Sunday, 25 August 2024

Improvement in the Fuel Consumption by Drag Reduction in Semi-Trailer Trucks

 

Abstract – Semi-trailer trucks are the backbone of any freight system throughout the world. Annually, a large portion of the cargo is transported through semi-trailer trucks. These trucks spend most of their useable life traveling at motorway speeds. Most of the energy consumed by the trucks is used to overcome the aerodynamic drag force. The power required to overcome the drag force must be supplied by the truck’s powerplant. Therefore, it is of the utmost importance to minimize the drag force. By using computational fluid dynamics, it is shown in this paper that the application of curved plate below the floor along with side-skirts on the semi-trailer truck can significantly reduce the drag force. We show that by removing the exposure of the blunt bodies on the underside of the semitrailer such as the axle, differential and suspension components to the oncoming air considerably reduces the drag force experienced by the semi-trailer truck. A reduction in the drag force will directly result in improved energy consumption of the truck. An improved fuel consumption would result in relatively cheaper goods for the consumers. An improved fuel consumption will also lead to a reduction in the in the emission of harmful gases, leading to greener and healthier environment.

                                                                                                                                                       I.            Introduction

It is shown by the drag equation (equation 1) that, while moving through the air the drag force relates directly to the square of an object’s velocity, for the present study the object is a semi-trailer truck. As a result, the power required from a truck’s powerplant to overcome the aerodynamic drag varies with the cube of the truck’s velocity (equation 2). For example, a semi-trailer truck cruising on a highway at 40 Km/h might require 10 bhp to overcome aerodynamic drag, but that same truck now travelling at 80 Km/h would require 80 bhp from the truck’s powerplant. The powerplant can be an internal combustion engine, batteries or a combination of both. In the usual real-world situations for example, when the truck is moving into a headwind, the dependence of drag force on velocity is even more critical. The headwind has to be added to the velocity of the truck before the velocity is squared for the aerodynamic drag calculation. Thus, it is very crucial to reduce the aerodynamic drag. It is shown by research that up to 65% of the fuel is used to overcome aerodynamic drag [17].

where  is the aerodynamic drag force,  is the fluid density,  is the fluid velocity,  is the surface area,  is the drag coefficient and  is the power required to overcome drag [18].

The idea to reduce drag and/or improve downforce on a vehicle using various aerodynamics devices such as side-skirts and fairings has been around for decades. In this paper, we present an explanation as to why placing a flat plate below the floor of the truck can be used to improve fuel economy. Crosswinds have a high impact on semi-trailer trucks. Crosswinds lead to an increase in the drag coefficient and direction instability which can lead to capsizing of the trucks in some extreme cases. In this paper we also study the effect the drag reduction devices in the case of a crosswind. Various passive flow control methods have been implemented and are under consideration for the purpose of drag reduction in semi-trailer trucks, as explained in the following paragraphs.

As a part of the research conducted by [3], experiments are conducted to study the wake in various portions of a tractor-trailer. The locations of interest in this study include the region at the back of tractor and front of trailer; and at the back of trailer. The Generic Conventional Model (GCM) is the truck used in this study. The modifications to the truck include a combination of side extenders and the boat-shaped tail. Up to 60% reduction in drag coefficient is noticed at higher yaw angles of around 10°. Various aerodynamics fuel saving devices are tested experimentally by [6] on heavy vehicles, primarily semi-trailer trucks. The results from wind tunnel experiments show that the fairings and side coverings reduce aerodynamics drag by up to 26%. It is found in this study that, among all the configurations tested, the modified truck with a fairing in between cabin and the container and side covers cause a maximum amount of reduction in the drag force.

Different drag reduction devices have been numerically tested by [7] using commercially available computational fluid dynamics (CFD) code STAR-CCM+. The results show that the largest drag reduction is achieved for the under-carriage treatment. The under-carriage treatment includes the addition of side-skirts, wheel covers and complete underbody redesign. The study concluded that it is not beneficial to change the shape of trailer, both aerodynamically and financially as simple addon devices can be added to the tractor-trailer with equivalent effect. Quantitatively, it is shown that side-skits and sealed wheels cause a maximum reduction in the drag coefficient of up to 77%.

