Natural Gas, a Mean to Reduce Emissions and Energy Consumption of HDV? A Case Study of Colombia Based on Vehicle Technology Criteria

In this study, the use of compressed and liquefied natural gas is evaluated for heavy-duty passengers (HDPV) and freight vehicles (HDFV). The evaluation is conducted considering the socioeconomic and vehicle fleet characteristics of Colombia. The energy consumption, the CO2, and the pollutant emissions of a baseline and four natural gas penetration scenarios are analyzed. The results show that the inclusion of natural gas reduces the energy consumption per capita of the HDPV and HDFV by up to 40% by 2050. Furthermore, PM2.5 emissions per capita are reduced up to 77% for HDPV and 90% for HDFV, while CO emissions per capita decreased by 82%. Additionally, the technological renovation of HDFV emerges as an effective way to reduce pollutant emissions in the medium term. The establishment of strategies to make HDFV cleaner and more efficient is imperative for the long term. Finally, a sensitivity analysis is conducted to evaluate the influence of the gross domestic product per capita (GDPc) over the indicators analyzed. The results show that higher GDPc demands more ambitious actions to decarbonize the transportation sector, since a considerable increase in energy consumption and emissions from heavy-duty vehicles is identified.


Introduction
Transport is one of the key economic sectors to propose strategies for reducing emissions. It is one of the primary producers of pollutants and greenhouse gasses (GHG) emissions. In 2018, this sector produced about 25% of CO 2 global emissions [1]. Likewise, the transport sector is one of the primary generators of pollutant gases that affect air quality in urban areas. Chemical species like particulate matter (PM), tropospheric ozone (O 3 ), nitrogen oxides (NO x ), and sulfur oxides (SO x ), are intricately linked to the operation of the transport sector. The use of cleaner fuels as hydrogen, biofuels, electricity, natural gas, and carbon synthetic fuels emerges as a strategy to achieve environmentally sustainable transport [2].
In this sense, socioeconomic parameters are simulated to evaluate the possible consequences of the introduction of new transport technologies in existing transport systems. [3]. Nevertheless, not all fuels are competitive in terms of performance or availability. The penetration of electric vehicles has been slow, principally in countries with developing economies [4]. Today, natural gas has emerged as a possible transition fuel to substitute the use of diesel and gasoline in road transport vehicles, since its carbon content is lower. Generally, natural gas can be employed in transport as compressed natural gas (CNG) or liquefied natural gas (LNG) [5]. A change from petroleum-based energy chains to natural gas-based energy chains could be a good strategy to both reduce emissions of fuels used Table 1. Studies related to the evaluation of natural gas in HDV in the world.

