Aerodynamic Analysis of Cooling Airflow for Different Front-End Designs of a Heavy-Duty Cab-Over-Engine Truck.
Reducing the harmful effect of vehicles on the environment is a key demand for vehicle manufacturers today. Constantly increasing oil prices and stricter emission legislation are also important factors for the vehicle industry to consider. There are many factors affecting the fuel consumption of a vehicle. One of the major ones is the aerodynamic resistance. In fact, above 80 km/h, the aerodynamic drag is the most significant resistance for a truck driving at constant speed, according to Logdberg . Hence the importance of working with the aerodynamic properties of trucks is apparent.
Cooling drag is a part of the total aerodynamic drag, which arises as a consequence of air flowing through the grille, cooling module, and the irregular under-hood area. Hallqvist  states that the trend is to constantly increase cooling requirements, both as a consequence of increased engine power, and also for the Air Conditioning (AC) system, exhaust after treatment systems, and auxiliary systems to work properly. Hallqvist  further states that the cooling drag component can be up to 8% of the total drag, which is a substantial amount. At the same time as the cooling demand is increasing, there are also discussions regarding how to be able to shut some cooling flows off, for example at highway speeds, when the full cooling capacity is not needed. In such situations, fuel consumption may be lowered as the cooling airflow is cut off.
The benefits of restricting the amount of cooling airflow entering the under-hood area are not only limited to reducing the aerodynamic drag. Pfeifer  states that the cold start behaviour; the noise level related to the combustion engine; and the fact that the cooling module is better protected from external, possible harmful objects, can be improved with a grille shutter system. Also, the cooling performance can be improved in some cases by directing the flow towards the cooling package, via a duct.
The European Union (EU) recently introduced a new directive, which would allow for slightly longer vehicles, the extra distance to be used for aerodynamic and safety purposes. The directive, (EU) 2015/719, states that extension of the total length of the vehicle is allowed for both the front and the rear of the vehicle, in order to make room for these improvements. There is however a stipulation in the directive that if the added devices extends 500 mm or if the cab shape is elongated beyond the limits for Directive 96/53/EC, these vehicles need to be type approved before they can be released onto the market .
The aim with this study was to investigate the potential of a Soft Nose (SN) concept from an aerodynamic and thermal management point of view. The cooling airflow and the flow through the under-hood was the main focus. One aspect was to investigate at what opening percentage of the grille area an acceptable heat rejection rate of the radiator was obtained, for a highway-driving scenario. Comparisons of the cooling airflow and aerodynamic properties of a cab-over-engine (COE) and SN model were performed and an analysis of the under-hood layout for these models was carried out.
In this section, a description of the different vehicle models and their modifications are presented.
Modelling of Closed Grille
Theoretically, it would be interesting to investigate the aerodynamic benefits of shutting off the entire cooling airflow with a completely sealed front. This would give an idealized figure of the possible reductions in aerodynamic drag.
However, this scenario would never exist in reality; depending on the design of the grille shutter device there will be some leakage into the under-hood area. Hence it is of more value to investigate a closed grille configuration with some degree of possible inflow through the grille. The intention of a sealed front is, naturally, to be able to reduce the aerodynamic drag when the cooling situation permits. But, what must also be considered is to ensure sufficient cooling capacity for the specific operating condition. In theory, a scenario where the truck is driving at a constant highway speed (80 km/h) on a level road would be the situation where the least airflow through the under-hood would be needed to ensure sufficient cooling. This however does not necessarily mean that it would be possible to reduce the cooling airflow to zero. The aim of this paper was to determine the minimum opening percentage of the grille area to ensure sufficient cooling capacity for the given driving condition. It was, however, decided to also simulate the entirely closed grille to be able to compare the differences in terms of aerodynamic properties.
Two methods for partially opening the grille were evaluated: one where horizontal slots (HS) were opened in the grille area and one where the grille area was continuously opened (CO) from the top, with the same total percentages as with the slots. It should be noted that, in this context, the expression "grille area" refers to the area of the grille where the Charge Air Cooler (CAC) has been projected onto the front, creating an area where the two opening approaches were applied. The two methods can be seen in Figure 1; the rectangular blue and white areas denoting the grille area.
It was found from simulations performed during this study that an opening of 17.5% of the COE grille area was sufficient to meet the target heat rejection for the radiator in the specified operating condition. This opening percentage was then used for evaluating the SN model as well. More information about the simulation set-up can be found in the numerical setup chapter and the results are further discussed in the results section. It should be mentioned that the parameter for which the grille opening percentage was designed was the radiator Top Tank Temperature (TTT).
