Don't know if the Japanese have thought of this already,(besides I am tired of researching to see if this is already done in the Shinkansen while I struggle to meet my daily quota ;) ) but I am sure that there is a lot of vibrational energy beneath the tracks and sleepers allowing for a harvesting setup that independently provides some chance of recovery.
After trying to look up energy recovery through hydrodynamic damping, I thought piezo maybe the way to go and found this piezo-generator very useful for a back of the envelop calculation!
http://www.piezo.com/prodproto4EHkit.html
Extract :
Piezo Energy Harvesting Circuit
The self powered Piezo Energy Harvesting Circuit collects intermittent or continuous energy input from the piezo generator and efficiently stores their associated energy in an on-board capacitor bank. During the charging process, the capacitor voltage is continuously monitored. When it reaches 5.2V the module output is enabled to supply power to an external (user) load. At this point 55 mJ of energy are available. When generator energy input is high, the output voltage remains ON continuously. Capacitor voltage is clamped at 6.8V. If external power demand exceeds generation, the output voltage decreases. When the output voltage drops to 3.1V, power to the load is switched OFF and is not turned on again until the capacitor bank has been recharged to 5.2V. The circuit accepts input voltages from 0V to ±500V AC or DC and input currents to 400 mA.
So what will a possible system look like if more energy is to be harvested -
and how much energy is available ?
Here's a back of the envelop calculation -
The piezo is about 3 inches (7.62 cms) , which allows us to lay 13123 pieces per km, which means about 9.8 (say 10) million pieces below the entire track length (presumed here to be 750 kms including the maintenance and shunting yards). Multiply this by 4 (because you will have at least 4 rails for 2 tracks one going up and the other down) you have 40 million pieces capable of generating-
40 million x 55 mJ = 2,200 million mJ or 2,200 Mega Joules
Now, Joules per second is Watts , so the potential is 2200 Megawatts!!! or am I dreaming again with my time integrals?
Ah! I know my wise friends will scream "cost", like the proverbial Faust, as they would have scrolled to the bottom of the above link and found the cost per piece for a quantity of 100 pcs to be $ 349, and adding another $50 for the system we round to say $ 400 then, we have $ 160 Million as investment for harvesting this energy!!
My experience with component level costing tells me that this price should fall to 1/2 or even 1/3 of the price indicated at 100 pcs when production is ramped to a million pieces and more importantly it is "Make in India" - I know one company that is capable of this in India, but are they watching, its anybody's guess.
In the end, this need not be under the track, it can be under the sleepers, it can be under the thrust pads of the bearings on some of the pillars and also, it can be thoroughly tested to full implementation before the second leg of the bullet train is operationalised anywhere in India or the world!