Using Thermal Simulation to Model the Effects of Wind on the Mars Curiosity Rover

Figure 1. The Curiosity Mars Rover

The Centro de Astrobiología (CAB), a joint center of Consejo Superior de Investigaciones Cientificas and the Instituto Nacional de Tecnica Aeroespacíal (CSIC-INTA), designed and built the weather monitoring station contributed by the Spanish government as part of the payload of the Mars Science Laboratory (MSL) rover, Curiosity, which landed on Mars in August 2012 (Figure 1).

Called the remote environmental monitoring station (REMS), the collection of weather-monitoring instruments was intended to operate on Mars for the equivalent of two Earth years, acquiring data on Mars’ wind speed and direction, air and ground temperature, atmospheric humidity, pressure, and UV irradiation, from the surface of the crater on Gale Mountain (Figure 2). These measurements enabled analysis of the day-to-night cycle and seasonal environmental variations and provided the first measurements of UV radiation incident on the Martian surface.

Figure 2. The Gale crater. (Photo courtesy of NASA-JPL and the University of Arizona)

The Curiosity rover is the size of a small car, and, along with the REMS, it is equipped with geological, geochemical, and atmospheric instruments. It includes a robot arm that can dig and recover soil samples to be analyzed later in situ. It has cameras for navigation and scientific purposes. The overall mission was to determine the past habitability of the red planet, by studying the water markers, minerals characteristics, and sediments of the Gale Mountain area [1] .

Venus, Earth, and Mars are the only inner solar planets with atmospheres relevant to life as we know it. Venus has a hotter and denser atmosphere than Earth, whereas Mars has a much colder and less dense atmosphere than our planet. The Mars atmospheric pressure averages 6 mbar (compared to Earth's average at sea level of 1,013.25 mbar). Temperatures at the Gale crater range from -80 to 5°C.

Despite these differences, the atmospheric phenomena on Mars are similar to what happens on Earth. The Martian day duration is 40 minutes longer than our day, and a complete turn around the Sun is equivalent to two Earth years. Mars also has four seasons, and the second half of the Martian year is the windy period, when dust storms usually occur.

The goal of REMS was to provide insights into habitability, atmospheric processes, and surface-atmosphere interactions. The Gale crater provides a far more complex meteorological environment than that sampled by earlier missions because of its large and varied topography. The crater is 154km wide and is located in the north-eastern portion of the Aeolis quadrangle on the boundary between the southern cratered highlands and the lowlands of Elysium Planitia. Both its location and morphology make the Gale crater an interesting place from a meteorological point of view.

The REMS’ suite of sensors includes two wind sensors. These sensors were installed in two small booms, located in the mast. Before the rover left Earth, we needed to make sure that accurate measurement of winds flowing around Curiosity could be attained. To do this, we needed to know what the effects of the rover itself, its shape and movement speeds, would be on the airflow. Before launch, we used 3D thermal simulation (FloEFD computational fluid dynamics software [2] ), to analyze the effects of the rover on the wind stream flowing around it.

Most of the electronics for the Curiosity rover and its instrument are mounted in its central body. The locomotion system has six motorized wheels and a suspension system built for the rough Martian mountainous terrain. The rover is powered by a radioisotope thermal generator and can communicate directly with one of the Deep Space Network of NASA antennas or using the Mars Reconnaissance Orbiter or Odyssey satellites for relaying data to Earth. The remote-sensing mast supports some cameras for navigation and science, including the ChemCam instrument and most of the REMS sensors.

Because of the general constraints of Curiosity's design, the two REMS booms [3] are not long enough to be out of the rover fluid volume. We needed to identify the wind directions that would have the maximum effect on the rover and would affect the REMS measurements. We also needed to understand how the wind stream would be perturbed by the rover. Wind-tunnel tests were not feasible because of time and facilities restrictions so we decided to use CFD simulation.

Thermal Simulations

Figure 3. CAD model (left) of the rover from the Jet Propulsion Laboratory used in the thermal simulation software. Close up of the two wind sensors on the mast (right).

