Test 17

Aircraft ditching

By Matthieu de Leffe (Nextflow Software, France)

(Download full test case data files here: SPHERIC_TestCase17.zip)


The ditching is the emergency procedure that leads the aircraft to land on water. Although being a rare event, aircrafts need to be designed and certified at ditching in order to minimize the risks for the occupants and to allow a safe evacuation.

The objective of the benchmark presented is to verify the ability of the SPH method to correctly simulate the flow taking place during an emergency ditching of an aircraft. The validation of the method requires to capture the global loads which will then allow to correctly predict the evolution of the plane: will it bounce on the surface, or will it sink? The local pressure values are also a determining factor in predicting the behavior of the aircraft's structure.

In order to validate this, we have at our disposal the experiments realized within the H2020 European project SARAH by Iafrati et al. (2019) at the high-speed ditching facility of the CNR-INM.

Fig. 1. High-speed ditching facility at CNR-INM.

Fig. 2. Simulation by SPH-flow code.

Flow phenomena

During the early stage of the impact phase, the hydrodynamics of the free surface flow is strongly nonlinear, and the highest loads are generated. A preponderant effect is the suction present at the back of the surface. With the SPH method, this is an important point because negative pressures can lead to the well-known tensile instability. Depending on the speed and the shape of the fuselage, cavitation and ventilation can also occur.

Fig. 3. Low speed on the left, high speed on the right with cavitation and ventilation.


In this test, we are only interested in the small part of the fuselage that will firstly impact the water (figure 4). The impact occurs in a tank which is large enough to allow neglecting the presence of the boundaries during the impact stage.The body geometry and the position of pressure probes are given in the attached igs file.

Fig. 4. Representation of the double curvature specimen as a portion of the whole fuselage.

Fig. 5. Bottom view of the pressure probes positioning and numbering.

Boundary conditions

The specimen surface is smooth aluminium and can be modelled through free-slip boundary conditions since the flow Reynolds number is very high.

Initial conditions

The specimen is attached to a carriage whose speed at the moment of impact can be considered constant. The impact occurs with an oblique velocity and a certain pitch angle (the angle between the longitudinal symmetry axis of the fuselage and the water surface). For the case here considered the pitch angle is 6° and the vertical to horizontal velocity ratio V/U =0.0375. In the present benchmark two velocities are proposed, but you can find other configurations in the reference paper by Iafrati et al. (2019).

For the low speed test case: U=26.83 m/s, V=-1.01m/s For the high speed test case: U=40.06 m/s, V=-1.5m/s
The latter case is particularly challenging as cavitation and ventilation phenomena occur in the rear part of the specimen.
To compare and synchronize in time the simulation and the experiments, a good reference is the instant of impact on the P17 sensor.
Water has a density of 1000kg/s and is initially at rest. The gravity acceleration is 9.81m/s².


Experimental data are provided only.

Results specification

The result files (High_speed_expe.txt and Low_speed_expe.txt) present experimental results by CNR-INM, in particular the pressure values on the surface axis (probes 4, 9, 13, 17, 21, 24, 28, 30) and the vertical and horizontal force components.

Results format

The ASCII file containing the result has the following format:

column 1: time (s)

column 2-31: pressure (Kpa) measured at probes 1 to 30

column 32: total vertical force Fz (N)

column 33-34: vertical force measured, respectively, on the rear and forward sides

column 35: total horizontal force Fx (N)

Benchmark results

In the figure below an example of the horizontal and vertical force recording and the pressure measurements at probes  4, 9, 13, 17, 21, 24, 28, 30 for the low speed test.

Fig. 6. Experimental results from Iafrati et al at 26.83m/s.

SPH Publications using this Case

    If you have published results for this case, please email the webmaster to have your papers added.