We solved this problem in our physics class back in 1981/82. The professor had a model he built some years before, though it was only an axle with square wheels.What I thought was cool was that when you tilted the track with the inverted catenaries, the wheels started to roll. He even had a chain (Lat. catena) that was pegged to an arc of the track, so that when the track was inverted, the chain matched the curve, "proving" it was a catenary. Always been a favorite for me.