CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (

John C. Chien September 1, 1999 23:14

(1). As the foot steps of the century is coming to the end, for some people, the future of CFD becomes an important question. (2). As we have just celebrated the 30th anniversary of man on the moon, can we learn something from the history? And ask, What is the future of Appolo moon landing program? (3). Looking back, there were 10 million dollars super-computers for CFD back in late 70's. What are the future of those super-computers? (4). As the new fighter gradually sinking into the sea of over-cost, would you like to propose CFD for cost-reduction? Or should the CFD work should be cut to save the cost? (5). The fact is, similar to the space program, the main stream CFD disappeared at the end of 80's. The left-over somehow was transformed into various shapes of commercial codes. Under the skin of the life-less code, there are signs of struggle for life, for survival, for recognition, for identification. (6). You can hear it, you can feel it, and you can smell it. Like a small weed in the hot desert sun, dried and waiting for a summer afternoon rain storm. (7). To be realistic, wise persons still around. But they know how to hide their knowledge for the bad time to come. It is useless to spend every night to pack it into a CFD code. (8). If a 10 dollars recipe book can turn a man into a French chef, French restaurants would be long gone many centuries ago. (9). As the century is coming to an end, is it really that difficult to find a French restaurant with a good French chef? The same is true with the future of CFD. (10). As the door is closing, it will seal that precious knowledge into the bottom of human brain. (11). As the summer afternoon rain storm come at the begining of the next century, something like desert yellow flowers will again cover the desert land. Will the code still be there?

X. Ye September 2, 1999 09:15

About the future of CFD, I have a vision:

Now a CFD package does normally only in the following route:

geometry + boundary conditions ---> solution of flow field

I hope a CFD package in the future reverses this route as follows:

boundary conditions + hoped flow field + expert system ---> optimized geometry

I know it is very difficult but it is an efficient way.

X. Ye

Fabien Coppens September 2, 1999 09:35

John, some of your postings take on an interesting poetic form, and your metaphores remind me of zen bhuddism descriptions of nature and life in general. But please, can we bring this forum back to mostly TECHNICAL discussions in the field of CFD ? Thank you !

Peter Menegay September 2, 1999 10:00

I think it's obvious that CFD, and the whole field of computational physics, is here to stay and will grow. The death that you speak of at the end of the 80's was the end of the cold war. Naturally the defense driver doesn't exist the way it used to. But commercial applications are growing. And with the rise of the next superpower, the defense driver will come back. I'd be a little more optimistic about the future of CFD.

Janna J. September 2, 1999 10:53

agree and we started to try, very preliminary though.

Demselles Hollowajiawoak September 2, 1999 11:36

I would like to know why your e-mail address has suddenly changed. I have the sneaking suspicion that you are not who say you are or perhaps you have gone off the deep end. If the former is true please state your true name, and not putrify someone else's name.

John C. Chien September 2, 1999 11:55

(1). Thank you for your interest. (2). On Internet, cfd-online is not cfd-online. It is a bunch of numbers only. (3). On Internet, the e-mail is free, the software is free, and the computers are also free. (4). So, I guess you are new to cfd-online. Why not try the search option on this forum and do a name search on John C. Chien or Chien. It is a good idea to dig into the archive section. (5). Nothing is real on the Internet. That is why they are called virtual environment.

John C. Chien September 2, 1999 12:02

(1). Good idea. (2). But, I think it has been used in industries for many years already, for example, the inverse design of turbine blades. (3). There are codes available and they also have been used in actual design processes for many years. (commercial codes? no.)

John C. Chien September 2, 1999 12:03

(1). Good idea. (2). But, I think it has been used in industries for many years already, for example, the inverse design of turbine blades. (3). There are codes available and they also have been used in actual design processes for many years. (commercial codes? no.)

John C. Chien September 2, 1999 12:27

(1). Thank you very much for your suggestions. But there are valid reasons for that opening speech. (2). The space was brought about not because someone like Godard was interested in rocketry. It was a simple decision of a president. A leader who took the technology into a new era. (3). Likewise, the aeronautical wind tunnel facilities were also created largely by some general in the Army. (4). Without the leadership, it is very hard to make any progress in technology. (5). So, the point I was trying to make was: the future of CFD needs a powerful leader. A good example? Former CEO of Chrysler, or the current CEO of Microsoft. Or like Navier, Stokes who invented Navier-Stokes equations more than 100 years ago. (6). Technology comes from the applications of physics, which is expressed in mathematics, which is derived from logics, and which is created from the mind of innovation. Let your mind float above all of these small detailes of everyday life. (7). I have not heard any voice from learders of commercial industries that CFD will play an important role in their future business decisions. Without that kind of support and vision, one can always work in one's backyard on his commodore64? or Amiga1000? AppleII... ( I heard that the former Chrysler CEO is selling electric bicycles. But, then we have to find the CFD problem for the electric bike first. I am quite sure that he is not selling CFD.)

