Upon landing on the Red Planet in May 2008, and while in operation for the next five months, the Phoenix Mars lander fed back to Earth a steady stream of data, including confirmation of the presence of water-ice in the Martian subsurface and more than 25,000 pictures from the first atomic force microscope ever used outside Earth.
Contributing to that data was the lander’s meteorological station (MET), which provided atmospheric measurements that complemented surface and subsurface data obtained by other Phoenix instruments for a complete picture of the water cycle from the planet’s regolith to the upper atmosphere.
Prior to launch, a team of Canadian scientists at the University of Alberta aided in development of the MET. With design and calibration experiments difficult and costly to perform, the university turned to virtual testing with fluid dynamics software from ANSYS.
The CFD software was used to simulate the Martian environment, quantifying the effect of winds and of the lander itself on the measurements to be performed by the MET. During their research, scientists discovered that, under certain wind conditions, heat emitted from the lander could cause a temperature sensor to show higher-than-atmospheric values.
"With space missions, there is only one shot at getting it right. Any minor flaw could result in the instantaneous loss of years of preparation and hundreds of millions of dollars," said Carlos Lange, Associate Professor of Mechanical Engineering at the University of Alberta. "Using simulation software from ANSYS, we learned that the internal heat generation and emission of radiation from the lander's surface could increase temperature measurements. Similarly, obstacles upstream from velocity and pressure sensors could alter readings of wind magnitude and/or direction."
Therefore, Lange and his colleagues calibrated the instruments through a large parametric study. After the lander touched down on Mars, the university team used the results of such simulations to evaluate the raw mission data and find instances when these wind directions occurred. This process was key to preventing misinterpretation of the data by the Phoenix scientific team.
Simulation turnaround time was a concern to the team as well. During the mission, limited time and power resources were allocated daily to the operation of specific instruments.
"To make decisions about the prioritization of data collection, strategic planners sometimes required input based on the results of the simulations. So we needed to quickly simulate new cases," Lange said.
This short time-response requirement was met by employing the high-performance computing capabilities of ANSYS software, running the simulations in parallel on a cluster to achieve, at times, super-linear speedup (speedup greater than an amount proportional to the number of processors used). The efficiency of high-performance computing capabilities and multicore hardware together enabled new simulations to be completed within the time frame required for decision making.
The success of the University of Alberta's work has allowed for additional simulations to be performed to aid in the explanation of certain phenomena found in the raw data.