In order to simplify the calculation of the open area of the cooling module against the grille opening, the geometry between the grille and cooling module was removed, in order to obtain a clean area between these two parts.
A range of the two different cab shapes, grille-opening configurations and yaw angles were evaluated. In addition, all configurations were also evaluated with active heat exchangers to investigate the differences between a COE and SN model in terms of cooling capacity. Table 1 shows an overview of the different configurations evaluated in this paper. The COE model run with partially closed grille was evaluated with different grille percentage openings and opening strategies (HS or CO). Note however that all the simulations from this sub-study are not shown here, only the final choice of design where an acceptable level of TTT was achieved.
Considering the yaw angles tested, two different angles were evaluated: 0[degrees] and 5[degrees]. The windward side in the 5[degrees] simulations was to the right-hand side, seen from the drivers' perspective.
In the Grille Configuration column, 'open' refers to a conventional grille opening with a grille pattern, while 'closed' means that the entire grille area is sealed (100%) or partially opened (17.5% HS). For the heat exchanger models, the 'active' choice denotes the inclusion of heat transfer in the simulation model, as opposed to the 'inactive' choice which states that no heat transfer was accounted for.
COE Vehicle Model
The baseline vehicle model used for the simulations was a fullsize 4x2 tractor model, equipped with a full aerodynamic package including roof deflector, cab side-extenders and chassis skirts. The cab was of COE design. The trailer fitted to the vehicle was a standard semi-trailer of 16.5 m, without any aerodynamic devices. This vehicle combination was evaluated in three main configurations; one with an open grille, one with a completely sealed front and one with different grille design strategies and opening percentages of the grille area. The open and completely closed configurations can be seen in Figures 2 and 3, respectively. In Figure 2, the open-grille configuration, the grille pattern is visible. In Figure 3 on the other hand, the grille pattern has been removed and replaced with a solid, impermeable surface to define the closed-grille configuration.
SN Vehicle Model
The SN model was basically an elongation in the longitudinal direction of the grille area of the COE cab. The nose was 200 mm longer than the COE front. The corner radius of the nose was larger than for the COE configuration. The SN cab is shown in Figure 4.
As with the COE model, the SN model was evaluated with opened and closed grille plus the final design of the partially opened-grille configuration, obtained from the simulation loop for the COE model. The geometries of the SN models with opened and completely closed grille, can be seen in Figures 5 and 6, respectively.
The investigations in this paper were performed using Computational Fluid Dynamics (CFD). The Computer Aided Design (CAD) cleaning of the vehicle geometry and initial surface meshing was carried out using ANSA by Beta CAE Systems , while the vehicle model was wrapped, surface meshed, volume meshed and run in STAR-CCM+ by CD-adapco .
Two different kinds of simulations were performed; purely aerodynamic simulations with inactive heat exchanger models; and aerodynamic simulations with active heat exchanger models. The reason for performing both kinds of simulations was to evaluate the influence of heating the air on the aerodynamic drag. The strategies for the two different simulation types are described in the coming sub-sections.
The entire meshing and simulation method was made in accordance with the results from a previous study by Martini et al. , where comparative studies between experimental and numerical results of external aerodynamics and cooling performance were performed. In this paper however, the methods for external aerodynamics and cooling performance have been combined in order to take into account cooling airflow within the external aerodynamics simulations.
The volume meshes for the simulations were hexahedral dominant and created using the STAR-CCM+ Trimmer model . For the surfaces where mainly attached flow was expected, a number of prisms layers were created closest to the vehicle surface in order to account for the viscous effects against the vehicle surfaces. For surfaces where significant separated flow was expected, one prism layer of 1 mm was generated.
A set of volume sources were used to locally refine regions in the flow field where higher resolution was desired, for example in the base wake, around the entire vehicle, below the vehicle, and around the rear-view mirrors. These refined flow regions were in-line with what is recommended in the SAE J2966 Standard . In addition, the STAR-CCM+ Trimmer Wake Refinement model was used , to enable a finer mesh growing downstream of selected surfaces to refine wake areas along the vehicle. A mesh dependency study was performed; it can be further viewed in the study by Martini . An example of the volume mesh is displayed in Figure 7.
The meshing approach for the 5[degrees] yaw simulations was slightly different. The calculation domain was rotated 5[degrees] and then the mesh was regenerated with this new orientation of the domain as the reference co-ordinate system. This approach was used in order for the free stream to flow axially in and out of the computational cell.
