Read the forum code of contact
By: - 21st February 2005 at 21:24 Permalink - Edited 1st January 1970 at 01:00
What exactly was simulated? If a simulator is logically correct, it will not make wrong predictions unless someone input the wrong data.
By: - 22nd February 2005 at 09:49 Permalink - Edited 1st January 1970 at 01:00
well, i want to know whats the level of aero-design simulation, i dont know very well, but i think that there some cases like the f23 and f18 in wich the computing asistance made wrong predictions, there are other cases???
I am into computer simulation, although not exactly CFD (computational fluid dynamics) rather than structural analyses. Just look at my Avatar.
Computing assistance can never do wrong predictions as long as all constraints, loads and all boundary conditions are logically correct
By: - 22nd February 2005 at 09:56 Permalink - Edited 1st January 1970 at 01:00
I am in similar business. If you give wrong input, you get wrong out put. What software they were using?
By: - 22nd February 2005 at 10:11 Permalink - Edited 1st January 1970 at 01:00
I am in similar business. If you give wrong input, you get wrong out put. What software they were using?
I would expect either CATIA CFD solver, ANSYS CFX or, most likely Fluent, to be honest...
By: - 22nd February 2005 at 10:32 Permalink - Edited 1st January 1970 at 01:00
There are some pretty good solutions from hindustan aeronautics /national aeronautics. LCA was created on them. I have seen them work. AFIK some EU outfit has purchased them.
By: - 22nd February 2005 at 10:37 Permalink - Edited 1st January 1970 at 01:00
There are some pretty good solutions fron hindustan aeronautics and national aeronautics. LCA was created on them. I have seen them work. AFIK some EU outfit has purchased them. Let me check out ansys.
Indians are Gods in simulations.. I don't know why, but on XANSYS forum every second member is an Indian... Makes no surprise to me that they try to spread the market with their own solutions...
Any idea what the name of their CFD is?
As with ANSYS, it is a very widespread tool that has large capabilities but also some limitations. It is into structural statics, dynamics, fatigue, buckling, and nonlinear, plus cast flow, electromagnetics, fluid dynamics, optimization, DOE etc. etc. Now it's even some kind of basic multi-rigid-body solver incoorporated with the new 9.0 release, haven't seen it yet, waiting eagerly for our updates.
By: - 22nd February 2005 at 10:59 Permalink - Edited 1st January 1970 at 01:00
Wanna outsource ? :diablo:
http://www.ada.gov.in/Activities/Software/cfd/cfd.html
CFD Tool Kit and its Application
Aeronautical development Agency(ADA) has taken up the task of design and development of the Light Combat Aircraft (LCA), In order to carry out this task, ADA has provided continuous thrust to develop CFD tools for the aerodynamic design/analysis of LCA since 1986, through both in-house efforts and sponsored projects.
Significant progress has been made in the development of CFD codes for design and analysis and also in the area of grid generation and past-processing, As a design tool, ADA has developed a panel code for complete aircraft with ground effect at small angles of attack and side slip. For transonic flow regime, full potential field integral code of ADA, which uses cartesian field grid with surface panelling, can handle complete aircraft including external stores and control surface deflections. Full potential code based on cartesian grid is a very useful tool for design and analysis of complex aircraft configuration since its turn-around time is small. For higher angles of attack and side slip at all speeds, multi-block Euler code capable of handling complex configuration has been developed. Further, specialised codes such as RANS codes for internal/external flows have also been developed. For unsteady flow, codes have been developed for subsonic and transonic speeds for wing-body configuration (both rigid and flexible shapes) using transonic small perturbation (TSP) theory and vartex lattice method (VLM).
Tool Kit at ADA
Panel code capable of handling complete aircraft configuration with ground effect at subsonic speeds.
Inverse design code for laminar super crittical airfoil.
Full potential field integral code using cartesian grid for complete aircraft with external stores at transonic speeds.
