1.
Abstract
Trading-off the system parameters, devising the
(optimal) ascend route/protocol & hence exerting optimal control throughout the flight
regime, are the most fundamental issues pertaining to an aerospace vehicle. With i.
payload ii. frequency iii. efficiency & iv. funding the name of the game for the new
millennium, grassroots innovation & trusted technology will hence rank as the primary
substance for a new shuttle vehicle. With the aero-shell & propulsion planform assumed
as fixed with the X-33, the design endeavor should hence focus on the
technology/innovation adjuncts; with optimization the primary objective.
Whereas the necessary procedures/techniques
(& the associated computational resources) to enumerate the desired presets have been
available for some time, doing so recursively/adaptively (or more so doing so in
real-time) may have come at an excessive cost/complications; or simply may not have been
feasible at all.
However, merging the attributes of
1) stochastic estimation & control
2) the separation principle
3) super-positioning
4) dimensional analysis
5) decentralization &
6) Dynamic programming (the “SD-3” synthesis),
a piecewise hierarchy is being rendered whereby
(exhaustive) CFD et al partials may simply be equated up front & stored/indexed as
look-up tables for subsequent adaptive enumeration; albeit with an order of magnitude
lessor computational resource.
2. The X-33/Shuttle Corollary
Focusing on the X-33 (assuming it’s a
LACE-type aspirating vehicle) the primary design considerations that have to be resolved
up-front comprise ia;
1) the degree of aspiration/LACE
2) O2 vs LA (liquid air) trade-off
3) the thrust/weight ratio
4) fuel options
5) the payload
6) the stress regime
7) the thermal envelope
8) the ascend path proper
9) the mode of take-off/landing.
However, the $10 question relates to the O2/LA
required to go orbital; eg should the O2/LA a) be ferried thru take-off or b)
“scooped” (& distilled ..sic) en-route. Additionally with relation to the
stated poser, if option b) is affirmative, at which/what stage of the ascend path & at
what rate ?!
3. The HVAC Analogy
However daunting the preceding problem
statement may seem, the likes in HVAC has already been solved on the equivalent of an
AT-PC thru the mid eighties by writer. Perceiving the essentials of stochastic/recursive
systems, super-positioning & the separation principle at UCLA in 1972, the concept was
promulgated thru the University of the Witwatersrand (Johannesburg S.Af) subsequently.
With Dynamic programming as the dominant hierarchy, “SD-3” as tools & the
(thermal) transients as the principal variable(s), a computational system was structured
whereby 1) the envelope threshold of HVAC systems may be enumerated (more specifically to
verify the validity of “free-cool HVAC”) 2) to optimize/rationalize component
selection & 3) as a BMS kernel.
Having run the Dp algorithm originally on a 1st
generation “Apple” in the late seventies, the comprehensive system was perfected
thru the mid eighties. In accordance with the Dp protocol, once the substance/range of the
primary plant/building related variables have been identified, the control/management
strategies are traced (& logged). The latter sequences (entrenching the desired
forward looking capability) hence became the heart of the control model, inasmuch as in
accordance with Bellman’s principle of optimality that ..whatever the initial state,
the remaining control must constitute an optimal one with regard to the state that results
from the initial control.
The relevance hence of the X-33 (or any
aspirating shuttle per se) is that akin to the HVAC problem, the context/extent of the
O2/LA may be optimized in a similar fashion as the transients (eg “flywheel”
effect of a structure) in an air conditioned building !! Ditto the ascend strategy &
flight management. With the HVAC application tried & tested, the “Shuttle”
analogy can be expediently developed. Refer enclosed exposition re the SD-3 elements. |