This section describes the methodology used for the simulations performed in this paper.
External Aerodynamics The external aerodynamics simulations were run based on a Reynolds-Averaged Navier-Stokes (RANS) approach, according to the results obtained in the study by Martini et al. . The model chosen was the standard ks-model with a realizability factor denoted as the Durbin Scale Limiter. This turbulence model was applied together with the STAR-CCM+ Two-Layer approach . The intention in this study was, even though the Two-Layer approach was used, that the y + values should be less than 1 in most flow regions. The boundary layer was assumed to be turbulent throughout the entire simulation model.
The truck and trailer models analyzed in this paper were evaluated with a control volume boundary condition to simulate the open-road condition; i.e. the calculation domain was considerably larger in order not to affect the simulation results. The tunnel inlet was modelled as velocity-inlet, outlet as pressure-outlet, and the domain walls were modelled as walls with a slip condition. The Reynolds Number in the simulations, based on the vehicle width, was 4 million.
Fan Model. The Multiple Reference Frame (MRF) model was used to model the fan rotation in the simulations. This model is a frozen rotor method. The user defines a control volume around the fan, and the model simulates the effect of rotation on a stationary mesh inside this control volume. The choice of using the MRF method instead of a sliding-mesh approach for the fan rotation was made according to the results from the study by Gullberg and Sengupta . The size of the control volume, or MRF region, is a sensitive parameter for the results. The choice of MRF region size was made in accordance with results from Gullberg et al. . In the simulation model, the cooling fan was set to a 'wind-milling' condition of 250 rpm throughout the entire range of simulations.
Modelling of Cooling Module. The cooling package of the vehicles in the simulations consisted of an AC condenser, a CAC and a radiator. The heat exchanger surfaces were modelled as porous media, in order to avoid the extremely large meshes that would be necessary to resolve the entire structure of a detailed heat exchanger.
The porous media formulation included parameters, which defined the effects of the porous media on the flow field downstream the heat exchanger. The parameters are included in source terms in the momentum equation. The parts that should be modelled as porous media are assigned their own flow region, where co-efficients of flow resistance are added. These co-efficients are then used to calculate the porous source term in the momentum equation.
The porous source term is defined as in Equation 1.
[f.sub.p] =-([P.sub.v] + [P.sub.v] |v|)- v Eq. (1)
In Equation 1, [P.sub.v] and [P.sub.j] are the viscous and inertial resistance tensors, respectively. The values of the inertial and viscous resistance were obtained by pressure drop measurements over each heat exchanger component and the values were obtained from the manufacturer.
Operating Conditions. The operating condition was intended to resemble a highway driving condition. Therefore, the inlet velocity of the air entering the calculation domain was set to 80 km/h, which is a common operating speed for trucks in Europe. The input values of the parameters needed for the heat exchanger models, for example the target heat rejection of the radiator, the mass flow of air entering the CAC and the volume flow of coolant through the radiator, were extracted from test data for the specific driving condition.
The temperature of the air entering the calculation domain was set to 10 [degrees]C, corresponding to the average daily temperature in Europe .
External Aerodynamics with Active Heat Exchanger Models This section describes the specific settings for the external aerodynamics simulations with active heat exchanger models, which were used in addition to the ones described previously in the Simulations section.
Heat Exchanger Model. To include cooling capacity analysis, a heat exchanger model was included in the simulation. The STAR-CCM+ Actual Dual Stream heat exchanger model was used to evaluate the heat exchange between the hot and cold cores. The hot core was defined as the part of the heat exchanger where the coolant and charge air was flowing, whereas the cold core was defined as the part where air passed through the heat exchanger to cool the coolant and charge air. The heat exchanger model also required information about the specific heat exchanger components that were used. Files including the mass flow through the hot and cold cores, and the corresponding heat transfer rate were therefore provided to the simulation model.
The fan model used to model the fan rotation in the simulation was performed in the same manner as for the external aerodynamics simulations without heat exchanger model.
Definition of Radiator Top Tank Temperature (TTT). The radiator TTT value was defined as the temperature of the coolant leaving the engine, but before entering the cooling module. The unit of the TTT is [degrees]C. The value of the TTT is a surface average of the temperature at the radiator inlet. The inlet tank is assumed adiabatic in the simulations.
This section presents the results from the different simulations that have been performed. Further analysis of the numerical results will be performed in coming sub-sections of the paper.
