Solar Powered Trains

I had the "bright" idea one day that trains could perhaps just power themselves on solar equipment, so I ran some viability tests. Average CSX train, employing a GE AC4400CW diesel-electric locomotive, employs 4,400 horsepower supplied via a single exceedingly large diesel engine and multiple electric traction motors. This setup can comfortably carry around 2000 tons, or 20 fully-laden 100 ton coal cars, while consuming around 200 gallons of diesel fuel per hour at cruising speed. Presuming non highway-taxed diesel is about $2 a gallon, that's $400 an hour of operating expenses. Now let's say instead of 100 tons, you had a passenger car weighing only 40 tons, and you carried only 20 of them. The load on the engine is down to 800 tons, and I'm going to presume that means half the horsepower is required. The footprint of a standard boxcar is 55' 5" x 10' 7". I'll simplify this to 10.5*55.5 square feet of area, or about 580 square feet per car. Presuming that is entirely filled with today's economical solar cells, which are typically around 17% efficient, in full bright sunlight that would output around 16 watts per square foot. Let's go with 10. Cloudy days, and nights, will obviously not work. With our example 20-car train, the total power output of the cars is 116,000 watts, which translates to 155 horsepower. Pathetic, but considering our theoretical engine is only using half of its' power because of the lessened weight restrictions, that's roughly 7% of the energy for free. 7% of the supposed fuel consumption of $200 an hour comes to $14 an hour in savings. For a 116,000 watt installation, that would likely cost about $1/watt of panel, that would take over 8,200 hours, or 341 days of constant operation, to pay itself back. If, instead, we went with a 5-foot wide strip of solar panels that went down the edge of the track and transferred energy via overhead wires or a third-rail setup, each train would require 6.2 miles of solar to power it's full 2200 horsepower requirement. Considering most trains, especially long-haul passenger trains, frequently operate more than three miles apart, that sounds viable to me. The installations would cost $264,000 a mile, and represent savings of much less than that. I'm really tired actually and this isn't very viable so I'm pretty done, thanks for reading. I want solar powered trains. Put sails on them too, why not.