We’re about two-thirds of the way into the Autumn Semester here at Maynooth and, by a miracle, I’m just about on schedule with both the modules I’m teaching. It’s always difficult to work out how long things are going to need for explanation when you’re teaching them for the first time.
One of the modules I’m doing is Differential Equations and Transform Methods for Engineering Students. I’ve been on the bit following the “and” for a couple of weeks already. The first transform method covered was the Laplace transform, which I remember doing as a physics undergraduate but have used only rarely. Now I’m doing Fourier Series, as a prelude to Fourier transforms.
As I have observed periodically, the differential equations and transform methods are not at all disconnected, but are linked via the heat equation, the solution of which led Joseph Fourier to devise his series in Mémoire sur la propagation de la chaleur dans les corps solides (1807), a truly remarkable work for its time that inspired so many subsequent developments.
In the module I’m teaching, the applications are rather different from when I taught Fourier series to Physics students. Engineering students at Maynooth primarily study electronic engineering and robotics, so there’s a much greater emphasis on using integral transforms for signal processing. The mathematics is the same, of course, but some of the terminology is different from that used by physicists.
Anyway I was looking for nice demonstrations of Fourier series to help my class get to grips with them when I remembered this little video recommended to me some time ago by esteemed Professor George Ellis. It’s a nice illustration of the principles of Fourier series, by which any periodic function can be decomposed into a series of sine and cosine functions.
This reminds me of a point I’ve made a few times in popular talks about astronomy. It’s a common view that Kepler’s laws of planetary motion according to which which the planets move in elliptical motion around the Sun, is a completely different formulation from the previous Ptolemaic system which involved epicycles and deferents and which is generally held to have been much more complicated.
The video demonstrates however that epicycles and deferents can be viewed as the elements used in the construction of a Fourier series. Since elliptical orbits are periodic, it is perfectly valid to present them in the form of a Fourier series. Therefore, in a sense, there’s nothing so very wrong with epicycles. I admit, however, that a closed-form expression for such an orbit is considerably more compact and elegant than a Fourier representation, and also encapsulates a deeper level of physical understanding. What makes for a good physical theory is, in my view, largely a matter of economy: if two theories have equal predictive power, the one that takes less chalk to write it on a blackboard is the better one!
Anyway, soon I’ll be moving onto the complex Fourier series and thence to Fourier transforms which is familiar territory, but I have to end the module with the Z-transform, which I have never studied and never used. That should be fun!
In the Dark 