SN vs. COE
The results from the simulations showed that the SN model was to be preferred from a drag perspective. The magnitude of the drag reductions were however modest; a maximum drag reduction of almost 6% was obtained for the closed grille configuration at 5[degrees] yaw.
The numerical results of the difference in drag co-efficient (CD) between the SN and COE models can be seen in Figure 8. In the graph, the SN model with open grille has been compared to the corresponding COE model with open grille and the SN model with closed grille has been compared to the corresponding COE model with closed grille. The same method was employed for the 0[degrees] and 5[degrees] yaw configurations.
Two main trends can be discerned from Figure 8: The first is that the
SN model was showing larger drag reductions compared to the COE model at 5[degrees] yaw; the second being that the closed grille configuration showed a generally higher reduction in drag for both yaw angles.
Cooling Airflow Characteristics
Part of the scope of this paper was to investigate the effect of a closed grille for a COE and SN, to see whether there was any difference in the drag reduction. Figure 9 shows the percentage difference between open and closed grille for the COE and SN models, referred to as cooling drag. The cooling drag is defined as in Equation 2.
[mathematical expression not reproducible] (2)
It was seen when looking at the data in Figure 9 that there was a considerable reduction in aerodynamic drag when closing off the entire cooling airflow. This effect was seen for both types of cab design. The cooling drag proportion was larger for the SN cab design. The amount of cooling drag was over 12% for the SN model, while it was 9.5% for the COE model. From these numbers it can be suspected that the under-hood flow for both the models was not ideal. This will be investigated further on in the paper.
A trend, which was clear when analyzing the results, was that the amount of air flowing through the heat exchangers was significantly reduced for the SN model. This was seen for all cooling module components; AC condenser, CAC and radiator (RAD). Figure 10 shows the data for the configurations at 0[degrees] yaw. The trends of the mass flows were however similar at 5[degrees] yaw. The absolute values of the mass flows over the heat exchangers are taken as the sum of the mass flow over each cell of the heat exchanger core.
It was seen that the amount of air flowing through each cooling module component became smaller and smaller as it moved downstream; the largest difference being for the radiator. Leakage of air was thought to be the main reason for this.
However, even though the mass flows through the heat exchangers were higher for the COE model, the mass flows through the grille were significantly lower for the COE model compared to the SN model. The mass flow of air through the grille for the COE model was approximately 45% lower than for the SN model. This phenomenon can be explained by viewing Figure 11, which shows the velocity magnitude in a z-plane cut through the grille and cooling package area.
Figure 11 confirms that the velocity magnitude was much higher at the grille inlet for the SN model. There was, however, also a strong effect of leakage of air around the cooling module, as seen in the lower picture in Figure 11. Air was leaking around the cooling module instead of flowing through it, hence the lower percentages of mass flow through the cooling module for the SN model in Figure 10.
Comparing the corresponding COE and SN models with inactive vs. active heat exchangers, it was seen that the velocity magnitudes were more or less identical between the two heat exchanger approaches. The values of the mass flow of air through the grille only showed deviations of a maximum of 3%; the inactive heat exchanger simulations showing slightly higher airflow rates.
It has been confirmed that there was a considerable level of leakage of air in the under-hood area for the SN model, where air was leaking on the sides of the cooling module. It was also seen that there was a significant leakage under the cooling module; making a substantial amount of the air entering the grille exit under the vehicle. Figure 12 shows the velocity magnitude in an x-plane just before the cooling module.
Analyzing Figure 12 it was seen that in the lower part of the pictures, the flow field differed substantially between the two models. The COE model showed very low velocities in this region, while the SN model showed the highest magnitudes of velocity in this region. This strong downward flow towards the vehicle underbody can also help to explain the lower amount of air flowing through the cooling module for the SN model.
So far, it has been seen that the potential for reduced aerodynamic drag by the use of a SN model is apparent; however it has also been discovered that the under-hood area is not ideally designed for an elongated nose. This has been confirmed by the magnitudes of cooling drag and the level of mass flow of air through the heat exchangers for the two cab designs. The single factor which was changed with the SN model was the nose exterior itself; the under hood area was unmodified, remaining identical to the COE model. Since there was considerable leakage of cooling airflow around the cooling package, it would be interesting to investigate a ducted cooling air channel, guiding the flow from the grille area to the cooling module. Another possible continuing analysis would be to move the entire cooling module forwards, to obtain a similar layout as for the COE model, possibly also adding exit ducts for the cooling airflow after the fan.