John C. Chien September 2, 1999 13:01

(1). Sorry to ask you this question. (2). In my thiry years of research work in CFD, I have never come across this new theory called " Zen bhuddism ". But I have seen books on "Zen" in business section of bookstores. I have never studied this kind of theory. Is it popular in France? I did visit some beautiful temples when visiting Taiwan. Do you think that there is some connecting between the "Zen bhuddism" and "CFD"? Is this the opposite side of "chaos" or "fractal"? (3). I am just curious about it. And you are the first person mention this theory here. If it is not technical, you don't have to answer it.

Janna J. September 2, 1999 16:34

They are those stubborn obstacles for the technology 'evolution' in industries, non-sense to keep only for themselves. They could make big fortune by selling them out. They still have the chance. Big 3 are dumping their own secret codes and turning to commercial ones. Most suppliers will be forced to follow. Commercial codes will have more money to hire more smart people to develope more sophisticated codes to the market at cheaper price. They also provide group brains for professional tech support. Secret keepers will soon fade out.

T. J. Wanat September 3, 1999 00:35

I think that CFD has an excellent future. As computing power becomes cheaper and algorithms become more efficient more industries can afford to take advantage of CFD and improve their designs. I seems that in the past two years that the number of commercial CFD codes has more than doubled (this is my perception, I haven't counted, it could just be better marketing). After some reflection, it may be that Mr. Chien is lamenting the end of the pioneering era of CFD and not the end of CFD altogether (though he may not realize it). From reading Mr. Chien's post's, it seems that he has been involved in CFD for many years, was involved in the defence industry and strongly advocates the ability of an analyst to be able to write his/her own code. He may have been involved at a time when there were no commercial codes expensive computers were required (which only the government, universities and very large corporations could afford) and it was the domain of a small elite group. Now anyone with a bachelor's degree in engineering who can point and click can call themselves a CFD analyst. This is similar to the original group of Internet users who feel it has been ruined by the unwashed masses who now litter the Usenet group and spam others en mass. Just my two (uhh, make that five) cents.

Zhong Lei September 3, 1999 01:29

This is one of the oldest problems in CFD. I think the final purpose of CFD is to design fluid dynamic systems and any of their component but not only to verify how the flow is or compute what the flow is. It falls in the category of optimization design and flow control. It has been used in industries for many year, for example Boing and MD. Although CFD companies just started to creat their CFD design software recently, I believe that it will extensively used in the industries in the near future. As I know, for commercial softwares, NUMECA Co. Ltd has developped a 2D version and they are planning to extend their work to 3D. Some other companies (FLUENT ?, STAR-CD ?) also begin to do this. By the optimization techniques, the shape is modeified to improve the performance of fluid systems to approach targets including prescribed pressure or velocity distributions, lift range, maximum lift, minimal drag, shock-free suction side in transonic flow and type of stall at subsonic speeds under geometrical constraints. Differently from the method of traditional linear design, the optimization design combines optimization thoery and CFD so that the nonlinear effect is introduced in the design process. The optimization methods currently used are gradient-based method, genetic algorithm, neural networks thoery, and so on. We can get a optimum shape by computer just in a few days. However the biggest problem is still the accuracy of CFD results. You may read J. of Aircraft, 1999.1. I am working in this field and currently doing supersonic aircraft design only for acadmic research. But if I got a good chance, I would like to work in a company to creat such a software.

Z. Lei

John C. Chien September 3, 1999 03:35

(1). You are always right. It is always important to leave room for readers imagination and draw his own conclusion. (2). Read the next posting of "Salary Analysis and CFD jobs". The future is always in your own hands, not in my words.

Janna J. September 3, 1999 09:47

try adapco or fluent. their core developing centers are in England though.

clifford bradford September 3, 1999 11:53

interesting concept. but why limit to 'hoped flowfield' (inverse design) in which case you get what you want which is not always the best and to 'expert system' which is only one technique of automatic design, why not say "automatic design system" which can incorporate many tecniques.

i agree with you though that analyses is not the prupose of design results are. as well as good analysis tools we need automatic design tools to produce better products faster. it is all well and good to have great CFD code (the structural people already have this in FEM) but what to do when system is complicated and redesign is tricky and time consuming. fluid flows are very sensitive and even a good engineer often has difficulty coming up with better design. this optimisation/automatic design field is growing but there's more to come

clifford bradford September 3, 1999 11:54

more than just inverse design John - we need optimization

clifford bradford September 3, 1999 12:03

i rather liked this posting. i think after you read john long enough you get to understand his postings. the first time i read his stuff i was bamboozled. now it's not so bad. i think in additon to the technical stuff we also need discussion on the philosophy of cfd. i remember talking to a gentleman at Pratt and whitney who said it's notupwind or central diffrence or lax wendroff or runge kutta its getting the job done as good as possible (i paraphrase). we do need the nitty gritty but sometimes we also need the zen to truly understand. otherwise we'll might end up as one of those people using 4 stage runge kutta with artificial dissipation to solve flow of water in pipes

clifford bradford September 3, 1999 12:08

more than anything else we need people who understand what the three letters in CFD mean together with each other. you can know C but without FD you don't know CFD and vice versa. sorry to say also that many have neither C or FD

All times are GMT -4. The time now is 00:35.