Mitigating Risk of UAV Collisions Using Explicit Dynamics

drone UAV explicit dynamicsUnmanned aerial vehicles, in short UAVs or drones, have become very popular both in the industrial and consumer space. With the number of units expected to reach 67 million by 2021 the potential for accidents and collisions with manned air vehicles is real. Understanding and mitigating the impact of UAV collisions using pervasive engineering simulation and explicit dynamics will be the key to helping accelerate the acceptance of drones into commercial airspace without sacrificing safety.

The working concept of drones was first developed in the early 19th century but they were not practical to fly, because it was very difficult to control them without an on-board control system. In the late 1990s and early 2000s, MEMS sensors and miniaturized electronic controllers became more affordable and proliferated, and were used for on-board attitude control. With an on-board attitude control, the pilots could easily fly them because they only had to maneuver the drones and not worry about stability issues.

There are innumerable applications for drones today, starting from parcel delivery, aerial photography, and surveillance etc. It is estimated that the consumer drones market alone will reach a market value of $5.0 billion by the end of 2021 from $1.9 billion in 2015. With such a huge increase in the number of drones flying in the sky, the threat to safety caused by them also increases at an equal pace, if not more. There have been multiple cases reported where drones were flown in restricted airspace and caused havoc.

In one such incident, two sightings of drones being flown recklessly close to Heathrow Airport in London have been reported. To evaluate the risks associated with drones in the vicinity of airports, the Department of Transport, in conjunction with the Civil Aviation Authority (CAA) have ordered tests, to be conducted by a private company, wherein, drones will be made to crash onto the fuselage and windows of aircraft in mid-air. They have committed more than £250,000 to pay for the tests.

With the recent advancements in engineering simulation, extending the use of the technology from digital prototyping to digital exploration and digital twin, ANSYS can offer seamless solutions for conducting virtual tests to accurately mimic the crash test sequence. Simulations can make the physical tests much more efficient by reducing the number of tests to be performed per drone, and therefore reducing the overall testing costs and the time needed to obtain the desired insight thus accelerating the regulatory process.

To demonstrate this a series of impact simulations were conducted using the explicit dynamics solve in ANSYS Mechanical Enterprise to study the effect of a high speed quadcopter propeller colliding with a static wall that could be used to represent the surface of another aircraft.

High-speed quadcopter propeller impacts have very small time-scales – of the order of micro-seconds. Given the complexity of the test rig (an aircraft mid-air and a quadcopter), it is extremely difficult to record the exact conditions of the DUT (Device Under Test) during the impact.

The simulations performed using ANSYS Mechanical Enterprise, however, offered great insights into the exact phenomenon occurring at the time of impact. To put this in perspective, we simulated the above case for 1 microsecond with a time-step size of 10 nanoseconds to accurately capture the physics during impact.

This can give immense information to the testing engineer. Further with ANSYS tools we could easily parametrize various design points like the propeller materials, wall materials, propeller rotation speed, propeller geometry etc.; this helped us evaluating the failure threshold for the combinations of design points. Going one step further, using the explicit dynamics solver we were also able to track the broken propeller fragments after the impact and analyze their trajectories.

This helped us in analyzing whether the high velocity propeller fragments can cause any damage (especially with the modern, high strength carbon fiber propellers) to the existing systems on-board the quadcopter, or by piercing into the wall and to assess the possible fragment ingestion.

To conclude, I would like to highlight that soon drones will deliver our goods, guard our gates and analyze our crops. It is apparent that we should get accustomed to them and start using engineering simulation to efficiently design safety measures to let the good drones do their work and keep the bad drones at bay.