When analyzing the flow field around the vehicles, it was seen that the overall flow behaviour was similar for the COE and SN models. Since the reduction of CD with the SN was very small, 0.2% for the simulation in 0[degrees] yaw, there were no major differences with respect to separated regions between the models. There was, however, quite substantial differences in the flow field when comparing the open and closed grille configurations. Here it was possible to detect considerable changes in the extension and position of the separated regions. Visualizing isosurfaces of total pressure equal to zero gives an indication of regions in the flow field where large energy losses occur. Figure 13 shows the SN model with open and closed grille.
Comparing the figures of the isosurfaces of Figure 13 it can be seen that the flow losses at ground level were reduced considerably when closing off the cooling airflow. This effect was very prominent also in the trailer area; the losses at ground level were reduced to a large extent for the closed-grille configuration. It was also seen that the two distinct separated regions originating from the split line and lower front of the open-grille configuration was reduced considerably for the closed-grille configuration.
Figure 14 shows a front view of the COE model, with isosurfaces for both open- and closed-grille configurations. The lighter shaded isosurfaces represent the open-grille configuration and the darker surfaces correspond to the energy losses for the closed grille configuration.
From Figure 14 it can be seen that when closing off the coolingairflow the extension of the energy losses decreases. The flow field stayed attached to the cab surface for the closed-grille configuration and hence lower losses were seen. One explanation for the better-attached flow for the closed-grille configuration was due to the fact that the split line was covered, hence reducing the separated region originating from the gap between the upper and lower part of the cab. This effect was clearly seen by comparing this area of both models in Figure 13.
Active vs. Inactive Heat Exchanger Models
One aspect of this paper was to investigate whether the usage of active heat-exchanger models influenced the results of CD calculations. Figure 15 shows the percentage difference in CD between the respective COE and SN models for the open-grille configurations.
As seen in Figure 15 the difference in CD was below 0.15% throughout the simulations. Hence it can be concluded that the presence of the heat exchanger models did not affect the CD calculations in a significant way in these cases. There were however, small changes in mass flow though the heat exchangers. As was seen in Figure 10, the differences in mass flow between the cooling package components were larger for the active heat-exchanger simulations. An explanation for this could be that the temperature changes of the surrounding air, which are accounted for in the active heat-exchanger simulations, affects the density of the air. As temperature increases, the density decreases, which in turn results in the mass flow also decreasing given that the velocity of the flow is the same. Due to the fact that the temperatures in the under- hood area increased for the SN model compared to the COE model, there would also be larger differences between the SN and COE models run with active heat-exchanger models. The inactive heat-exchanger model was based on a constant density approach and hence no changes in this parameter were accounted for.
For the simulations run with active heat-exchangers, it was possible to compare the cooling performance parameters to investigate the differences between a SN and COE model. Table 2 shows chosen parameters for the two cab designs.
It can easily be seen from the numbers in Table 2 that the cooling capacity was lower for the SN model compared to the COE model. The TTT values were in the region of 9 [degrees]C higher than for the COE model and the difference was consistent both for 0[degrees] and 5[degrees] yaw. The absolute values of the TTT values should not be seen as temperatures that would be obtained in real life. In the simulations performed here, the thermostat has been treated as fully open, which it would not be in reality for this particular driving condition. The thermostat opens fully at 82 [degrees]C, which means that for both these models, the TTT values would be about 82 [degrees]C but with different flow rates of the coolant in the radiator. The TTT values in Table 2 however states the difference in available cooling airflow for the two vehicle models.
The main reason for the SN model having worse cooling capacity was due to the massive leakage of air around the cooling module, which resulted in lower mass flows of air through the cooling module. This was discussed earlier in the results section. Since the driving condition analyzed in this paper is not a critical cooling situation, the TTT values were still within reasonable limits. However, if the cooling situation were more critical, it may be that the SN model may not fulfil the required amount of cooling. This was, however, not a part of the scope in this paper.
Partially-Opened Grille Configurations with Active Heat-Exchanger Model
When running a simulation of a COE model with completely sealed front with active heat exchanger models, no satisfactory cooling situation could be found. The TTT values were considerably higher than the permitted levels; these simulation results have been excluded from the paper. Due to the nonsatisfactory cooling airflow results, it was decided to investigate what minimum percentage opening of the grille area would result in TTT values within the permitted range. As stated in the Case Description section, a number of simulations were run with different grille designs (HS and CO) and degrees of opening of the grille area.
