JLR vehicle wade testing

LinkedIn +

Jaguar Land Rover has developed a novel approach to vehicle wade testing through simulation

The ability of a car to maintain its stability and functionality in a vehicle wading scenario – which is detrimental to underbody components, bumper cover, electronic circuits, air intake (causing hydro-lock) and engine – is crucial. Jaguar Land Rover has developed a new approach to vehicle wade testing through simulation in order to improve the wading performance of its vehicles.

Physical wading tests typically involve driving the car through different depths of water at different speeds. Often, the underbody design and placement of components, as well as the structural design of the chassis, have already been decided with simulation not being used. This leads to the late detection of failure modes, expensive design changes, and increased test costs and times.

An established CAE process for vehicle wade testing can identify failure modes at an earlier stage, provide insight into the structural integrity of underbody components, and analyze multiple designs with confidence, leading to an optimum design. The associated cost and time savings are also enormous.

Choosing the right simulation tool

Literature on best practices and the use of CAE in vehicle wading is limited. The work done by Xin Zheng, Xin Qiao and Fanhua Kong in their Vehicle Wading Simulation with STAR-CCM+ (which was presented at FISITA World Automotive Congress, SAE China, Beijing, in 2012) is the major reference for JLR’s development of its CAE process. Aside from this, JLR is the first OEM to publish literature on this topic. The need for this process was to understand the failure modes of underbody components early in the design stage and their effect on the vehicle performance and integrity.

The current testing procedure at JLR involves driving the vehicle over a ramp into a wading trough and using another ramp to exit the trough, with testing done at different speeds and water depths. Various combinations of speed and depth produce differing behaviors in stability, splash pattern and bow wave formation in front of the vehicle, and using simulation JLR aims to understand these different behaviors and optimize the underbody design.

With no historic literature or procedure available, JLR’s first challenge was to identify a computational tool capable of accurately modeling the motion of a vehicle through water. STAR-CCM+ was one of the contenders, in addition to a Smoothed Particle Hydrodynamics (SPH) code and LS-DYNA, another popular Navier-Stokes-based commercial code.

The CAE process needed to accurately simulate the transient pressure forces on the underbody components due to the motion of the vehicle and water relative to each other. To accurately identify failure modes, the tool needed to handle modeling the motion of the vehicle in a fully-transient analysis. After careful consideration, STAR-CCM+ was the clear winner due to its proven use in the automotive industry, overset mesh capability to model motion and a well-validated volume of fluid (VOF) model to capture the air-water interface during wading.

The motion modeling needed to be robust and as close to the test scenario as possible. The overset mesh capability technique involves two different mesh domains, one for the vehicle (overset region) and one for the background domain. This Chimera meshing technique will cut out the region of the background grid overlapping with the overset region, leaving the bordering cells (acceptor cells) between the two regions which can communicate with each other through interpolation. This enables handling of large motions in a robust, accurate manner.

Validating the overset mesh

Before applying the overset mesh approach to the vehicle wading simulation, it was imperative to validate this methodology for modeling an object motion into water. For this purpose, JLR scaled down one of their vehicles into a rectangular block to be tested in a towing tank. Six pressure sensors were placed on the block in testing to gather transient pressure data, which could be compared with the CFD results to validate the numerical approach. The box was 1,000mm x 400mm x 500mm and tests were conducted at water depths of 50mm, 180mm and 1,000mm and at speeds of 0.87m/s and 1.86m/s.

Figure 1 shows the overset mesh with hexahedral cells around the block in STAR-CCM+. The SST k-omega turbulence model in STAR-CCM+, well-validated in the marine industry, was used with the VOF model to capture the air-water interface. Pressure monitors were set up in the simulation at the exact locations of the six pressure sensors.

Figure 1: Mid-plane cross section of overset mesh and domain

Figure 2 shows the rectangular block at an immersion depth of 180mm and speed of 1.85m/s in both the towing tank and simulation. This shows good comparison of the water level around the block between test and CFD.

Figure 2: Simulation result at immersion depth of 180 mm and speed of 1.85 m/s

In Figure 3 the correlation of peak pressure data (in mm of water) between test and simulation at the six sensor locations is represented for 180mm and 1.85m/s. The difference between simulation and test results for all scenarios was within 10%, which was deemed acceptable. In addition, the water level height comparison between CFD (0.158m) and test (0.16m) was also satisfactory, establishing the validity of this simulation method.

Figure 3: Comparison of peak pressure data (in mm of water) at sensor locations for 180 mm, 1.85 m/s

Vehicle testing

With confidence in the simulation strategy established, JLR moved to the vehicle wading testing and modeling. A Jaguar XJ was used for the wading tests, conducted in the wading trough at Millbrook Proving Ground, UK, (Figure 4).

