Truck side skirts simulation
The truck skirts simulation tutorial looks at the effect of adding side skirts to a typical semi-trailer truck. The simulation is carried using our Computational Fluid Dynamics software SimWorks.
Before continuing with this tutorial, we recommend to read and complete at least one base SimWorks advanced tutorial to familiarise with the software.
The goal of the simulation is to evaluate the effect of the truck side skirts starting with a semi-trailer truck without any skirts and adding them progressively.
What are truck side skirts?
Side skirts are rigid panels added to the truck lower portion between the wheels. Their main purpose is to reduce the local flow aerodynamic recirculations ultimately reducing the truck drag. The aerodynamic resistance due to the introduction of aerodynamic losses is one of the major contributions to the overall truck fuel consumption and side skirts and other fairings are an effective solution in reducing the truck coefficient of drag.
Before starting the tutorial download SimWorks, validate your 14-day free trial contacting us and download the tutorial geometry:
Load the geometries
As already seen in the previous tutorials, we need to define a new project and define the new simulations:
- create new project and call it Truck aerodynamic study, define a new geometry calling it Truck skirts investigation and finally define 3 simulations calling them Truck without skirts, Truck with front skirts and Truck with front and rear skirts.
- Finally select each simulation and select the geometries: truck_no_skirts.igs, truck_front_skirts_only.igs and truck_front_and_rear_skirts.igs in the respective simulation.
- Finally load each one in the respective Geom viewer.
Define the outer domain
Since we are going to run the simulations in straight ahead condition, we can exploit the fact that the flow would be symmetric with respect to the longitudinal axis of the truck and therefore we are going to only model half of the truck. This also reduces the overall size of the computational grid, thus reducing the cost of the CFD simulation. Therefore, we can define an outer domain (OD) that splits the truck in half and set a slip wall boundary condition on the middle plane, as a symmetry condition.
- Define the OD with dimensions 140 6 15 and position 10 3 7.7
Base mesh parameters
From the Mesh tab in the simulation editor window define the following mesh parameters:
- Set the base size to 2
- Define the point in mesh location as -15 1 5 and its size as 0.5, the point will be shown in the geometry viewer, verify that is inside the OD but outside of the geometry
- Create two mesh groups for both the front and the rear wheels
- Create mesh groups for all the side skirts (in the simulations where those are present) and assign a mesh refinement of 6 6 and 3 prism layers with a starting size of 0.005m and a growth ratio of 1.2 for the side skirts group and the wheel group
- Leave the remaining mesh group containing the Cabin and the Trailer at a level of 5 5 but still add the prism layers with the same parameters as the previous point
- Finally add a refinement box which guarantees a smooth transition from the coarse OD base mesh and improves the mesh refinement around the truck. The refinement level will be 3, size 40 4 8 and origin 0 2 4.2 as this one will also be cut in half.
Boundary conditions definition
The inlet face of the OD will be defined as a Velocity Normal boundary condition with speed 25 m/s (corresponding to 90 km/h) the exit face will be a Pressure Outlet with a constant pressure of 0, since we are simulating just half of the truck the symmetry face which cuts the truck in half will be a Slip wall boundary condition (this is the default one so no need to do anything but makes sure that in similar exercises in the future you will use this boundary condition). The boundary conditions can be defined using the pre-processor in SimWorks by showing the Boundary conditions types layering in the layering menu, as seen in previous tutorials.
In the Output tab define 50 planes in X between -15 and 25, 30 planes in Y between 0 and 3 and finally 30 planes in Z between 0 and 5.
Finally, in the Setup tab, set the Reference area to 5.25 (this one is half of the truck frontal area used in the previous tutorials) and Reference length of 3 to make sure that all the calculated aerodynamics coefficients are correct.
Complete the simulation
Run the simulation on 4 processors for 1000 steps and execute all phases: Setup, Mesh and Run sequentially.
Post-processing of the results
Right click on the completed simulations and load the results of each simulation with Fields → Load. Under Selection, tick the sections normal to X and leave the truck in show.
- Select the Cp0 variable to assess the flow energy and define a range from 0 to 1
- Click on Show/Hide delta to create the delta images
- Define the delta Cp0 range from -0.1 to 0.1 to capture the flow field differences
For the subsequent analysis, the background has been changed to white in the Settings menu.
When looking at the effect of adding the front skirts, the difference in terms of overall aerodynamic wake is very small (1) and can only be appreciated looking at the delta, which shows a clear increase in energy in the front tyre wake region (2). Looking at the delta in 3D, the front skirts are missing (3) as this component is not present in both simulations, so the delta cannot be evaluated on it.
Finally, looking at the simulation with both the front and the rear skirts the effect on the front position of the aerodynamic wake is similar to simulation 2, as expected (4).
The delta plot is a very important tool to assess the effect of a geometrical change. We are now going to analyse the effect of adding the front skirts loading only the first two simulations in the post-processor. We are also going to use the square probing to calculate the increase in energy in the front skirts region:
The average Cp0 value in the window of interest goes from 0.428 to 0.465 with an increase in local flow energy of roughly the 10%.
Now it is possible to check the small difference on the side of the truck once again using the square probing tool:
In this case the effect on the average value is very small, of the order of 0.4%, showing that the main effect is a small repositioning of the losses building up on the truck side surface rather than a significant improvement or worsening.
Finally, looking at the top of the cabin (figure below), in both delta plots there is small difference in the top truck wake despite the geometry in the area being the same, and this is again confirmed by a square probing analysis
The reason for the latter two effects is that the wake behind a bluff body is inherently unsteady. The steady state simulation does not converge to a stable solution but still has some small oscillations, as those oscillations are not perfectly timed among the different simulations such a variation can become apparent. This is not a problem but rather an expected result the user should be aware of.
Introducing iteration averaging of the solution would help reduce the differences between two simulations with “unsteady” results.
Truck side skirts simulation results
To get reliable results it is important to carry out a mesh sensitivity study first and therefore have access to a suitable machine. Also natural solution oscillations have not to be taken into account in the final aerodynamic analysis as explained in the previous point. Future versions of SimWorks Manager will include an averaged solution of the latest n time steps naturally smoothing the final results:
Nevertheless adding the front side skirts and subsequently the rear one has shown a progressive increase in flow energy, by means of a general increase in local Cp0, carrying out simulations with a finer mesh can show in detail the effect on drag of such devices.
Download SimWorks and activate a free 14-day trial or contact us to find out more