Canada
Life cycle analysis (LCA) and dynamic simulation of LNG and CNG as alternatives to diesel in buses.
Through a life cycle analysis (LCA), when diesel was changed for CNG and LNG, there were GHG reductions of 4.8% and 8.1%, respectively. [22] Poland Cost-benefit analysis with the use of CNG in public transport buses.
Purchasing CNG buses is more profitable when the price of CNG per m 3 is 55% below the m 3 cost for diesel. The LCA of a CNG bus is reduced by 400 kg of GHG and 52.5 kg of PM. The environmental benefits are sufficient to justify the use of CNG in buses by transport companies. [20] Taiwan Study of the carbon footprint of an urban bus in the city of Taiwan, based on the international regulation ISO/TS 14067:2013.
Carbon footprints in the lifecycle of buses measured from highest to lowest: LNG 63.14 g CO 2e /km; diesel, 54.6 g CO 2e /km; liquefied petroleum gas, 47.4 g CO 2e /km; plugin electric, 37.82 g CO 2e /km, and hydrogen gas cell, 29.17 g CO 2e /km. [23] China Identification of polluting emissions of buses using a mean distribution deviation (MDD) for different fuels on the road.
On the road, bus stops, and intersections, the diesel vehicles with standard Euro IV and Euro V have higher CO and NO x emissions than natural gas buses. Nevertheless, for CNG buses, the levels of CO 2 and HC are higher.
There is no significant difference in velocity or acceleration. [7] LCA of LNG and CNG in HDV, taking into account the gas leakage in the supply chain.
Trucks that use LNG and CNG can reduce GHG emissions by 11.17% and 5.8%, respectively, in comparison to diesel vehicles in the lifecycle. [24] CO 2 , NO x , CO, THC emissions analysis in real operating conditions using a portable emission measuring system (PEMS) in semi-trucks that use diesel and LNG in China.
The emission factors of THC of the LNG vehicles were 2.01 to 10.63 g/km higher than diesel vehicles (0.03 to 0.19 g/km). Likewise, vehicles operating with LNG had higher NO x and CO 2 emissions than diesel vehicles. [25] Engine combustion and emission fueled with natural gas: A review Compared to gasoline, natural gas vehicles have a lower production of THC, CO, CO 2 , and PM. However, NO x emissions are higher. In natural gas vehicles, methane is the main component of THC emissions. It can be reduced by the addition of hydrogen and after-treatment technology. [26] United States Optimization of the impact of the evaluation of the lifecycle of bus fleets using alternative fuel. CNG vehicles had 23% less GHG emissions than diesel vehicles. [27] Analysis of GHG reductions that can be achieved through increasing engine efficiency and transition to low-carbon fuels.
Compared to gasoline, natural gas provides a 20% reduction of GHG emissions in the internal combustion engine and up to 50% when it is used as a raw material to produce hydrogen or electricity. [24] Comparison of the complete lifecycle of transit autobuses with diesel engines and five types of buses using alternative fuels.
The number of buses fueled with B20, CNG, and LNG has high participation in today's transit operations, but their CO 2 emissions do not reveal a significant reduction compared to diesel buses. LNG-fueled transit buses showed the worst results in terms of CO 2e emissions during their lifetime of 12 years for the Manhattan operation cycle compared to other cycles. [25] Brazil Economic analysis and regulation barriers for the inclusion of natural gas as a fuel in HDFV.
The availability of LNG vehicles on the market reduces the possibility of including natural gas as a combustible in HDFV. The cost of fuel makes it possible to incur investments of up to 17% higher than in diesel fuel vehicles. [8] Analysis of the increase of use of LNG in the transport sector in different parts of the world.
The use of CNG and LNG have increased in different types of vehicles in various countries, among them Italy, New Zealand, the United States, Australia, Iran, and France. Italy is the country with the most developed program using natural gas as a fuel, where almost 300,000 vehicles circulate with CNG. Countries like Brazil, Egypt, Canada, and Bangladesh implemented programs to use natural gas for internal transport needs. [28] United Kingdom Energy and environmental well-to-wheel analysis of LNG as a transition fuel in heavy-duty vehicles.
GHG emissions reductions of the LNG trucks only happen when there are efficiency improvements in the vehicles. The LNG vehicles perform 17% less efficiently than diesel, provoking 7% more GHG emissions. [10] South Korea General vision about the LNG and CNG characteristics for automobiles.
For a tank of fuel of a determined capacity, a vehicle that runs on LNG can travel up to 2.4 times the distance of its CNG counterpart. [16] Colombia Energy and emission impact of natural gas in heavy-duty vehicles: A case of study of Colombia based vehicle technology criteria The introduction of natural gas for HDPV and HDFV reduces energy consumption by up to 40% by 2050. Pollutant emissions such as CO and PM 2.5 decrease up to 80 and 90%, respectively. The use of LNG in HDFV increases NO x emissions by 30% in 2050.
-In this article, the authors want to assess if natural gas can be considered as a transition fuel in the transport sector in Colombia based on the energy and environmental effects that result from the inclusion of this fuel in the operation of the HDPV and HDFV fleet. For this, models that relate the socioeconomic, energy, pollutant emissions, vehicle technology, and energy supply infrastructure parameters are proposed for road transport systems. The models are applied in the case study of Colombia; nevertheless, they can be replicated in any region based on data of the fleet composition, population, and the use of the energy sources in the transport sector.
An estimate and projection are made of the number of commercial transport vehicles and tank-to-wheel (TTW) pollutant emissions resulting from the operation of the HDPV and commercial vehicle fleet in the country through the year 2050. The above provides an evaluation of the environmental and energetic effectiveness of the use of passenger buses powered by CNG in public transport and HDFV powered by LNG, considering the current HDV technologies, based on diesel and gasoline fuels, that are being commercialized in Colombia. Based on the model, several scenarios of the inclusion of natural gas in HDV can be proposed and analyzed. The results could be used to draw insights, support, and argue the development of national and regional policies, as well as to project the required actions in the medium and short term.
Finally, a sensitivity analysis is executed to identify what factors influence the projections of energy consumption and pollutant emissions in the case study, and medium and long-term regulatory recommendations are established.