As a part of this study [8], several devices to reduce drag force on class-8 tractor-trailers is analyzed. This study is completed via wind tunnel testing of reduced scale models of the tractor-trailers. By combining various aerodynamic devices, a reduction in the drag coefficient of 11.5% in the highway fuel economy is noted. The wind averaged drag coefficient improvements of 22 to 23% are observed for the aerodynamic (modified) trailer. The modifications include the installation of a trailer forebody device and side extenders of variable lengths.

Drag reduction using platooning is analyzed numerically by [9] for semi-trailer trucks. In this study, the effect of varying the offset of the trailing truck with the leading truck and the linear distance between the trucks are investigated in detail. It is found that platooning configuration in which the two trucks are aligned with each other results in most drag reduction. The study also concluded that as the distance between the trucks increases, the drag reduction decreases. Furthermore, it is noted that the stagnation pressure on the rear truck increases and the pressure at the rear of the front truck decreases as the distance between the trucks increases.

Various trailer attachments for the semi-trailer trucks are considered by [10] in search of an optimal configuration w.r.t. maximum drag reduction. CFD software Fluent is used to simulate flow around a Mercedes Benz Actros truck. The results of the simulation show that adding relatively inexpensive addons to the truck is more beneficial as compared with redesigning of the truck and trailer bodies. A substantial increase in the fuel economy of 17.74% is noted. Moreover, regions of low pressure behind the trailer are observed to have reduced contribution to the stability of the truck.

Drag reduction for Heavy Goods Vehicles (HGVs) is the subject of research conducted by [11]. The drag reduction devices used in this research work include a boat tail, a deflector and a fin. As a result of adding these modifications, the drag coefficient decreased by 7.2%. As a part of research conducted in [12-13], a deflector plate of variable height is designed and implemented on a semi-trailer truck. Under the proposed control system, the adjustment to the drag coefficient of the truck can be done online. Quantitively, when in the optimized position, the deflector reduces drag coefficient by 7.4%. An additional 2% reduction in drag coefficient takes place with the implementation of an active system.

Vortex generators are added to the rear of the semi-trailer trucks to investigate their effects of the aerodynamics drag by [16]. The study is carried out numerically using a commercially available CFD package. The authors investigate size, location and geometry of the vortex generators. It is found that by the addition of vortex generators, an increase in the surface pressure can be observed. The vortex generators also reduce the amount of recirculation in the flow in the aft of the vehicle. The results from this study indicate that drag can be reduced by up to 10% by addition of these devices.

The effects of investigating drag reduction by injection of a low velocity jet of air near the floor of a vehicle is investigated by [19]. The effects of flow rate and cavity position and distribution have been investigated. It is found that a major portion of drag force reduction arises from the presence of a cavity in the model. A drag reduction of up to 13% is achieved in this study [19]. Various fuel-saving potentials of drag-reducing devices installed on heavy vehicles are experimentally explored by [20]. The devices include a combination of a cab roof and side fairings, trailer-front fairings, tractor and trailer side-skirts, boat-tails and base-flaps, and gap splitters. It is found that up to 9% of the fuel can be saved in case of a vehicle being driven for an annual mileage of 80,000 miles. It is also found that vehicles operated on motorway speeds save almost twice as much fuel as compared to the vehicles being driver inside the cities.

We explain our methods and modifications to the GCM in the methodology section. In the results and discussion section we present our results while comparing the two configurations both quantitatively and quantitively.

