Their is one huge glaring problem with your whole theoretical framework

COSTS != VALUE

The cost to create BTC's has no link what so ever to their market price.  This is a classic 'regression theorem' error by BTC zealots, just because something valuable is destroyed to make a product dose not make the product worth the cost of the inputs.  People make foolish production decisions every day which produce products worth less then the input costs, the free market weeds these people out eventually so in the LONG RUN the products are worth at least the inputs (and hopefully some profit margin if your a good business person).  But it is by no means an absolute that can be applied here.

By adding up all expenditures involved in mining BTC all you have determined is how much minimum value we would have to apply to BTC as a  service in order to consider the production to have been a sound investment.  The actual value we DO assign could be vastly higher or vastly lower then that production cost, it is a fully independent variable.  Your number IS useful (I can't speak to accuracy) don't get me wrong, your just calling it by the wrong name and drawing erroneous conclusions from it.

Also you may find this site to be a treasure trove of data, this guy has been doing analysis of the network for some time and has tracked the ASIC transition and produced estimates of electrical consumption.  http://organofcorti.blogspot.com/