The results obtained from the simulations with partially-opened grille showed that rather small differences in opening percentage had a large impact on the radiator TTT values. There was also a significant difference between the two methods: HS resulted in a lower opening percentage being acceptable compared to the CO approach. For the simulations performed here, the fan speed was kept at a constant level of 250 rpm, which is a typical speed for a fan in a "wind-milling" condition. The wind-milling fan condition is defined as the speed of the fan when the fan is disengaged and its rotation is dependent on the oncoming airflow and the slipping of the fan clutch. No studies of the effect of different fan rotations have been performed in the analyses here.
Figure 16 shows [delta]CD values and mass flow of air through the grille area as a function of the opening percentage of the grille for the HS and CO cases. The [delta][C.sub.D] values are calculated as the difference between each HS and CO configuration compared to the COE model with conventional grille design at 0[degrees] yaw, with active heat-exchanger models.
It was seen that the mass flow through the grille area showed a rather linear increase as the grille opening increased. The [C.sub.D] values showed more diverse behaviour; it was however, clear that the CO grille cases had a flatter curve than the HS. It can hence be stated that the two grille opening approaches resulted in different behaviour of the air distribution around the vehicle.
The parameter of most interest for this sub-study was the radiator TTT value. As opening up the grille area, and letting more air into the engine bay, it was expected that the TTT values would be lowered. A coolant temperature of 100 [degrees]C entering the radiator was considered as an acceptable level and this was the target for the sub-study. Figure 17 shows the TTT values and the corresponding mass-flow rates of air through the radiator for the two different grille-opening approaches.
It can be seen from Figure 17 that the TTT values for the CO cases were consistently higher throughout the entire range of grille opening percentages. The difference in values between the approaches was also quite consistent between the simulations; however the curves of mass flow through the radiator diverged with increasing opening percentage of the grille, similar to the mass flow through the grille area shown in Figure 16, but with a more prominent effect here.
According to the results from this sub-study, it was beneficial from a thermodynamic point of view to have horizontal slots, in terms of TTT values only. For the HS cases, TTT values were below the target level at 17.5% grille opening, while the corresponding value for the CO cases was 25%. However, the aerodynamic gains were very small; the CD values for the HS with 17.5% were very similar to the 25% case for the CO arrangement. Hence the CO method would be preferable from a purely aerodynamic perspective. But, during the analysis of the results, it was found that this method resulted in a very non-uniform temperature field in the engine bay compared to the scenario with HS. The HS approach was also thought as a more realistic representation, since air will, most likely, always leak through the grille cover construction over the entire grille area; and not, as for the CO case, be entirely sealed in the lower grille area.
Apart from the irregular temperature field in the under-hood area, the general flow field on the inside of the grille was also very different between the two grille-opening strategies. Figure 18 shows streamlines coloured by the velocity magnitude for the CO and HS configurations with a 17.5% opening of the grille area.
From Figure 18 the difference in flow behaviour inside the grille area is apparent. For the CO configuration, there is a very strong downward motion of the flow on the inside of the grille, which was not seen at all for the HS case. It was concluded that this downward flow was one of the main reasons for the irregular flow- and temperature fields in the under-hood area for the CO cases; and hence the resulting higher levels of the TTT values. A large part of the flow field entering the grille area was, instead of flowing through the cooling module, exiting under the vehicle.
As mentioned above, the HS method was considered superior to the CO. With these results as a reference, the SN model was also simulated with horizontal slots at an opening of 17.5%. The numerical results for both the COE and SN model are presented in Table 3.
As can be seen in Table 3 the TTT values for the SN case was unrealistic and indicates that this opening percentage of the grille would not be sufficient to ensure the required cooling airflow. The results were however expected since, according to the analyses earlier in the paper, the flow field in the under hood area of the SN model was not satisfactory and needed to be developed in order for the concept to be attractive.
The results from the simulations performed in this study showed that there is clearly a potential with having a SN model with a smoother shape than the COE model. However, it would not be beneficial to only change the exterior of the cab without also considering the layout of the components in the engine bay. Closing off the entire cooling airflow for the SN model showed that the aerodynamic properties were better for this model compared to the COE model. The CD of the SN model was reduced by 3.6% with a closed grille compared to the COE model at 0[degrees] yaw; and the corresponding value at 5[degrees] yaw was almost 6.6%, which really emphasizes the potential for this kind of cab design. At 0[degrees] yaw for the open-grille configurations, the absolute values of CD was more or less identical for the COE and SN models. This indicated that the flow field in the under-hood area was not ideal for the SN. Visualizing the flow field for the two models showed that a large portion of air was leaking around the cooling module for the SN, yielding significant flow losses as a consequence. It would hence be of interest to develop the under-hood area, to maximise the potential that this extra space presents with the addition of the nose. There is a great opportunity to re-think the concept of the layout in the under-hood area, even by a relatively modest elongation of the cab by 200 mm. Innovative solutions of the engine-bay layout, present possibilities for both improved aerodynamics and thermal management. Examples of interesting areas of development are not only ducted inlets, but also outlets, of the cooling airflow to and from the engine bay. Moving the cooling module forwards in the SN model, closer to the grille area, would be another possibility for an improved cooling airflow. Thinking even further, possibilities for splitting the cooling module into separate systems with separate channels for a more customized cooling airflow system are seen.