We used 3D thermal simulation software [4] , with our own external boundary condition domain that recreated Mars’ atmospheric conditions, to generate and activate the laminar and turbulent flow features (Figure 3). We also included conduction in solids and free-forced convection in the model.

The fluid computational domain size for the simulations was 4 × 4 × 2 m. A faithful reproduction of Curiosity's geometry was achieved using a fine mesh. The basic mesh dimensions of Nx = 123, Ny = 123, and Nz = 123 created an initial mesh size of 3 cm. In addition, local meshes were used to refine the solid cells in the rover: “wheels,” “mast and camera,” “deck,” “WS1,” and “WS2.” Also, a solid ground was added to simulate the effect of the rover on the terrain.

Figure 4. Solid cells used for simulating the rover.

The resulting Cartesian mesh consisted of 3.2 million cells, which included 2.3 million fluid cells, 590,000 solid cells, and 300,000 partial cells. Figure 4 shows the solid cells in FloEFD.

The general settings for the simulations were: external analysis, 713 Pa atmospheric pressure, 243 K temperature, and air as the fluid. Wind speed and direction were set in the velocity parameters box using three wind speeds: Vx, Vy, Vz.

Heat conduction in solids was enabled. With this function, we were able to simulate more realistic conditions of how the rover is heated and interference with the atmosphere and wind stream.

Figure 5. Detail of the refined mesh for one of the wind sensors.

Computation time varied from 10 to 15 hours, depending on the wind velocity or direction chosen for the simulation using a 24-CPU server machine.

The wind direction and speeds simulated were: 5, 10, and 20 m/s of speed, yaw angles over range 0, 30…330°, and pitch = 0°, 30.

The 24 wind sensors are based on hot film anemometry [4] . They are 1.5 × 1.5 × 1.5 mm in size, located in the front end; one on the left, the other on the right. Local fine meshes were used for each sensor to mesh the solid area properly (Figure 5).

Wind Speed Results

Figure 6. Simulation of the wind velocity around the sensors.

We used the simulation results to determine the perturbation of the rover on the wind sensor for different directions and speeds. We needed an extended number of simulations but, once the model was prepared, we could send batch simulation runs by modifying the initial parameters. We extracted detail xyz velocity information from the set-point parameters (Figure 6). These parameters were created near each of the wind sensors and around both of the booms.

Figure 7. Front winds for both booms at 10 m/s.
Figure 8. Rear wind (top) and side wind (bottom) for both booms at 10 m/s.

Figures 7 and 8 are detailed cut plots of the perturbation of the free-stream, wind-speed profile at 10 m/s for the wind directions 0°, 90°, 180°, and 270°. These illustrations show the perturbation that the rover creates around the wind-sensor area.

Thermofluid Simulations

Figure 9. The effect of the heat plume with no wind (left) and with 5 m/s rear wind (right).

The simulations were also modeled with conduction in solids and free-forced convection. This enabled us to heat different parts of the rover, providing a more realistic perturbation of the free-stream velocity (Figure 9).


Before sending the Curiosity out on its mission to Mars, we were able to extract from the simulations how the incident wind on the rover from different directions perturbs the reading on the wind sensors. Using thermal simulation and analysis, we were able to better understand which wind sensor is closest to the free-stream velocity and understand the different wind directions and speeds that could perturb both sensors because of the geometry of the rover.

The success of the design has been proven as Curiosity's mission time was doubled, sending important data that has helped our understanding of the near-Mars atmosphere. A good understanding of Martian lower atmospheric processes and the ability to simulate them with climate models is necessary for the potential planning of future safe operations of humans at the surface, as well as robotic mission operations [5] .

This article was written by Dr. Josefina Torres, Dr. J. Gomez Elvira, S. Navarro, M. Marin, and S. Carretero, Center for Astrobiology, National Institute of Aerospace Technology (Madrid, Spain). For more information, Click Here .