Methodology
An analysis of HDV is conducted, since this fleet is one of the primary causes of NO x , SO x , CO, and PM emissions in urban areas [29]. Currently, Colombia has about 341,100 HDV (26% HDPV and 74% HDFV) in operation. 11% of HDPV uses gasoline fuel, meanwhile 86% and 3% uses diesel and natural gas respectively. On the other hand, 9% of HDFV uses gasoline and 91% uses diesel.

Vehicle and Population Projections
The size of the HDV fleet in Colombia by 2050 is projected through the Gompertz function. It has been widely used for modeling the growth of vehicle fleets considering the market and the socioeconomic characteristics of a specific population. This function models the relation between the number of vehicles and the economic development of a country. It establishes the number of vehicles per capita in the function of the gross domestic product per capita (GDPc), as shown in Equation (1) [30].
where: V is the motorization index, in number of vehicles per capita. ff and fi establish the form of the Gompertz function. These parameters are determined using a linear regression method that allows the curve to be adjusted according to the specific conditions of the region, based on the historical data of the vehicle fleet.
fl: is known as the saturation level and determines the maximum point of growth of the fleet. It is expressed in vehicles per 1000 inhabitants and depends, principally, on the population density and urbanization rate of the country [30]. Initially, it is determined for Light-Duty Vehicles (LDV), according to Equation (2) [30]. Later, the saturation level of HDPV and HDFV are determined based on the relation between the number of automobiles and the number of vehicles in each category.
where: γ max is the maximum saturation level of automobiles. As a reference, it is considered the γ max of the United States, with 652 automobiles for every 1000 inhabitants [30].
D i and U i relate population density and the urbanization rate, respectively, and are calculated according to Equations (3) and (4), respectively.
where: D US and D i are the population densities of the United States and the country of study, in this case, Colombia. For these, values of 36 and 45 inhabitants/km 2 are considered, respectively [31]. U US and U i are the urbanization rates of the United States and the case study, Colombia. Values of 82.7 and 81.4% are considered respectively [32].
Once the saturation level is determined for LDV, the saturation level of HDPV and HDFV is estimated according to the historic relationship between the number of automobiles and the quantity of HDPV and HDFV. In the last 15 years, an average relation of 3.2 HDPV and 8.47 HDFV per 100 automobiles is established. Table 2 shows the parameters of the Gompertz function. The level of saturation γ is established based on the ratio between the number of automobiles of the country and the quantity of HDPV and HDFV, knowing the level of saturation of automobiles in the country. It is calculated based on what Dargy, Gately, and Sommer [30] established. From this model, the quantity of HDPV and HDFV from 2020 to 2050 is projected, based on the actual data from 2010 to 2020, and assuming an average annual growth of GDPc of 4% each year as established in the report "The World in 2050" [33]. In Figure 1, the projection of the number of vehicles per capita is presented. In which an elevated growth is identified from 2030 and greater growth for HDFV than for HDPV, reaching 28 vehicles per 1000 inhabitants by the year 2050 for HDFV as compared to a total of 7 per 1000 inhabitants of HDPV by the year 2050. The substitution of diesel fuel and gasoline for natural gas is considered, specifically HDPV powered by CNG, and HDFV powered by LNG. Four scenarios of natural gas penetration are proposed:

•
The baseline scenario follows the current trend of natural gas use in HDV. Therefore, it considers a minimal penetration of natural gas, with a participation of 95% diesel and 5% natural gas by 2050.
• In a low scenario, greater participation of diesel compared to natural gas by 2050 is assessed (75% diesel-25% natural gas).