Even though highway driving, on level roads, at a constant speed is not considered as a critical cooling condition, it has been seen that entirely closing off the cooling airflow is not a scenario that would work in real life, at least not for any longer periods of time. Hence it was of great interest to investigate at what opening percentage of the grille resulted in reasonable values of the TTT values for this particular driving scenario. These results can be used in a further design process of grille-shutter systems.
It was seen from the simulation results with partially-opened grille that rather small differences in grille opening resulted in large changes in TTT values. The [C.sub.D] changes for the same grille percentage changes did not however change much; a grille opening of 10% with HS resulted in a 7% decreased [C.sub.D] value compared to the conventional grille configuration, while a 25% opening percentage still had a [C.sub.D] decrease of over 5% compared to the same case. Hence the pure [C.sub.D] values were not very sensitive to these opening percentages. It was also discovered from the partially-opened grille simulations that the type of opening method of the grille was important for the cooling capacity. The CO cases, where a single rectangular- shaped area was opened from the upper part of the grille, showed considerably higher TTT values for the same opening area as the HS cases. The reason for this was due to the irregular flow field inside the grille that occurred and hence the HS method was considered better, both for uniformity reasons, and due to the fact that the HS design was considered as a more realistic representation of a grille shutter system.
In this paper, two different cab designs have been evaluated in terms of aerodynamic properties and cooling capacity. It has been shown that there is potential for an SN concept, both in terms of reduced CD and improved cooling capacity, to develop or re-design the under-hood area to make use of the space for drag-reducing strategies instead of leaving it unchanged.
The layout of the particular SN model and under-hood area used in this study was not optimised to make use of the extra space obtained by the elongation of the cab. A large portion of the air entering the engine bay leaked around the sides of the cooling module, resulting in considerably higher TTT values for this configuration.
It has been shown that aerodynamics and cooling capacity can be evaluated in the same simulation, with active heat exchanger models, without affecting the CD results in a significant way. This was shown for both the COE and SN models, for both yaw angles tested.
A continuation of this analysis would be to add a duct behind the grille of the SN to better guide the flow between the grille and cooling module and hence decrease or even eliminate the leakage of air around the cooling module.
Another interesting continuation of the analysis of the SN model would be to move the entire cooling module forwards, towards the grille, to study the effects of having a similar layout in the under-hood area as the COE model. This approach also opens up possibilities for exiting the cooling airflow via ducts in a more controlled way than is the case today: where the air is simply exiting the fan and flowing freely around the engine.
As a final remark, the authors believe that a better-designed under-hood area of the SN model would result in both improved drag values and better cooling capacity.
Chalmers University of Technology Applied Mechanics SE-412 96 Gothenburg
The authors would like to acknowledge Vinnova/FFI for the funding of this study. Great thanks go to Volvo Group Trucks Technology (GTT), Gothenburg, for valuable support during the work. Also, the Road Vehicle Aerodynamics group at Chalmers University of Technology is acknowledged for their support.
AC - Air Conditioning
CAC - Charge Air Cooler
CAD - Computer Aided Design
CD - Drag Coefficient [-]
[C.sub.cooling] - Cooling Drag Coefficient [%]
CFD - Computational Fluid Dynamics
CO - Continuously Open
COE - Cab Over Engine
[f.sub.p] - Porous resistance source term [kg/[m.sup.2][s.sup.2]]
GTT - Group Trucks Technology
HS - Horizontal Slots
MRF - Multiple Reference Frame
m - Mass Flow [kg/s]
[P.sub.i] - Inertial resistance [kg/[m.sup.4]]
[P.sub.v] - Viscous resistance [kg/[m.sup.3]s]
RAD - Radiator
SN - Soft Nose
TTT - Top Tank Temperature [[degrees]C] V - Local velocity [m/s]
1. Logdberg, O., "Turbulent Boundary Layer Separation and Control," Ph.D. thesis, KTH, Stockholm, 2008.
2. Hallqvist, T., "The Cooling Airflow of Heavy Trucks-A Parametric Study," SAEInt.J. Commer. Veh. 1(1):119-133, 2009, doi:10.4271/2008-01-1171.
