One of the first questions people will ask when you tell you’re planning to launch a rocket is: “But where does it land?”. If you are flying to a few hundreds of meters altitude this question may be easy to answer using some simple calculus. When you however start to mix in supersonic speeds, weather effects, possible malfunctions and manufacturing tolerances on the flight hardware and need to answer how certain you are of your answers, you’ll quickly find that you need to develop the most envied skill in human history: predicting the future.

Nicolai , Aaron, and Remon have taken up this daunting task and in doing so are the only group of people who have actually already launched the Stratos II+ a few times. Over half a million times to be exact. In addition they became native Python speakers in the process of writing their custom simulation software.

Simulations team members Aaron and Remon fearlessly handling tens of thousands of simulations.

Monte Carlo

The reason for running so many simulations is that they need to predict where the Stratos II+ will land and with what probability. Think of this problem like predicting which side of a coin you’ll see when you toss is. For a coin this is a simple; you can tell that there should be a 50/50 chance for seeing each side of the coin. But how do you calculate, or predict the probability of where a rocket will land, also taking into account possible failures and possible manufacturing errors? Using statistics to try to directly calculate the answer would be an absolute nightmare. Luckily there is the Monte Carlo method to help out.

Monto Carlo Casino after which the Monte Carlo method was named

When the first computers became available at the end of the ’40 performing calculations became very fast and cheap. Some smart people realized that instead of directly calculating the probability of something happening, one could also simply perform an experiment a large number of times and determine the probability of certain outcomes based on the outcomes seen in the experiment. If one for example would physically toss a coin a thousand times, one would end up with approximately 500 times head and 500 times tails. The code name for this method was Monte Carlo after the Monte Carlo Casino in Monaco. The same can be done for the launch of Stratos II+.

Portable super computer

By simulating the launch of the Stratos II+ a large number of times and taking into account all sorts of possible failures and manufacturing defects the simulations team came up with an answer on the question in which area the rocket will land with very high certainty. It took 500.000 simulations to provide the launch site provider INTA with a definite answer on how large a part of the air and sea space around the launch side needs to be closed. The safety zone is approximately cone shaped with a diameter of around 50 km.

Visualization of the simulated Stratos II+ trajectory in Google Earth.

This week a final simulation run of 25000 cases were performed to select a safety region where it would be required to activate the FTS (Flight Termination System) in order to reduce the required safety region. The team however was in Spain and thus did not have access to their usual computing facilities. By building a super computer on site using the laptops of a large number of team members the required computing time was reduced to only a few hours.

Stratos II+ campaign

The simulation effort ahead of the Stratos II+ launch was by far the largest ever done within DARE and a lot of knowledge was gained by the the team members. As we launch larger rockets and reach further to space accurate predictions become more and more important to ensure safety. In addition the simulations team plays on important role in the Advanced Control Team (ACT). The knowledge and expertise on simulation will ensure safe launches now and in the future. Join us live on Wednesday the 14th of October when we launch the Stratos II+ rocket to 50 km altitude.