The fact bitcoin will be so difficult to earn should make the price go significantly higher.
The fact LN and maybe similar solutions will be used on a larger scale may make tx fees even go lower.
So imho comparing the block reward with the tx fees is useless/counterproductive. I think that block reward + tx fees should be compared with the fiat cost of mining a block, and then we would see how much one Bitcoin worth.
Of course, various modifiers have to be taken into account, from the more performant machines by then to many more using very cheap energy from the sun. Also don't forget the inflation of fiat.
But that's all. I don't understand why you've tried to overcomplicate the equation, but I see no problem with block reward going lower than the tx fees, even though I expect that happen significantly later than your predicted time frame.