Separately to the lightning network, the security (hashrate, energy usage) is still not directly linked to the number of on-chain transactions that are processed. You can compute "a practical energy/block" specifically for bitcoin but it does not translate to a meaningful energy/transaction value because the quantity of energy used is not directly caused by transactions. You can have a less-secure blockchain that does many more transactions than bitcoin with low energy usage (altcoins exist) - if these are low value transactions then people won't even be incentivised to attack it. The relationship is NOT just energy/transaction. Your estimated energy/transaction value will keep changing because that's not the relationship; the new coins in the block reward will eventually drop to zero, the value processed by the network will change, the ASIC development situation will change, etc. It's like finding a runner that eats icecream and computing icecream/kilometre.
>Again, if you don't believe this, you should ask yourself what stops the bitcoin blockchain from processing 1 billion transactions/second instead of the current pitiful ~2 hundred transactions.
Nothing is stopping it. Run your own lightning node and you can do more transactions than what you are erroneously claiming