There are issues with scaling up a core past the flow aerodynamics and temperatures.
There are serious issues with things like vibrations and bearing fatigue as well.
Give multi-variate multi-dimensional transfer functions involved, it can be modelled with monte carlo simulations to some degree....to give sense of where human resources are best arranged and used in 5 - 10 year time chunks. No idea if TEI has set up an experienced team to do that kind of thing, what is the overall computing power they have steady access to and process flow and handling for this. Lot of it comes only with time and some level of mistakes made, there is lot current PW builds upon from earlier PW team foundations etc and all documentation related to (real or likely) dead ends and why.
There is an issue with lot of montecarlo sims too the more complex they become....tied to things like the sample error reducing by a magnitude (10) only if you increase the sample by twice that magnitude (100)...and what this cascades to for computing (to direct and optimise lab work and engineering later).
Large part of my current job it to revisit a number of old code we have running reliably to sandbox and optimize them better (especially some of its random number generators in the old FORTRAN that need updating now) because processing time is always at premium and given way montecarlo and other stochastic sims can cascade wildly to begin with, its important RNG intrinsic to their inputs are actually as random as possible.
There are serious issues with things like vibrations and bearing fatigue as well.
Give multi-variate multi-dimensional transfer functions involved, it can be modelled with monte carlo simulations to some degree....to give sense of where human resources are best arranged and used in 5 - 10 year time chunks. No idea if TEI has set up an experienced team to do that kind of thing, what is the overall computing power they have steady access to and process flow and handling for this. Lot of it comes only with time and some level of mistakes made, there is lot current PW builds upon from earlier PW team foundations etc and all documentation related to (real or likely) dead ends and why.
There is an issue with lot of montecarlo sims too the more complex they become....tied to things like the sample error reducing by a magnitude (10) only if you increase the sample by twice that magnitude (100)...and what this cascades to for computing (to direct and optimise lab work and engineering later).
Large part of my current job it to revisit a number of old code we have running reliably to sandbox and optimize them better (especially some of its random number generators in the old FORTRAN that need updating now) because processing time is always at premium and given way montecarlo and other stochastic sims can cascade wildly to begin with, its important RNG intrinsic to their inputs are actually as random as possible.