                                                                                                                                                   II.            Methodology

The GCM designed and developed by the National Aeronautics and Space Administration (NASA) exclusively to examine in detail, the semi-trailer truck aerodynamics is used as a validation case for this paper. We analyze the GCM using a commercially available CFD code, Flow Simulation Premium [14]. The Flow Simulation Premium employs cell-centered finite volume method. The κ-ε model with damping functions is used to model turbulence (equations 6-7). The solution algorithm employed is SIMPLE-R (modified). The discretization of the convective terms is obtained by using the Upwind scheme with second level order of accuracy. Meanwhile, the diffusive terms are discretized using the Central-Differencing scheme. Time is discretized using the implicit first order accurate Euler scheme. The Favre-averaged Navier-Stokes equations (equation 3-5) are used to predict turbulent flows.

where,  is a mass-distributed external force per unit mass due to a porous media resistance , a buoyancy  is the gravitational acceleration component along the i-th coordinate direction) and the coordinate system’s rotation , i.e.,  [14]. The subscripts are used to denote summation over the three coordinate directions. While,  and  are source terms. The subscripts are used to denote summation over the three coordinate directions. The simulations are performed to predict three-dimensional steady-state flow around the GCM.

1.       Validation and Verification

The data from [1-5, 15] is used for validate the computational mesh for the present study. A thick flat plate is added on below the wheels of the GCM to simulate the functionality of a road. Slip boundary condition is applied to a portion of the flat plate to prevent a boundary layer from forming at the leading edge of the flat plate while no-slip boundary condition is applied to the portion of the flat plate directly under and after the rear of the GCM to allow for the formation of the boundary layer. This is done to keep the CFD simulation as close to the physically occurring phenomenon as possible.

The slip region is marked by red squares in Fig. 1. Far-field pressure conditions are applied to the boundaries of the computational domain. The computational domain with the boundary conditions applied is shown in Fig. 1. Within Fig. 1, the blue arrow represents the direction of the ambient velocity, with no cross-wind. The ambient velocity is kept at 115.875 mph (51.5 m s-1), similar to [2]. To study the effect of the cross-wind, the ambient velocity direction is varied.

Fig. 1, Computational domain and the boundary conditions

A Cartesian computational mesh with immersed boundary method is employed in the present work. Separate mesh controls are employed to refine the mesh in the areas of interest such as between the road and the truck floor, the gap between the cabin rear and the container front, the wake of the truck etc. to ensure solution accuracy. The mesh around the GCM is shown in Fig. 2, along with the mesh of the GCM superimposed. It should be noted that the results of the present simulations are within 4% of [1-5]. The drag coefficient from the present simulations comes out to be 0.421 while the drag coefficient from literature is 0.398 [1], 0.402 [2], 0.427 [3] and 0.399 [5], respectively. This drag coefficient is for the case with no cross-winds. The drag coefficient from the present simulations with cross-winds at 8° is at 0.759 while the values from experimental data are at 0.725 [3] and 0.700 [5]. This places our simulation results within 7% of the published experimental data, at high cross-winds.

Fig. 2, The computational mesh

The streamwise velocity contour from the present simulation is compared with the contours obtained from [3-4] in Fig. 3. It is clear from Fig. 3 that the results from present simulation are in close agreement with the previously published experimental and numerical simulation results [3-4]. A comparison of the vertical velocity contours between experimental data [3] and the present simulation is shown in Fig. 4. The results are in also close agreement with each other. The location of the plots w.r.t. semi-trailer truck, used in Fig. 3-4 is shown in Fig. 5.

 

Fig. 3, L-R, Streamwise velocity contours from literature. Experimental [3], numerical simulation [4] and present study

Fig. 4, L-R, Streamwise velocities contours from experiments [3] and present simulation

Fig. 5, L-R, Locations of the plots

2.       Mesh and Time Independence Test

The final mesh cell count is selected keeping under consideration that a very negligible change is noticed in the drag coefficient after increasing the cell count. It should also be noted that the amount of core hours required to complete the solution with a higher cell count increase considerably, as explained in the following paragraph.