Multi-block finite volume Euler code for realistic aircraft configuration.
Finite volume RANS code for intake duct
Implicit turbulent RANS code for external flow over 3-D bodies.
Grid generator with surface modeller.
Utilisation of CFD Codes at ADA
Wing design
Fuselage shape optimization
Evaluation of incremental changes in configuration
Pylon reshaping for reducing interference drag
Estimation of optimal deflection of leading edge devices.
Prediction of Aerodynamic loads
Calibration of Air data sensors
Trajectory prediction of external stores
Evaluation of performance of intake duct
other stuff
By: - 22nd February 2005 at 11:55 Permalink - Edited 1st January 1970 at 01:00
Wanna outsource ? :diablo:
I am not exactly into CFD too much, but yes, a professional deformation it could be called.
By: - 22nd February 2005 at 12:19 Permalink - Edited 1st January 1970 at 01:00
;) I know its a touchy subject in US. We do so much outsourced stuff. Anyway that was just a reply to Indians being gods in simulations. :D
By: - 22nd February 2005 at 13:52 Permalink - Edited 1st January 1970 at 01:00
;) I know its a touchy subject in US. We do so much outsourced stuff. Anyway that was just a reply to Indians being gods in simulations. :D
Finite Elements Modeling is dated back into early eighties, the mathematical theory goes as far back as with Gauss, Euler, Descartes and Fourier, definitely nothing that could be even called US-made, even if the first widespread codes on the market were from Pittsburgh.
I remember one of the first really optimised objects was the tennis racquet, which has been given the twin-beam shape after the optimization performed by a German aerospace Schlenker scientist using self-made FEM-code. Interesting enough, he was using the Yin-Yang principle of transfering 'hot' and 'cool' vibrational energy.
By: - 22nd February 2005 at 13:57 Permalink - Edited 1st January 1970 at 01:00
Is there anything openly available for Computational Electromagnetics (CEM)?
By: - 22nd February 2005 at 14:34 Permalink - Edited 1st January 1970 at 01:00
how are the chinese in simulation????
hello flex, brute gorilla well very interesting replies, but is also true that in a simulation you must chose the math model, sometimes diferents math models shows diferents solutions, aerodynamics -i think- is still an very empirical science (wind tunel, experiments, etc)
By: - 22nd February 2005 at 14:55 Permalink - Edited 1st January 1970 at 01:00
how are the chinese in simulation????hello flex, brute gorilla well very interesting replies, but is also true that in a simulation you must chose the math model, sometimes diferents math models shows diferents solutions, aerodynamics -i think- is still an very empirical science (wind tunel, experiments, etc)
Errrrrr... not really.. There are usually several methods for the same problem available, for example for a computational modal analysois you can use Block-Lanczos method, Subspace, Powerdynamics, Reduced or Unsymmetric matrix, Damped or QR Damped method, but it is not because they provide very different results, it is because the particular methods are coupled with different solvers, juts like Direct Elimination, Frontal Direct, Sparse Solver, or Preconditioned cojungate solver (whatever there are tons of those), which are suited for different problems. For instance, if I know that the system I want to solve is overdamped, I will immediately switch to QR Damped w/o any thoughts.
Aerodynamics is as well explored as any other fluid dynamics, it is by far not purely empirical science, but of course, experiments are always an important part of the whole chain, that will never change.
In general, first the FEM and MBS models are built and computed, then experiments and measurements with scale models are being performed to validate results, after those the models are updated and tweaked to match the measured reality.. After that you have a realistic model available and you are ready to solve various product modifications and to predict future capabilities and features with a solid accuracy. A simulation w/o experiment validation is often only worth being splashed down the toilet, but a validated computational CFD analysis is able to deliver you a later prototype that does not need practically any configuration modifications.
By: - 22nd February 2005 at 15:00 Permalink - Edited 1st January 1970 at 01:00
there some cases in wich the real performance didnt match with the simulation??and experiment?? in operative aircraft (at least in specifical conditions)??
