Lessons Learnt: Systems Engineering Failures | X-33 | Hubble | Delta 180 Interceptor
Lessons Learnt from three space missions; X-33, Hubble and Delta 180; to understand the different Systems Engineering failures in a complex aerospace system.
Systems Engineering is one of the five specialisations of Aerospace Engineering. The importance of Systems Engineering is sometimes overlooked. In today's Lessons Learnt article, we look back towards three space missions to understand the Systems Engineering failures in a complex aerospace system. A 2011 paper, "Learning from Failure in Systems Engineering: A Panel Discussion" examines the examples of X-33 Reusable Launch Vehicle, Hubble Space Telescope and Delta 180 Powered Space Interceptor.
X-33 Reusable Launch Vehicle
X-33 was a reusable launch vehicle program started in 1996. The program had many objectives. Because America had last invested in a new launch vehicle technology during the shuttle design, to catch-up, they set to reach for new technology in every subsystem. One major technical challenge was the linear aerospike rocket engine. The team had alternate designs in case it faced development or testing problems. Though it was good to have alternates in your pocket, it was not possible to have equivalent alternatives to each and every subsystem. It was assumed that ground tests would have failures. Ultimately, one of the other subsystems failed in a test. The reached far, but the program was cancelled in 2001.
Estimate all resources from the start.
While the initial design strategy was to set up new technology in every subsystem, an estimate of resources, money and facilities should be made.
A constant vision and persistence, even in times of failure, is needed.
The vision to achieve new technology in every subsystem was challenging to sustain with increasing costs and delays.
Hubble Space Telescope
A team was tasked to design and develop a backup system for the primary fine guidance sensor that would work on different principles. There were many "eye-opening" problems identified by this team.
People working on solar arrays were not coordinating with people working on the control system.
"Therefore, as the solar would swing and out of sunlight, they would irrevocably excite the satellite in return, and there was no image motion compensation or effective correction in the control loop."
The electronics, which were planned to be used, had launch-vehicle heritage rather than spacecraft which generally operated for much longer mission life.
Finally, there was no end-to-end test of the telescope.
Despite all this, every contractor could prove their design met the requirement given to them.
Communication is the key to a sophisticated design.
All the different teams should be able to communicate their ever-changing requirements with each other. The designs are often evolving, and therefore a constant communication is needed.
Despite every individual subsystem proving the requirements set onto it, an end-to-end test is of utmost importance. There are hundreds of interfaces between different subsystems which are tested only in integrated trials.
Delta 180 Powered Space Interceptor
The Delta 190 was a success. It was late after the launch when someone discovered a problem with the control system. The simulations had not modelled stiction, a friction force in the radar homing device. This oversight resulted in a slight lag of control which produced an error of a few feet in the trajectory. Fortunately, the interceptor technology was not good enough, resulting in a very big design. The Delta 180 still ended in a hit-to-kill blow. Three separate teams worked independently on the simulations, followed by a meeting to discuss the simulations. The launch was approved only when consolidated results were within 0.5 m error probables. Everyone missed the importance of stiction in the vacuum.
Despite an overall success, all the intricate subsystems parameters should be reviewed.
While the learnings presented above are based on my own understanding of systems engineering, the paper also offers a few essential learnings.
Failure is an option
Going against the famous words of Apollo 13, the panellists mentioned that failure should always be an option. In examples, only X-33 failed at the final level. Hubble had failures right until the 1993 servicing mission, but it still became the icon of space exploration.
Process is not the solution
In the case of Delta 180, every process was done right. Multiple independent teams worked and checked the simulations, and it still resulted in an oversight. The solution to failures in systems engineering is not increasing process. While the process is essential, teams which don't understand the system or the process fail. A simple example stated was that giving a checklist to fly an aeroplane cannot make everyone a pilot. Process, in the end, is just a tool. Humans and their creativity and intuition, along with the leadership, are what sets apart a great systems engineering team.
As we saw in the last lessons learnt article, it was humans who found creative solutions to the problems. Leadership and human resources management are, therefore, essential aspects of any program.
If you know about any other systems engineering failures, do let me know and I'll cover them in my subsequent posts.