The mesh employed in the validation of the numerical simulation comprises of 2,113,458 cells. There are 224,741 cells on the truck’s body. As a result of increasing the mesh cell count to 5,213,889 cells with 405,242 on the truck’s body, the drag coefficient improved by less than 1%, as compared to the experimental data [1-3, 5] and the initial mesh. It should be noted that the solution time for the mesh with a higher cell count is more than twice as compared to the selected mesh. Keeping the initial mesh, a transient simulation is performed to ensure that there are no transient effects in the study. It should be noted that the difference in the drag coefficient from the transient analysis and the steady-state simulation is less than 1%. An initial time-step of 6.95e-6 s employed for the transient analysis is gradually increased to 3.39e-1 s, as the solution approached the steady-state. It should be noted that the simulations employing an adaptive time-step are, by their nature, time-step independent.

3.       Modification to the GCM

The proposed plate along with side-skirts is added at the bottom of the truck’s floor, in a single assembly. The added plate completely covers the blunt and drag producing components on the sides and the underside of the truck. The modification is shown in as shown in Fig. 6.

Fig. 6, T-B, GCM and the modified GCM geometry

                                                                                                                               III.            Results and Discussion

The numerical simulations predict that the baseline configuration produces 8.36% more drag than the modified configuration. This improvement is noted at the highway cruise speed and without any crosswinds. At the higher yaw angle, the baseline configuration is predicted to have 6.3% more drag as compared to the modified configuration. Thus, the addition of a simple and cost-effective passive flow control device i.e. a flat plate with side skirts, has the potential to significantly improve heavy vehicle fuel economy at a variety of flow conditions.

The qualitative results from the numerical simulations are presented in this section. The three major fluid properties of interest discussed include vorticity, pressure and velocity. At first, we present the results at zero-degree yaw angle followed by a discussion at the higher yaw angle.

1.       Reduced Vorticity

The contours showing vorticity magnitude superimposed by the velocity streamlines are plotted in Fig. 7. The contours are plotted along the centerline of the GCM. The direction of the flow is from the left to right of the reader. The scale within Fig. 7 ranges from 0 – 1,000 s-1. The differences in vorticity are quite clear. The large amounts of flow separation in the GSM without modifications is evident at two important locations. The first region where the flow separates is at the position where the tractor body meets with the trailer section. A large vortex of magnitude ~500 s-1 can be seen in the baseline GCM while the same vortex is absent in the modified GCM. The addition of the flat plate below the floor of the GCM forms a continuous wall below the modified tractor-trailer thus eliminating the blunt backward-facing step like geometry from the flow path. This results in the elimination of the vortex, which in turn leads to reduced drag force.

The second location where vortex formation takes place below the baseline tractor-trailer can be seen at the locations where fluid interacts with the blunt bodies like axle, differential and suspension assemblies. In the baseline tractor-trailer the vortices of magnitude ~1000 s-1 can be seen in the wake of the rear axles. While at the same locations in the modified GCM, no such vortices can be observed. The addition of flat plate removes, shields, these bluff bodies from the flow path thus contributing towards reducing the drag force. It should be noted that the various small vortices observed in the modified configuration have significantly less strength as compared to the one’s observed in the baseline configuration.

This effect can be clearly observed in Fig. 8. Within Fig. 8, the normalized vorticity is plotted against normalized trailer length. The starting point on the x-axis represents the point where the tractor body ends and the trailer body starts. The line is plotted horizontally, along the center of the axles of the tractor-trailer. The two rises in the vorticity for the baseline configuration are the two axles. It should be noted that these axles are shielded from the oncoming airflow by addition of the flat plate in the modified configuration, hence the missing peaks in vorticity.

Fig. 7, T-B, GCM and the modified GCM

Fig. 8, T-B, GCM and the modified GCM

The vorticity around the axis normal to the road, superimposed by velocity streamlines is shown in Fig. 9. The scale in Fig. 9 ranges from -500 to 500 s-1. The images in Fig. 9 are taken in a plane along the center of the axle of the two configurations and parallel to the road. It can be seen from Fig. 9 that the addition of the flat plate below the eliminates the pair of large and high energy vortices from behind the tractor body in the modified configuration. The magnitude of vorticity is also low for the modified configuration as compared to the baseline configuration at the rear of the trailer, specially near the axles and suspensions components. The reduction in size and strength of unwanted flow separation leads to less drag force on the modified GCM as compared to the baseline GCM, even in the case of a crosswind.