•
In the medium scenario, equal participation of both fuels diesel and natural gas is considered by 2050.

•
In the high scenario, greater participation of natural gas compared to diesel fuel is analyzed (25% diesel-75% natural gas). • Finally, the total gas scenario considers that all the HDV fleet uses natural gas by 2050.
In all the analyzed scenarios, the withdrawal of used vehicles with more than 20 years of useful life is considered. To this end, an historical assessment of the registrations of new HDPV and HDFV in Colombia is conducted. Besides, the penetration of natural gas vehicles is analyzed progressively, considering a linear growth for each year. Under the Gompertz model, a projection of 260,000 HDPV is expected by 2050. In the base scenario, about 13,000 vehicles will use CNG. For the low, medium, and high penetration scenarios of natural gas, about 65,000; 130,000; and 195,000 vehicles will operate using natural gas. In addition, the total gas scenario will consider all the vehicles operating with CNG by 2050.
On the other hand, about 1.27 million HDFV are projected by 2050. In the base scenario, 63,700 vehicles will use LNG by 2050. In the low scenario, about 318,500 LNG-HDFV are projected; about 637,000 for the medium; and 955,500 for the high penetration scenario.

Energy Consumption
To estimate the energy consumed (EC), vehicle fuel consumption, and lower heating value (LHV) of fuels must be considered. These variables are related to the vehicle kilometers traveled (VKT) using Equation (5), through which the EC by class and vehicle technology is obtained. The VKT is calculated knowing the average of kilometers traveled per vehicle (AVKT) and the quantity of vehicles (Veh), as is shown in Equation (6). For this study, the AVKT is calculated based on the registry of total trips each year made by vehicles on the country's roads, and the number of kilometers traveled on each of the roads. Thus, the AVKT has been established at 40,576 km for HDPV and 37,877 km for HDFV [34]. Table 3 presents the mean fuel consumption of diesel and gasoline HDV fleet in Colombia under real operational conditions. In addition, this table reports the mean fuel consumption (FC) of LNG HDFV for China [19]. Table 4 presents the lower heating value of fuels and CO 2 (EF CO 2 ) emission factors.

Pollutant and CO 2 Emissions
The pollutant emissions (PE) generated by a vehicle fleet depend on the VKT and the emission factors (EF). The EF refers to the total emissions of a pollutant and can be expressed per unit of energy (g/MJ) or distance traveled (g/km). For this study, emission factors (g/km) are used, based on available information regarding emission standards with which vehicles must comply, which are presented in the Appendix A.
Therefore, as each vehicle of the fleet can accomplish a specific emission standard (ES), the PE is determined by Equation (7). It considers the participation of each ES in the fleet, in which the counter n = 1 is equivalent to the Euro I standard, through n = 6, equivalent to the Euro VI standard, for diesel and gasoline vehicles, respectively.
The actual participation percentage of ES is obtained from the baseline, and its projection is estimated assuming a vehicle lifespan of 20 years, as is indicated by the regulations of the case study country. Moreover, it is assumed that the EF in vehicles does not vary over the lifespan of the vehicle. These two considerations are made based on Colombia's regulations, which establish that diesel HDV must accomplish the Euro II standard since 2010, Euro IV since 2015, and Euro VI since 2023. On the other hand, gasoline HDV must accomplish the Euro II standards since 2015, Euro IV since 2025, and Euro VI since 2030 [29,37].
Furthermore, for the study, Euro VI is the maximum reference standard because of the lack of information on the values and implementation dates of forthcoming standards.
Carbon dioxide emissions (E CO 2 ) are calculated using Equation (8). It relates the CO 2 emission factors of (EF CO 2 ) in g/MJ, as shown in Table 4, and the used fuel through the energy consumption (EC) of the vehicles. E CO 2 = EC * EF CO 2 (8)

