A rule of thumb is you could do 105% of that for 20min so 252w.
In "The 33 - Bike Racing" forum, lots of people seem to arrive at more like 90-92%. Anything from moi as a Cat 5 to some of the Cat 2 guys I follow on there.
Going shorter down the ladder in duration from 1/2 distance to 10mi TT, you're likely to see less improvement in someone who doesn't have as much anaerobic focus (true triathlete vs. true bike racer). If you're fast in a 10mi TT, you're relying on more anaerobic capability and can probably rely on it over some roller hills and then recover a touch more than in a longer event.
I think the exception would be single-sport British time triallists who do both the 15k stuff and the 50mi TT stuff (think Obree, anything from World pursuit champion up to British 50mi TT champion).
It's good to set goals..........I'm just saying this stuff because I've been there. I tried to apply math, percentages of power, and lots of mental gymnastics to set goals that were NOT based in reality.
Instead, recently, I got a baseline off of a few real world 15k and 40k rides for my TT goals.
Now, I'm basing my goals off of that real world performance. Actual power and HR data from those rides. How much wind and elevation actually affected the ride.
I can then use online tools to get an idea of a strategy and goals for what I do.
It should go:
real world trial (at desired distance) -> data -> apply a model or calculator to data -> make a plan for a ride from data-> ride -> repeat
It probably isn't as good to:
Zwift ftp test -> apply a model or calculator to data from a different
ride -> make a plan for a ride -> ride
Geeez.......a 15k TT is going to be like 40 TSS and 25min of your time. A 40k, probably just at/over an hour. Etc...people spend that much time talking on Slowtwitch.