Fig. 9, L-R, GCM and the modified GCM

2.       Reduced Stagnation Pressure and Pressure Difference

In terms of the pressure field around the two configurations, three major differences can be observed between the baseline and the modified configurations.

The contours of relative pressure are shown in Fig. 10. The zoomed in view of the area of interest is shown in Fig. 11. The location of the contours is same as in Fig. 7. The scale in Fig. 10-11 ranges from -500 – 200 Pa. The bluff bodies, rear axles in this case, can be clearly seen to have their own stagnation regions marked by high relative pressure as these bodies are directly in the path of the flow. This phenomenon contributes towards the pressure drag. While the pressure distribution in the same regions in the modified GCM configuration is much more uniform. This reduces the pressure drag component of the total drag and thus contributes to the overall drag reduction observed in the case of the modified GCM configuration.

It can also be noted from Fig. 10 that the region of relatively high-pressure at the aft and below the centerline of the baseline configuration is far away from the body of the GCM. Meanwhile, in the configuration with the plate installed at the bottom of the tractor-trailer floor, the high-pressure region moves closer to the body of the GCM. This helps to reduce the over all pressure difference between the rear and front of the GCM. Thus, contributing to the reduction in drag force. As shown in Fig. 12, more pressure for the modified configuration indicates less pressure difference between front and rear of the tractor-trailer assembly, resulting in less drag force.

The pressure difference between the top and bottom of the GCM can also be clearly seen in Fig. 10. The higher-pressure difference creates a net lift force which is more in the modified configuration as compared to the baseline configuration. This phenomenon has two advantages. One being the less load on the suspension of the modified GCM, increasing the service life of the vehicle’s suspension. The other one being, the modified GCM being able to carry additional load, equivalent to the lift produced, while consuming less amount of fuel than the baseline configuration.

As is the case with no cross-winds, the addition of the plate at the bottom of the tractor-trailer completely eliminates the stagnation zones at the rear of the trailer in the case with the cross-winds as well. The pressure difference between rear and the front of the tractor-trailer assembly is also reduced. The pressure difference is compared using the relative pressure and is shown in Fig. 13. The scale for Fig. 13 is same as is for Fig. 10-11.

Fig. 10, T-B, GCM and the modified GCM

Fig. 11, T-B, GCM and the modified GCM

Fig. 12, The relative pressure plots

Fig. 13, The relative pressure plots

3.       Improved Momentum of the Flow

The contours of the component of velocity in the direction of flow is presented in Fig. 14. The scale used in Fig. 14 ranges from -55, dark blue, to -1 m/s, red, which is similar to scale chosen by [3]. The location of contours in Fig. 14 is in the wake of the trailer. It can be seen that magnitude of the component of velocity in the direction of flow is more in the case of modified configuration as compared to the baseline configuration. This is indicated by the less amount of region colored red in the wake of the trailer. The flow on the underside of the GCM in the modified configuration is able to maintain significantly more momentum in the desired direction. This less amount of restriction to the oncoming flow is because of the plate that has been added in the modified configuration. The component of velocity in the direction of flow is plotted in the wake of the semi-trailer truck in Fig. 15, confirming the phenomenon. The negative sign within Fig. 15 indicates the magnitude of velocity in the negative – z direction i.e. the direction of flow.

