0.316.

For the nitrogen oxide emission rate (Figure 22), the value of the determination coefficient was greater than zero only for the vehicle equipped with a gasoline engine—where it was equal to 0.174. For the remaining tested vehicles, the determination coefficient value was less than zero, which indicated the lack of any meaningful relationship between the nitrogen oxide emission values determined in two different tests. As for the particle number emission rate (Figure 23), the situation was similar to the previous one, where the value of the determination coefficient was greater than zero only for the vehicle equipped with a gasoline engine (nevertheless still as low as 0.257). For the remaining tested vehicles, this value was around zero, which again showed the lack of any meaningful relationship between the values of the particle number emission rate determined in the two different tests.

All the values of the determination coefficients have been summarized and divided into test phases (

Table 4. Values of the determination coefficients obtained for the tests under consideration in individual phases and in the entire test (values higher than 0.7 were marked in bold).

		Urban	Rural	Motorway	Total
	CO2	0.957	0.975	0.930	0.867
	CO	0.261	0.867	0.570	0.141
	NOx	0.023	0.468	0.509	0.174
	PN	0.923	0.597	0.293	0.257
	CO2	0.801	0.964	0.953	0.876
	CO	0.189	0.963	0.883	0.316
	NOx	0.157	0.307	0.047	0.024
	PN	0.504	0.733	0.354	0.014
	CO2	0.835	0.985	0.885	0.835
	CO	0.740	0.821	0.742	0.333
	NOx	0.012	0.428	0.256	0.038
	PN	0.064	0.357	0.535	0.003

). According to the resulting data only a few correlations were found to be characterized by a coefficient greater than 0.7, which was considered the minimum value suggesting the data was indeed correlated. It was found that the carbon dioxide emission rates were the only ones whose values could be estimated in individual phases and in the entire RDE test based just on the WLTC test results, while for the other tested harmful exhaust gas compounds such an estimate would be unfounded. A value of the determination coefficient of over 0.7, which was deemed sufficient for estimating the results in the RDE test based on WLTC results, was obtained for the following:

• The carbon monoxide emission rate (in the rural phase) and the particle number emission rate (in the urban phase) of a vehicle equipped with a gasoline engine;
• The carbon monoxide emission rate (in the rural and motorway phase) and the particle number emission rate (in the rural phase) of a vehicle equipped with a diesel engine;
• The carbon monoxide emission rate (in the urban, rural, and motorway phases) of a vehicle equipped with a hybrid drive.

The analysis of the emission rate of individual exhaust gas components allowed us to conclude that it was possible to estimate carbon dioxide emission rate value in the RDE test based on the values obtained in the WLTC test for all types of drives, while for other compounds it was possible, but limited only to a very few cases.

6. Conclusions

The negative impact of deteriorating air quality is noticeable not only in large and small cities but also increasingly in motorway areas. This was confirmed by alarming data available not only in scientific works published in various parts of the world but also in reports on air quality assessment. Such information confirms that it is still necessary to focus special attention on exhaust emissions from, among others, the road transport sector. This type of exhaust emissions is important mainly due to the structure of the automotive market, where vehicles powered by combustion engines still play a dominant role. It is true that this structure is gradually changing in favor of electric vehicles, but it will still take many years to achieve full electrification of road transport. This especially applies to countries with more unstable economic situations.

An analytical tool was developed as part of this research enabling the comparison of homologation and road tests in terms of road emission results not only in the research tests but also in their individual phases. This was the result of an analysis of the vehicle’s operating conditions in research tests, which turned out to be very similar. However, when calculating the correlation of the emission rate of harmful compounds at individual operating points related to the vehicle the relationships were not entirely clear.

The quantitative conclusions regarding the obtained road emissions values for the tested vehicles in the entire WLTC_RDE and RDE test were as follows:

• For a vehicle with a gasoline engine, the relative difference between the obtained test results was as follows:
  • for carbon dioxide—1%;
  • for carbon monoxide—4%;
  • for nitrogen oxides—58%;
  • for particle number—2%;
• For a vehicle with a diesel engine, the relative difference between the obtained test results was:
  • for carbon dioxide—18%;
  • for carbon monoxide—7%;
  • for nitrogen oxides—130%;
  • for particle number—10%;
• For a vehicle with a hybrid drive, the relative difference between the obtained test results was:
  • for carbon dioxide—2%;
  • for carbon monoxide—13%;
  • for nitrogen oxides—10%;
  • for particle number—13%.

The proposed procedures for estimating road emissions in a road test based on dynamometer tests allow the results to be reflected with high accuracy, but one such universal procedure cannot be proposed.

Table 5. The potential of estimating road emissions in a road test based on the stationary dynamometer test (WLTC₁₊₂ with combined phases 1 and 2, WLTC_RDE a braking test + RDE procedure).

Vehicle Type	Exhaust Compounds	Potential of Estimating Road Emissions in the RDE Test Using	Determination Coefficients (R2)
Gasoline	CO2	WLTCRDE	0.855
	CO	WLTCRDE	0.936
	NOx	–	–
	PN	WLTCRDE	0.480
Diesel	CO2	WLTC1+2	0.852
	CO	WLTC1+2	0.788
	NOx	WLTC1+2	0.999
	PN	WLTCRDE	0.982
Hybrid	CO2	WLTCRDE	0.923
	CO	WLTCRDE	0.980
	NOx	–	–
	PN	WLTCRDE	1.000

shows the potential of estimating road emissions in the RDE test based on the WLTC test.

It follows, that it is not possible to estimate all the parameters of the RDE test using only the results obtained in the WLTC test for the tested vehicles. However, it seems to be possible to estimate individual parameters (according to the WLTC₁₊₂ or WLTC_RDE procedure), where for some parameters the accuracy of this estimation can be expected to reach over 90%, and for others much less. Nevertheless, the issue was deemed by the Authors to have enormous potential and definitely deserves further, more in-depth analysis.

The proposed procedures for estimating exhaust emissions in road tests are an issue that should be developed as follows:

• make an assessment for a larger group of vehicles of a given type and also extend it to plug-in hybrid vehicles;
• compare data using other correlation methods, taking into account the probability of the results obtained;
• extend the possibility of estimating road emissions to additional harmful compounds, ammonia, aldehydes, and nitrous oxide, which are to be introduced in the coming years (in the Euro 7 standard);
• introduce a larger number of road tests differing in driving dynamics, so as to expand the range of dynamic parameters of the tests performed;
• introduce restrictions related to the time or duration of tests that may affect the results of road pollutant emissions (e.g., by shortening the urban phase to minimum values).