As an ex-CEGB engineer, I couldn't resist checking the numbers on this.

'Hard' coal is widely reported as having an energy density of 24 Million Joules per kg. This sounds a lot, but 1 Joule of energy will only deliver 1 Watt of power for 1 second, so you need 1000 Joules to deliver 1kW for 1 second and 3.6 Million Joules to deliver 1kW for 1 hour. This means that a kilogramme of coal contains 6.67kW hours of energy.

The next factor to consider is that the energy stored in the coal is converted to thermal energy when it is burnt and the process of converting thermal energy to electrical energy is quite inefficient. Padiham power station had two 120MW generators installed. My memory doesn't stretch to remembering the efficiency of the CEGB 120,000 kW generators, but a bit of searching suggests they were not far off the much larger 500,000 kW/660,000 kW units and could run at up to 33% overall efficiency. This means, of course, that you need to put in 3kW hours of thermal energy to get out 1kW hour of electrical power out. A kilogramme of coal will therefore only produce 2.2kW hours (i.e. units) of electricity - enough to boil a kettle continuously for an hour.

There are some useful details of Padiham power station posted on line. These suggest that Unit 1 was converted to oil firing in 1974 and was only used occasionally in the 1980's (due to the increase in oil prices relative to coal). Unit 2, however, remained as a coal burner and is reported to have generated more than 620 million 'units' (i.e. kW hours) annually during the early 1980's. Based on the calculation above, that would have required 280 million kg of coal - in other words 280,000 tonnes each year. Assuming this arrived at Padiham in old fashioned 20t mineral wagons, this would have required 14,000 wagon loads each year or 500 trains of 28 wagons each.

To draw these ramblings to a close, the figure of 15 million MW hours generated per annum is clearly wrong, the station could only generate 2.1 million MW hours if both units ran continuously (240MW x 8760 hours). With one unit (of the two) burning oil, the theoretical maximum coal consumption would be based on around 1 million MW hours of electrical output. Considering the actual output, Bald Rick's estimate of 1-2 trains per day is supported by the maths and was spot on.