Fig. 14, L-R, GCM, modified GCM

Fig. 15, Significantly more momentum in the flow for the modified configuration

The streamwise velocity superimposed by streamlines is shown in Fig. 16. The scale within Fig. 16 ranges from -86, blue, to 30 m/s, red. The location of the images shown in Fig. 16 is same as that of Fig. 9. The streamwise flow within Fig. 16 is from the bottom to top. It can be said that the addition of the plate below the tractor-trailer delays stall. It can be seen form Fig. 16 that the size of vortices in the case of the modified configuration is significantly less in the wake of the right wheel assembly, wake of the trailer and around the bluff bodies. The vortex beside the right wheel of the baseline configuration for example, has size of ~0.06w where w is the width of the trailer. Meanwhile, at the same location, the vortex is nonexistent in the modified configuration and the flow is directed in the streamwise direction. The vortices in the wake of the trailer beside the right wheel and the bluff bodies i.e. suspension and axle assemblies, are almost completely eliminated as a result of the proposed modification. Meanwhile, the vortices in the wake of the tractor are also significantly reduced. The bulk of the flow in the wake of the tractor-trailer can be clearly seen to have been directed towards the streamwise flow. This reduced flow separation results in less drag force in the modified configuration.

Fig. 16, L-R, GCM, modified GCM

Conclusion

In the present study we propose the addition of flat plate with side-skirts plate to shield bluff bodies present on the underside of the semi-trailer trucks from the oncoming flow. CFD simulations predict a drag reduction of 7.71% in the modified GCM as compared to the baseline design. The CFD simulations also predict that the addition of flat plate would also mitigate the drag force in case of cross-winds. A reduction of up to 5.92% reduction of drag is predicted by the CFD simulations for the case with a cross-wind. The reduction in drag force in the modified GCM configuration as compared to the baseline configuration can be attributed to the reduced vorticity, a reduced pressure difference between front and rear of the truck and in general a more streamlined flow in the underside of the modified truck. The reduction in drag force would directly result in a reduced fuel requirement by the vehicle’s engine. This research has special significance for author’s home country where a number of light and heavy-duty commercial vehicles would benefit from the implementation of the proposed changes to the semi-trailer truck body work. This research can easily be expanded to include smaller vehicles like the 4-6 wheelers and even 18-wheeler heavy vehicles, which form the back bone of transport within the many countries throughout the world.

References

[1] Dale Satran, “An Experimental Study of the Generic Conventional Model (GCM) in the NASA Ames 7-by-10-Foot Wind Tunnel,” Proceedings of the United Engineering Foundation Conference on The Aerodynamics of Heavy Vehicles: Trucks, Buses, and Trains, Monterey, CA, December 2002.

[2] Pointer W.D., Sofu T., Weber D., “Commercial CFD Code Validation for Heavy-Vehicle External Aerodynamics Simulation”, Lecture Notes in Applied and Computational Mechanics, vol 19. Springer, Berlin, Heidelberg, 2004.

https://doi.org/10.1007/978-3-540-44419-0_42

[3] Heineck J.T., Walker S.M., Satran D., “The Measurement of Wake and Gap Flows of the Generic Conventional Truck Model (GCM) using Three-Component PIV”, Applied and Computational Mechanics, vol 19, Springer, Berlin, Heidelberg, 2004, pp 173-184.

https://doi.org/10.1007/978-3-540-44419-0_19

[4] Wood, R., "Reynolds Number Impact on Commercial Vehicle Aerodynamics and Performance," SAE International Journal of Commercial Vehicles, Volume 8(2), 2015, pp 590-667.

https://doi.org/10.4271/2015-01-2859

[5] Bruce Storms, Dale Satran, James Heineck and Stephen Walker, “A Study of Reynolds Number Effects and Drag-Reduction Concepts on a Generic Tractor-Trailer”, 34th AIAA Fluid Dynamics Conference and Exhibit, Portland, Oregon, July 2004.

https://doi.org/10.2514/6.2004-2251

[6] Harun Chowdhury, Hazim Moria, Abdulkadir Ali, Iftekhar Khan, Firoz Alam and Simon Watkins, “A study on aerodynamic drag of a semi-trailer truck”, Procedia Engineering, volume 56, pp 201 – 205, 2013.