3. Pfeifer, C., "Evolution of Active Grille Shutters," SAE Technical Paper 2014-01-0633, 2014, doi:10.4271/2014-01-0633.
4. European Union, "Directive (EU) 2015/719 of the European Parliament and of the Council," http://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32015L0719&from=F R, Nov. 2015.
5. BETA CAE Systems S.A., http://www.beta-cae.gr, Nov. 2015.
6. CD-adapco, "CFD and CAE Engineering Simulation Software," http://www.cd-adapco.com, Nov. 2015.
7. Martini, H., Gullberg, P., and Lofdahl, L., "Comparative Studies between CFD and Wind Tunnel Measurements of Cooling Performance and External Aerodynamics for a Heavy Truck," SAE Int. J. Commer. Veh. 7(2):640-652, 2014, doi:10.4271/2014-01-2443.
8. CD-adapco, "STAR-CCM+ Manual," Version 9.06.009 Edition.
9. SAE International Surface Vehicle Recommended Practice, "Guidelines for Aerodynamic Assessment of Medium and Heavy Commercial Ground Vehicles Using Computational Fluid Dynamics," SAE Standard J2966, Rev. Sept. 2013.
10. Martini, H., "Perspectives of Aerodynamic Drag and Cooling Airflow for Heavy-Duty Trucks-Reconsidering European Total-Length Legislation," Ph.D. thesis, Chalmers University of Technology, Gothenburg, 2016.
11. Gullberg, P. and Sengupta, R., "Axial Fan Performance Predictions in CFD, Comparison of MRF and Sliding Mesh with Experiments," SAE Technical Paper 2011-01-0652, 2011, doi:10.4271/2011-01-0652.
12. Gullberg, P., Lofdahl, L., Adelman, S., and Nilsson, P., "An Investigation and Correction Method of Stationary Fan CFD MRF Simulations," SAE Technical Paper 2009-01-3067, 2009, doi:10.4271/2009-01-3067.
13. EURO4M, "European Temperature," http://www.euro4m.eu/downloads/euro4m CIB temperatu re.pdf, Nov. 2015.
Helena Martini. Chalmers University of Technology
Peter Gullberg. Volvo Group Trucks Technology
Lennart Lofdahl. Chalmers University of Technology
Received: 09 Mar 2018
e-Available: 07 Apr 2018
TABLE 1 Analysis matrix of the configurations evaluated in CFD. Front-End Heat Grille Yaw angle Design Exchanger Configuration Model COE Inactive Open 0[degrees] COE Inactive Open +5[degrees] COE Inactive Closed - 100% 0[degrees] COE Inactive Closed - 100[degrees]% +5[degrees] COE Active Open 0[degrees] COE Active Open +5[degrees] COE Active Closed - 17.5[degrees]% open (HS) 0[degrees] SN Inactive Open 0[degrees] SN Inactive Open +5[degrees] SN Inactive Closed - 100[degrees]% 0[degrees] SN Inactive Closed - 100[degrees]% +5[degrees] SN Active Open 0[degrees] SN Active Open +5[degrees] SN Active Closed - 17.5% open (HS) 0[degrees] TABLE 2 Numerical results of the chosen cooling performance parameters for COE and SN, open grille. Parameter COE 0[degrees] yaw 5[degrees] yaw Radiator TTT 59 [degrees]C 59 [degrees]C Parameter SN 0[degrees] yaw 5[degrees] yaw Radiator TTT 68 [degrees]C 68 [degrees]C TABLE 3 Numerical results of the chosen cooling performance parameters for COE and SN, 17.5% opened grille area (HS). Parameter COE SN 0[degrees] yaw 0[degrees] yaw Radiator TTT 99 [degrees]C 195 [degrees]C
|Printer friendly Cite/link Email Feedback|
|Author:||Martini, Helena; Gullberg, Peter; Lofdahl, Lennart|
|Publication:||SAE International Journal of Commercial Vehicles|
|Date:||Mar 1, 2018|
|Previous Article:||CFD Windshield Deicing Simulations for Commercial Vehicle Applications.|
|Next Article:||Methodology for Developing a Diesel Exhaust After Treatment Simulation Tool.|