Results
Despite the use of natural gas has been analyzed widely, several authors have found controversial results related to the benefits of its use. This is because results are closely linked to the characteristics of the vehicle fleet under analysis and its operational conditions. Particularly for Colombia, Table 5 shows the behavior of energy consumption and CO 2 , CO, NO x, and PM 2.5 emissions per capita in the proposed scenarios compared to the baseline scenario for HDPV and HDFV by 2050. In general, the use of natural gas generated a positive impact on energy consumption and pollutant emissions is evidenced for HDPV and HDFV by 2050. Nevertheless, results show a "negative" reduction in NO x emissions of HDFV fleet for each analyzed scenario. It means that the use of LNG may cause the NO x emissions rates to increase by up to 30% in 2050 when compared to the baseline scenario. This behavior could be related to the fact that in real operating conditions the NO x emission factor is greater for LNG HDFV than Euro VI diesel ones. It occurs because the higher exhaust temperature of LNG vehicles affects the efficiency of the selected catalytic reduction system, causing a great rise in NO x emissions [25]. Nevertheless, there are ample benefits related to PM 2.5 emissions, given that LNG is considered the cleanest form of natural gas and has a lower sulfur content than gasoline and diesel [38]. Therefore, studies have shown that the use of LNG has generated reductions between 90% and 97% in PM emissions [39,40].  Likewise, the results reveal that from the year 2035, proposing complementary actions to significantly reduce NO x and CO emissions is important, principally, in HDFV. In other words, for this class of vehicles, a quicker transition to zero-emissions than for the HDPV is necessary, where new technologies like hybrid, electrical, or fuel cell trucks, not only could reduce emissions, but also provide economic benefits and fuel savings [41][42][43][44].

Sensitivity Analysis
The results represent the impact in emissions and energy consumption as a product of the inclusion of natural gas in heavy vehicles, varying the percentage of participation of fossil fuels. Nevertheless, a variation in the GDPc values, population growth, vehicle fuel efficiency, or emission factors can directly influence the results. Therefore, in this section, a sensitivity analysis of the GDPc and emission standards is executed to observe the impact of the number of vehicles, EC, and emissions.
The GDPc is directly related to the number of existing vehicles in a population, as the greater the economic development of a country the greater the purchasing power of its inhabitants. For effects of analysis, the average annual growth percentages of the GDPc were varied, with values of 2%, 4%, and 6%. Thus, it was observed that with an average annual growth of 6%, the total fleet analyzed would come close to its saturation value by 2050, with 61 heavy vehicles for every 1000 inhabitants. On the other hand, with a growth of 2% of the GDPc, there would be 16 heavy-duty vehicles for every 1000 inhabitants by 2050. Said behavior can be observed in Figure 3. Moreover, in Figure 4, the quantity of emissions of CO 2 , CO, NO x , and PM 2.5 per capita that would be generated by varying the GDPc is presented. In this figure, baseline scenario emissions with average growth percentages of the GDPc of 2%, 4%, and 6% are presented, observing a significant difference from the minimum to the maximum case. As can be observed, despite the reduction of the tons of emissions emitted due to the technological renewal of the fleet, CO, NO x , and PM 2.5 emissions reach a point of new increase. This shows that the reduction in emission factors is not enough to offset the total emissions of the number of vehicles. Additionally, a direct correlation between the GDPc and GHG and pollutant emissions has been identified. This is because a greater GDPc represents the greater purchasing power of the population to buy vehicles, which will induce an increase in the size of the fleet, thus, a higher level of emissions. This leads to the assessment that countries with greater levels of wealth should generate more rigorous strategies of decarbonization of heavy vehicles.
Additionally, Figure 5 shows how EC varies with different values of the GDPc. By the year 2050, there would be a total EC of 5.97 GJ per capita with a 2% annual growth of the GDPc, whereas the energy consumption would be 13.18 GJ per capita for a 4% annual growth. Finally, with a 6% annual average growth of the GDPc, there would be a total EC of 22.28 GJ per capita.   Figure 6 shows the influence of the GDPc in total emissions per capita in different periods in natural gas inclusion scenarios. As can be observed, CO 2 emissions are most affected by the GDPc, generating increases of up to 375% by the year 2050. Likewise, as evidenced, independent of the GDPc, there is an initial reduction in CO and NO x emissions; however, an increase in these pollutants is generated by the year 2050. The tendency that the greater the GDPc, the greater the generation of these emissions, has been identified. It was also identified that, for PM 2.5 emissions, independent of the GDPc, the inclusion of natural gas technologies generates a significant reduction of the emission of this material.