https://doi.org/10.1016/j.proeng.2013.03.108

[7] Christoffer Hakansson and Malin J. Lenngren, “CFD Analysis of Aerodynamic Trailer Devices for Drag Reduction

of Heavy-Duty Trucks”, Master’s Thesis, Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden 2010

[8] Ronald E. Schoon, “On-Road Evaluation of Devices to Reduce Heavy Truck Aerodynamic Drag”, in the Proceedings of SAE 2007 Commercial Vehicle Engineering Congress & Exhibition, SAE Technical Paper number 2007-01-4294, 2007

https://doi.org/10.4271/2007-01-4294

[9] Wael Mokhtar, Kyle Dinger, Shardul Kachare, Md Razib Hossain and Colin Britcher, “A CFD Study Investigating Drag Reduction for Different Offset and Inline Distance for Two Trucks Platooning”, Journal of Energy and Power Engineering, volume 14, pp. 143-150, 2020

https://10.17265/1934-8975/2020.05.001

[10] Anu R. Nair, Fred Barez, Ernie Thurlow, Metin Ozen, “Method to Reduce Drag Coefficient for Fuel Efficiency in Semi-Truck Trailer and Trailer Stability”, in the Proceedings of ASME 2018 International Mechanical Engineering Congress and Exposition, IMECE2018-86584, 2018

https://doi.org/10.1115/IMECE2018-86584

[11] Anil K. Madhusudhanan, Daniel Ainalis, Xiaoxiang Na, Isabel Vallina Garcia, Michael Sutcliffe, David Cebon, “Effects of semi-trailer modifications on HGV fuel consumption”, Transportation Research Part D: Transport and Environment, volume 92, 102717, 2021

https://doi.org/10.1016/j.trd.2021.102717

[12] Georgios Papageorgiou, Jason Barden, James F. Whidborne and Kevin P. Garry, “Extremum Seeking Control for Truck Drag Reduction”, in the proceedings of 2018 UKACC 12th International Conference on Control (CONTROL), Sheffield, UK, 2018

https://doi.org/10.1109/CONTROL.2018.8516808

[13] Jason Barden, “Active Aerodynamic Control of Heavy Goods Vehicles”, PhD Thesis, Cranfield University, United Kingdom, 2013.

http://dspace.lib.cranfield.ac.uk/handle/1826/9293

[14] Mentor Graphics Corporation, Flow Simulation Technical Reference, 1st ed., 2016, pp 4,81.

[15] Bruce L. Storms, Dale R. Satran, James T. Heineck, Stephen M. Walker, “A Summary of the Experimental Results for a Generic Tractor-Trailer in the Ames Research Center 7- by 10-Foot and 12-Foot Wind Tunnels”, NASA/TM–2006-213489, July 2006

[16] Abdellah Ait Moussa, Justin Fischer, and Rohan Yadav, “Aerodynamic Drag Reduction for a Generic Truck Using Geometrically Optimized Rear Cabin Bumps”, Journal of Engineering, Volume 2015, paper number 789475, 2015

https://doi.org/10.1155/2015/789475

[17] David Manosa, lvas-Kjono, Josh Dawson, Eduardo Molina, Travis Carrigan, “Understanding Tractor Trailer Aerodynamics Using a Time Accurate Approach”, Webinar, https://info.pointwise.com/understanding-tractor-trailer-aerodynamics-using-time-accurate-approach, 2020

[18] Anderson, John D. Jr, Aircraft Performance and Design, 3rd ed., Tata McGraw-Hill Edition, New Delhi, 2010, pp 58

[19] Jeff Howell, Andrew Sheppard and Alex Blakemore, “Aerodynamic Drag Reduction for a Simple Bluff Body Using Base Bleed”, SAE Paper 2003-01-0995, March 2003

[20] Zulfaa Mohamed-Kassim, Antonio Filippone, “Fuel savings on a heavy vehicle via aerodynamic drag reduction”, Transportation Research Part D: Transport and Environment, volume 15, issue 5, 2010, pp. 275-284