Managing your Risk through Constants and Variables
Icaraus flew too close to the sun. While a myth, there’s a reason new pilots are told the ancient Greek myth of Daedalus and Icarus. To recap - the two men were stranded on an island. Daedalus realized he could fashion wings out of bird feathers and wax, so he made two sets. One was for him, and the other for Icarus. The two put on their wings and managed to escape. However, Icarus, overcome with the joy of flight, kept flying higher and higher until he got too close to the sun and his wings melted. Icarus plunged to his death.
Even today, though we don’t make airplanes out of wax and feathers, aviation is a dangerous activity. This is why the FAA has instituted mandatory risk management training as part of its pilot training syllabus. Those of us fortunate enough to take airplanes into the air are given multiple tools to manage the risk associated with flight. These tools range from the mandatory use of checklists, to currency requirements and required maintenance cycles. All of these are designed to buy down risk for routine flight operations.
Research aviation, or flight test, adds additional layers of risk to flying. The FAA has a published handbook for managing risk. Aerospace giants such as NASA, the Air Force, and Boeing have their own handbooks for risk management in flying and flight research programs. There are multiple tools available, but perhaps the most popular is the five-by-five risk management chart. This chart breaks down risk by likelihood and impact. The higher these two numbers are, the higher the risk. From there, it converts into a stoplight chart to identify whether the risk is high, medium, or low. When used properly, this tool is a great way to categorize identified risks and begin to decide how to mitigate them.
At its heart though, risk management is no different than the tools used to fly an airplane. When I was flight instructing, I would teach my students that flying is nothing more than managing your constants and your variables. Want to level-off from a climb? First set your pitch with the yolk, then set the power with your throttle. Make them constants; and then set your trim. Your trim then becomes your only variable as you continue to fine-tune it as you fly through different air. Moving into cross country flying and instrument flying we use the same methods - plan your alternate airports and have contingency plans so when bad things happen you already have a plan. You can even turn the greatest variable of all, the weather, into a constant through the institution of weather minimums. As you slowly turn your variables into constants, you start to decrease either the likelihood or the impact of the risk; and they slowly start to move from red to yellow, or even to green.
When we add test operations, or flight research, we find that we can apply similar techniques to manage the risks of the new technology or technique. My favorite example of this goes back to my time at NASA and the Low Density Supersonic Decelerator project. In a nutshell, this was a new method of atmospheric re-entry using inflatable re-entry vehicles. In order to properly test this, we needed to get our test vehicle up to 120,000 or so feet before we could launch it. The test vehicle itself
was a flying saucer, consisting of an Orion heat shield, the inflatable re-entry system (called the SIAD), and a small rocket motor normally used to put satellites into their final orbit. Here it would be used to accelerate the flight vehicle to Mach 4 and simulate our Martian re-entry. The vehicle would trigger the SIAD, a ballute, and a recovery parachute in succession before splashing down in the ocean and giving us our valuable flight test data.
As you can see, none of this is standard. Our normal tools to manage our variables and constants are essentially moot. There is no way to have pilot currency, no maintenance cycles, and no real training that can be given for a one-off mission. Additionally, we are getting way outside of normal operating procedures because we are essentially carrying a bomb on a balloon. This alone required a brand new method of launching the flight vehicle rather than the well-practiced, well-rehearsed, and well-known method.
If you’ll permit a small tangent, its important to understand how high altitude balloons are normally launched. Normally, there are two vehicles used. The first vehicle is the spool vehicle used to hold the balloon in place while they fill it with gas. The second is a vehicle that is attached to the payload, and maneuvers the payload while the balloon is being launched so that it won’t swing like a pendulum when the payload is released and hit the ground damaging it.. In this instance, because we were essentially carrying a bomb that could explode in the vehicle driver’s face, using that second vehicle was not an option. Therefore, the balloon team developed a launch system using a stationary tower that could trigger a payload release remotely.
This effectively eliminated our variable of ‘what if the rocket motor explodes on the ground?’. However, it created a new variable by introducing a new launch system. So how did we turn that mega variable into a constant? We started by chair-flying the procedures with the team as the tower was designed and built.
Once created, the launch team ran a series of practice launches first with tethered, and then with free-flying balloons in order to vet their procedures and build confidence in the system. Then, the system was certified for our flights. On the day-of, the launches were all performed flawlessly. In total, two LDSD flight tests were conducted each one year apart. Murphy’s Law, though, still reared his ugly head by destroying the parachutes of both systems when they were deployed at altitude. While the parachutes were tested in the same way they test parachutes for Mars missions, they were considered at the least risk of failing since it was essentially tried-and-true technology. This just goes to show that one cannot take anything for granted; we flew too close to the sun on that one.
So how do you keep yourself from suffering the same fate as Icarus? The next decade of flying will see the introduction of two revolutionary technologies. The first will be the deployment of Artificial Intelligence (which NX Aviation has already been involved with flight testing), and the second is urban air mobility vehicles sometimes colloquially called ‘manned drones’. Managing the ratio of constants and variables will be of paramount importance to those companies looking to field their technology. Air Mobility faces similar challenges to the ones we faced on LDSD; that is the creation of a new type of flying machine, and the laying of new infrastructure to support it, for which there is no training, no established protocols, and no institutional risk management. Similar methods of managing risk, such as baby steps testing systems not just for maiden flights, but over time and over routes, will be important to ensure that they are as safe as possible moving into this brave new world. The same can be said for the inevitable introduction of artificial intelligence.
One way to buy down your risk, and add more constants, is through the use of flying testbeds. What these allow is a way to fly systems utilizing existing and well-proven airframes. These airframes can be manned or unmanned, but should ideally be a tried-and-true system. For example, if you’re flying a new flight control system on your UAS, be able to flip back to a manual or stock system if something goes wrong. Or, for AI, place it on a manned aircraft and give a real, flesh-and-blood pilot control over when to engage and disengage it. Methodical testing of the systems and infrastructure using existing technologies is the clear gateway to enabling the biggest revolution in aerospace this century. NX Aviation offers many of these testbeds for customers, as well as the expertise to help manage risk in your projects moving forward. How can we help you?