Conclusions
In this article, an evaluation of the inclusion of natural gas as a transition fuel to reduce the use of diesel and gasoline in heavy-duty passenger (HDPV) and freight vehicles (HDFV) has been executed. Taking into consideration their operating conditions, the use of compressed natural gas (CNG) in HDPV and liquified natural gas (LNG) in HDFV are considered.
Four substitution scenarios were proposed considering different natural gas penetration rates. For the evaluation, a baseline of vehicles in Colombia was used as a case study. Vehicle projections were made based on the socioeconomic criteria of the population, annual vehicle kilometers traveled (VKT), fuel economy, and vehicle emission standards.
Based on the results, we can conclude that the inclusion of natural gas in heavy transport road vehicles significantly reduces CO and PM 2.5 emissions per capita. By 2050, potential reductions of around 82% and 77% of these pollutants, respectively, are identified in HDPV. On the other hand, a reduction of PM 2.5 emissions of the order of 90% can be expected by 2050 in HDFV, compared to the baseline scenario. The results show that the use of natural gas in HDV can generate relevant environmental benefits in the medium and long term. However, it is necessary to promote the renewal of the fleet.
In the case of HDFV, there is evidence of a significant reduction in CO, NO x , and PM 2.5 emissions until 2035. Nevertheless, from this year, a net increase in the generation of these pollutants is identified. This is a product of the significant rise in the number of vehicles in operation. Thus, the inclusion of LNG in the HDFV fleet generates environmental benefits medium-term; however, long-term benefits will require more ambitious and complementary strategies of the decarbonization of the vehicles.
Based on the results of the period from 2034-2036 (medium-term), strategies should be developed to reduce pollutant emissions in the future. This said, addressing challenges in transport requires an integral approach that implements greater actions to diversify energy sources in the transport sector. Long-term, the most important action would be a transition to electric vehicles (EV), which provide a more viable road to clean transportation with zero emissions.
A sensitivity analysis of evaluated indicators concerning the GDPc was performed. It was identified that with a higher GDPc, there is a higher purchasing power of vehicles, which induces growth in the vehicle fleet in operation. Thus, this triggers a significant increase in GHG and pollutant emissions. This tells us that greater wealth-a greater GDPc-demands the deployment of more ambitious energy and environmental policies in the transport sector.
Several authors disagree on the effects of the use of natural gas in HDV since some studies reveal that CO 2 emissions and energy consumption are reduced, others report increases on these parameters. This is because results are closely linked to the characteristics of the vehicle fleet under analysis and its operational conditions. The results showed in this article are based on the historical data and information of the HDFV and HDPV fleet of Colombia and the energy demand reported on local studies. Nevertheless, the projected behavior of pollutant emissions, such as NO x is aligned to the results reported in other studies, where the increase in the emissions of this pollutant is reported when LNG is used in HDV.
The results presented in this study are based on limited information on the actual operation and technological characteristics of the HDV fleet in Colombia. Thus, to generate technical inputs closed to the reality of the country is required for supporting the enforcement of policies to reduce energy consumption and pollutant emissions on HDV. Thence, the existing gaps between real and reported data on pollutant emissions, fuel consumption, and the VKT of HDPV and HDFV need to be closed. In addition, modeling the impact of geographical conditions and the age of the vehicles on fuel consumption and pollutant emissions will allow us to perform more reliable projections. Finally, to compare natural gas-powered vehicles with electric and hybrid technologies is necessary to establish a whole roadmap to move the HDV fleet towards sustainability.

Acknowledgments:
The authors want to acknowledge the Gestión Energética Research Group (GENERGÉTICA), the Universidad Tecnológica de Pereira, and the Universidad Católica de Pereira for their contribution to the development of this research. Besides, the authors want to acknowledge to Ministerio de Ciencia Tecnología e Innovación de Colombia for the financial support to this project and to our "Jóvenes Investigadores".

Conflicts of Interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Table A1. Emission factors for HDPV and HDFV [45][46][47].