Figure 4: Wade testing in the wading trough

Sixteen waterproof pressure transducers were fitted on the underside panels (Figure 5) and bumpers, and protective stainless steel meshes covered the sensing diaphragm. The data acquisition and signal conditioning system were set up in the rear of the vehicle, with shielded electrical signal wires to minimize contamination of test data. Different speeds and wading depths were tested. The vehicle started from standstill and data acquisition began before the vehicle entered the water and stopped when it came to a halt.

Figure 5: Sensor locations (white marks) on the vehicle undertray.

CFD modeling of vehicle wading

For accurate modeling of the test environment, a CAD representation of the vehicle and the wading trough was built and cleaned in Hypermesh and ANSA, and brought into STAR-CCM+. The vehicle was aligned with the ramp entry with the wheels floating to enable rotation (Figure 6). A rectangular domain around the vehicle was created to be the overset region which moves and the rest of the domain was modeled as the static background region. The cool packs (intercooler, condenser and radiator) were modeled as separate domains to solve for porous physics along with normal physics, and were connected to other regions by internal interfaces through which data interpolation takes place. A hexahedral trimmed mesh was automatically generated with proper refinement around the cool packs, water region and the motion path of the vehicle. The final mesh count was around 40 million cells.

Figure 6: Motion definition of vehicle and wheels and initial water in STAR-CCM+

The Segregated Implicit Unsteady solver was used to resolve the flow field and the VOF model to solve the multiphase flow physics. Turbulence was modeled using the SST k-omega model and experimental data supplied the inertial and viscous resistance coefficients for the porous flow physics. A velocity boundary condition was chosen at the domain inlet and the side and upper faces were designated as pressure outlets. A rotating (while entering the trough) and translating motion were prescribed for the vehicle to model test conditions, and tangential velocity boundary conditions were given at the wheels using local rotation rate. Sixteen pressure monitors were set up in the simulation at the same locations as the test to compare the results.

Figure 7 shows the comparison of transient pressure data in sensor 2 (undertray) between CFD and testing at 450mm and 1.944m/s. The transient pressure data from CFD in all scenarios was within acceptable limits in comparison to test data, especially on stiff components like the undertray. In flexible components such as the aeroflips, the numerical results were significantly higher compared to test data. This is to be expected since these were modeled as rigid bodies in the simulation while in testing, the deflection from loading leads to reduced pressures. The front bow wave structure also corresponded well between CFD and experimental results.

Figure 7: Comparison of transient pressure data in Sensor 2 (undertray) at 450 mm, 1.944 m/s

One of the benefits of using STAR-CCM+ is the fully coupled, two-way, co-simulation capability with Abaqus, a leading finite element analysis (FEA) structural solver from Simulia. Pressure data from STAR-CCM+ was mapped at various time intervals to Abaqus and the loads at various fixtures and high stress areas were obtained. This information is crucial in assisting the underbody design at an early stage. JLR modeled one-way coupling between the fluid and structure but future work will model two-way coupling. The Von Mises stresses on the undertray at a time step of 0.675 seconds from Abaqus are seen in Figure 8.

Figure 8: Von Mises stresses on undertray at T=0.675 sec

A full multi-physics procedure is being validated currently on a simplified model using STAR-CCM+ and Simpack, a multi-body simulation (MBS) software using a coupling tool called Multiphysics Code Coupling Interface (MpCCI). This will enable the forces and torques from STAR-CCM+ to be transferred to Simpack to calculate jumping behavior when the vehicle enters water. Simpack then transfers the corresponding velocities back during jumping behavior to STAR-CCM+.

Visionary approach

JLR has developed a revolutionary process for vehicle wade testing using simulation, the first published work of its kind among OEMs. The overset mesh capability of STAR-CCM+ and advanced physics models have helped JLR successfully integrate virtual testing into its process, giving better insight into the underbody component loading and potential failure modes at an earlier design stage. Future work involving FSI and MBS in addition to CFD will result in an accurate virtual test bed for wade testing.

The benefits are many, and include the early detection of failure modes, the ability to investigate multiple designs, a reduced cost of testing, lesser delays in program timing and better wading capability.

Dr Prashant Khapane, manager durability and reliability CAE, and Uday Ganeshawade, senior CFD analyst and architect,

Jaguar Land Rover – vehicle engineering

Mobile: +44 7788 302231;

Email: pkhapane@jaguarlandrover.com

July 22, 2015

Share this story:

About Author


John joined UKi Media & Events in 2012 and has worked across a range of B2B titles within the company's automotive, marine and entertainment divisions. Currently editor of Automotive Testing Technology International, Crash Test Technology International and Electric & Hybrid Marine Technology International, John co-ordinates the day-the-day operations of each magazine, from commissioning and writing to editing and signing-off, as well managing web content. Aside from the magazines, John also serves as co-chairman of the annual Electric & Hybrid Marine Awards and can be found sniffing out stories throughout the halls of several of UKI's industry-leading expo events.

Comments are closed.