By: - 22nd February 2005 at 15:09 Permalink - Edited 1st January 1970 at 01:00
there some cases in wich the real performance didnt match with the simulation??and experiment?? in operative aircraft (at least in specifical conditions)??
The last time I remember incident like that was the Q-5 Fantan static prototype breaking up under 70% predicted allowable load :)
Oooops..
By: - 23rd February 2005 at 05:18 Permalink - Edited 1st January 1970 at 01:00
Finite Elements Modeling is dated back into early eighties, the mathematical theory goes as far back as with Gauss, Euler, Descartes and Fourier, definitely nothing that could be even called US-made, even if the first widespread codes on the market were from Pittsburgh.I remember one of the first really optimised objects was the tennis racquet, which has been given the twin-beam shape after the optimization performed by a German aerospace Schlenker scientist using self-made FEM-code. Interesting enough, he was using the Yin-Yang principle of transfering 'hot' and 'cool' vibrational energy.
French have bought a ADA software called "AUTOLAY." A-380 was made on Autolay.However Autolay is a Autocad software, http://www.ada.gov.in/Products/autolay/autolay.html
By: - 23rd February 2005 at 14:09 Permalink - Edited 1st January 1970 at 01:00
AUTOLAY is used by Airbus, BAe, Antonov, Hughes, HDW (Germany), Daimler Chrysler and Aerospatiale. It is used for the conversion of a theoretical design of a laminated composite structure to a practical product.
ADA's CFD capabilties include Panel/Full potential/Euler/Navier Stokes codes, Inverse design and unsteady CEM codes, subsonic/transonic/supersonic flows, viscous effects, Grid generation and adaptation, pre and post processing tools etc
By: - 23rd February 2005 at 18:51 Permalink - Edited 1st January 1970 at 01:00
As long as, (of course you know), the Navier - Stokes equations are not solved in 3 dimensions, experiments will take place each time a new project appears. Any way, I couldn't say that given the cost of developing new software and hardware is cheaper to simulate in PC than making experiments in the wind tunnel.
Limitations like following scale laws (eg Buckingham rule etc), or making assumptions (eg assuming constant Reynolds number) have been studied so much that makes wind tunnel tests even more reliable and actually I suppose the "no risk" philoshophy will allow both methods, computational and scale simulation to co-exist in the future.
By: - 23rd February 2005 at 19:48 Permalink - Edited 1st January 1970 at 01:00
Err, Ansys!
I believe that some parts of aerodynamics is quite hard to simulate, especially moment coefficients such as Cmz etc. Also, all kind of transonic simulation can go very very wrong due to the extremely difficult Navier-Stokes equations that has to be solved. These problems occur with all kinds of CFD codes, including Fluent, that's why most military jet manufacturers still uses windtunnels and flight testing.
regards,
Castor
By: - 24th February 2005 at 10:40 Permalink - Edited 1st January 1970 at 01:00
The main reason why any development company still uses measurements is very simple - any simulation has to be validated.
I have solved dozens of projects with ANSYS, starting with small rubber shim optimization up to a harmonic response analysis of a complete servo-powered steering of the upcoming BMW 5 and BMW X3 generation, but there never was a case I did not need any empiric data. And second, there never was a case we have offered a solution without actually measuring a prototype. I mean, who can afford depending solely on simulation results and screwing something totally up?
As with CFD, I am only into structural, I have very little or no experience with the complexity of the fluid dynamics analyses, we have an external partner for all coupled analyses regarding vibrations resulting from processes in fluid dynamics (yes, we do optimize vacuum cleaners and dryers, too :))
Posts: 1,838
By: over G - 21st February 2005 at 19:26
well, i want to know whats the level of aero-design simulation, i dont know very well, but i think that there some cases like the f23 and f18 in wich the computing asistance made wrong predictions, there are other cases???