Matthew Parsons, US Fulbright Advanced Student, ITER
I love making plans, and I do so almost obsessively. If I take a trip somewhere, you can bet I’ve planned enough things to do and see to occupy every hour of the day. With my research, I can tell you what I expect to accomplish this month, this year, over the next five or ten years, and how it ties into what everyone else in my field all around the world is working on. While I have always found it helpful to give myself a direction to go, I have also found that the most exciting and memorable things that happen are the things that aren’t planned.
At the end of my first week as a Fulbrighter in Aix-en-Provence, I was able to move into the apartment where I would be living for the rest of my stay in France. However, before I even got settled into my new room, I found myself helping out with various tasks around the apartment. The woman who was renting me her spare bedroom had apparently decided it was time to redecorate. After some painting tasks (and then getting settled in), I helped to reorganize the garage. The reorganization was partially motivated by her decision to donate her electric piano to the neighbors upstairs. In mentioning that I actually play the piano, I accidentally volunteered myself to give lessons to the neighbor’s children. I was hesitant in confirming my availability since I had never given piano lessons before, but I figured even if my French wasn’t so great that it would be fine since music is really a universal language.
Well… in the abstract it is. In preparing a cheat sheet of English/French music terms for my first lesson, I discovered that there is a fundamental difference in the way that musical notes are named in languages of Germanic vs Romanic origin. Luckily, to give beginner lessons to a 4- and 5-year-old, I didn’t need to be too familiar with this different convention. So far our weekly lessons (between holidays and travel) have been going really well and we have just begun to work on playing with both hands at the same time, quite a challenge! I honestly hadn’t expected to even touch a piano during my stay in France, and certainly hadn’t planned on giving piano lessons, but so far it has been a great experience!
[Sainte-Victoire, as seen on a beautiful mid-October day from one of the local trails which begins at the Bimont Dam.]
Another of my favorite experiences in France has also been unplanned, though perhaps it’s a bit embarrassing to admit that I had no idea what I was walking (hiking) into. When I first arrived at the airport in Marseille, I picked up a map at the information desk that showed some of the scenic areas near Aix. A beautiful lake in Saint-Marc-Jaumegarde caught my attention, and finally in mid-October I found a weekend to take the bus out to the nearby dam that holds this lake to spend an hour or two enjoying the warm, sunny day. I was surprised to find a small parking lot almost completely full when I arrived, which was my first clue that there was something more interesting nearby. Before reaching the dam, there were some signs that described its history and also marked the local trails one could access on the other side of it. There appeared to be quite an extensive system of trails in the area, so I decided to explore a short one that was marked as being scenic.
The quiet trail ended with a direct view of a nearby mountain and the countryside around it. On my map I could see that I wasn’t too far from another trail if I went straight up the bare, rocky hill that I was on, so I decided to go up for a better view. From this new trail, I could see the dam where I started from, as well as the nearby mountain. This trail was pretty busy with hikers and runners going by every five minutes or so, and it seemed to continue on toward the top of the mountain. Unfortunately, I had not come prepared for this longer hike, and with my water supply already half way gone I decided to just head back home. Of course it was only later when I did some research online that I realized the mountain was Sainte-Victoire, an icon well-known as the subject of many paintings by Paul Cézanne. While my “discovery” of the mountain and the extensive systems of trails that cover much of France has led to other adventures since then, that first unwitting excursion was probably my favorite single day in France so far.
[Two weeks after my “discovery” of Sainte-Victoire, I returned with more supplies to make the 6-hour hike on the GR9 trail near Vauvenargues across the mountain to Puyloubier. In this picture I’m standing behind the Croix de Provence, and you can see the lake and the Bimont Dam, as well as Aix, off in the distance to the left.]
Of course, I am also doing some of the actual work that I proposed to come to France for in the first place. And of course, the best results that I’ve gotten were from something that wasn’t even part of my research plan. Here in France I am working at ITER, which is the largest nuclear fusion experiment of all time. When you heat hydrogen up to millions of degrees, you can get the hydrogen nuclei to collide and “fuse” together, and this process releases energy. This is exactly how stars produce their energy. It is also a really promising source of alternative energy because it is carbon-free, has an effectively limitless fuel supply, and doesn’t produce any long-lived radioactive waste. Although other fusion experiments have produced millions of watts of fusion power before, ITER will be the first to produce more energy than it takes to start up.
When you heat a hydrogen gas up to millions of degrees it actually becomes a new state of matter: a plasma. A plasma forms when there is enough energy for the negatively charged electrons surrounding an atom to be stripped away, exposing the positively charged nuclei. Since charged particles experience a force when they pass through a magnetic field, you can actually use magnets to hold your hundred-million-degree plasma in place and keep it away from the walls of your reactor. However, just like any other fluid a plasma can experience turbulence, and this turbulence can build up and cause the plasma (and the fusion reaction) to suddenly extinguish.
My research is to take data from experiments around the world and to use machine learning algorithms to produce models that predict when a plasma “disruption” will happen. This model could be plugged into ITER’s control system to monitor the plasma and give the probability of a disruption occurring, or at least a simple “yes/no” prediction of whether one is coming. While my research plan is to build these models and test whether they make good predictions, I actually came up with a new way to interpret the models to tell you how to avoid an approaching disruption and therefore keep the fusion reaction going. I think that my new method to interpret these models could fundamentally change the way that they are applied, and change the way that physicists think about machine learning in general. In just a few short weeks I went from a simple idea to a fully fleshed-out analysis ready for publication, and I think this just goes to show the importance of the things that aren’t planned!