We used POD to compress result files, i heard some people adopted it for post-processing of large chunky CFD data which was not possible to process (a few dozens of TB if i remember right).
Actually, it is similar with
mode superposition method applied in vibration. A flow mode is analogical to vibrational mode.
We need both experiments and simulations. It is another level of discussion that even profs in conferences cynically judges each other over this
.
Excellent summary. We have employed some of this in current projects I am working on.
You are indeed correct we use POD and DMD in post-processing in a few specialised ways, given sampling rates of CFD w.r.t assumptions that may not hold (and limit further useful use) if you take discrete snapshot directly.
In either case would I be wrong if I said understanding this topic/piece would be an important step in developing UAV, and perhaps an engine as well.
Yes, but not directly applicable (more a mid or late stream concept for specific application to a particular system).
You would have to treat/isolate a feature/problem etc of an existing base design. How to make a broader module (with well understood limitations) thats adaptable to whole range/class of systems will need research and development too.
i.e you would have to start from first principles way back, and at some point in design refinement you see the advantage to isolating a particular feature/environment interaction for the system for further resolution.
As just one example, one of my doctoral advisors was heavily involved in a specific (but pretty serious) LERX buffeting mode that diminished the performance of the early F-18 (back when it was being heavily tested) in a particular regime of flight (that didnt show up in the modelling earlier during those days...and even today).
They basically took approach something like this paper to gain the proper resolution of the issue and implement a patch fix for it (after considering all the tradeoffs w.r.t changing the design fundamentally versus "patching") that still works to this day (carried over to SH even) and is near forgotten about basically. I could talk at some length on the matter, but it will be boring.
Its actually a more relevant (in somewhat earlier iterative step in development) process for Engines like you said. Thats my line of work...using a range of tools and research methods (some macro, but mostly micro analysis using good resolution discipline and stochastic control) to improve jet engine combustion stability and performance....because the flows inside can get quite complex (they are still not understood that well esp when you increase resolution) and there is lot of room for continual improvement given nature and legacy of iterative small % improvements in this field.
But it is really mid and final tier stuff, most of the early development for engines for any country/company is more geared towards setting the scope for materials research, thermodynamics/turbomachinery research and fluids research.
I haven't read the whole piece (it's pretty complicated for a non engineers), but wouldn't air currents behavior change depending on the altitude.
Broadly yes...but there are lot more air current features (of early basic relevance to UAV performance) to incorporate than this specific one done in this paper. This is more if you want to keep improving a UAV for example because you found a specific regime of aerodynamics it faces of relevance past the initial scope.
Would the air current from an engine effect the UAV ? where should the engine be placed to create least effect on a UAV.
Depends on number of design tradeoffs. Ideally you want engine as far away from lifting/aerodynamic surfaces as possible and simply interacting only with propulsion realm (why its there for). But there are several design considerations (weight distribution, basic geometry, orientation etc) that come into play that counter-influence and the engineers thus need to come to a compromise. These would all have far much more play in the early basic design of aircraft compared to a refined iterative improvement later. Its somewhat rare you would have POD or DMD process setting an early or even mid design envelope for example.
