CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Noise (

Heinz Wilkening January 17, 2000 08:26


in dense populated areas like larger cities, or near airports noise is a major emmission to the people living there. For this reason there are noise lemits defined by law. But often it is not know in advance how much a new street/airport will increase the noise leavel in a certain area.

Often noise generation is cloosely connected to vortex generation which can be simulated by CFD, but can CFD also be used to simulated how the noise is transported? I do not know anything much about that, only that noise/sound can be described by an own set of wave equations. Can these be connected to an CFD-Code.

Thanks for anything like some remarks or references.

Ciao Heinz

Jonas Larsson January 17, 2000 09:57

Re: Noise
Yes, you can do a CFD simulation (preferably an unsteady LES simulation) and enclose it within a surface. Then outside this surface, where you can assume that all the sound-generation and non-linear effects are negligeble, you can just solve the linear wave-propagation equations to find the far-field noise distribution.

Getting this to work is very difficult though. To capture the acoustics waves correctly in an LES simulation you need good boundary conditions, a good scheme, ... It is very much still an area of research.

We recently had a MSc student here who did his thesis on something like this - an LES simulation of jet-noise. Results were not that good. But it was just a first shot.

Another approach is to make a RANS computation and somehow model the sound-generation.

Victor Serov January 17, 2000 12:34

Re: Noise

from your point of view does it make sense to do a transient RANSE calculation of e.g a valve and estimate whether one geometry variation is better than another concerning noise emission by looking for periodical pressure fluctuations AND evaluating the level turbulent kinetic energy. (higher k -> higher noise emission !?!)



John C. Chien January 17, 2000 13:32

Re: Noise
(1). I have noticed that the car speed and the pavement material of the road can affect greatly the noise generation. So, the tire design, the road surface material can be improved, along with the lower car speed limit. At the same time, clean hybrid car powerplant must be used to reduce the pollution and noise. (2). For the airport, the best solution is to relocate it to a remote area. In this way, it will be safer for the airplane to land. And the connection between the airport and the city can be served by the underground trains. (3). The motor cycles should use battery power, this will reduce both the noise and the pollution. (4). Without these changes, people in a big city will move out, or die of cancer eventually. And a new city will be built somewhere with no cars and airplanes allowed.

Alton J. Reich, P.E. January 17, 2000 16:02

Re: Noise
It is possible to use a CFD code with some post-processing to get far-field flow noise, and it has been done. I worked in a previous life (job) with someone who did just that as his PhD thesis, and the company I work for now has a post-processing module that computes flow induced noise. Both work best as a tool to compare designs (ie: a relative comparison rather than an absolute).

I don't think that anything about our noise prediction tool has made it to our web site yet, but I can fax you a paper if you'd like. Alton

Jonas Larsson January 17, 2000 17:27

Re: Noise
Makes sence to me, but let me emphasize that I'm not at all an expert on aeroacoustics. All I know is what I've learnt from my colleague, who is an expert. I think that you would need to validate the method for your application though. Concidering how inaccurate most turbulence models are at predicting turbulence levels (even much more inaccurate than predicting turbulent viscosity) I'd say that you really have to know what you are doing and have a lot of experience about possible pitfalls. I guess I'm saying obvious things here so I'd better stop. Perhaps someone else with more hands on experience can give you a better answer.

Frank Y. Wang January 17, 2000 19:15

Re: Noise (VKI Lecture Series)
This may be of interest to the readers of this thread:

Juin Kim January 17, 2000 21:28

Re: Noise of John C. Chien
Hello, Mr. Chien

Please re-read(at least 5 times) Heinz's e-mail. Your reponse is totally NOT related to his questions whatsoever. It is very similar that what you suggested to someone who needed some helps to tranlate French codes. You said "Find a French lady".

I don't understand why you keep posting that kind of messages in this dicussion group.

Please suppress youself and please post your ideas when you are sure and to where you have some experiences.

People like me usually expect some levels of technical helps from this discussion group. Usually not your life story.

I have been thinking about sending this type of message to you from last year. If you keep doing that, I will contact with the webmaster of this discussion group.

John C. Chien January 17, 2000 23:20

Re: Noise of John C. Chien
(1). First of all, I have to say that I am pleased to receive your posted letter, because you are reading my posted messages. (2). On the other hand, I have to say sorry that you were not given a chance to express your feeling about my style of posted message. Since the forum is rather dynamic, you should take actions right away to divert the discussions in the direction which you can benefit most in receiving the information. (3). At my level, I am only concern about the general health of the forum, thus, the small technical details are often not the most important issue. It is the direction, the approaches, and the common sense, which are most important. (4). Since I am retiring from this forum, the webmaster is spending more time answering the questions, which is a good sign. And I am also looking at the possibility of having a webpage of mine in the future. But right now, I am too busy to do either one. So, I think, you will not be able to hear my message sometime in the future. (5). The rest of the message is related to your questions, this is because you stated that " I don't understand why you keep posting that kind of messages in this discussion group". (6). The original message by Heinz talked about the important issue of noise pollution on people liveing around the densely populated city and the airports. And we both agreed on this issue. So, there is no conflict. (7). He said " for this reason, there are noise limits defined by law". This is one way to deal with the noise pollution only. Along this public health policy, I went further to propose several viable solutions, which not only will greatly reduce the noise pollution problem on people living around the big city and airports, it will also provide future directions for those who are wroking in CFD areas, because these new technology will require help from CFD engineers in the future. Thus , I have opened new areas of research for CFD in several positive directions. These are tire -aerodynamics- ground interaction with moving object and elastic structures. (8). Actually, the noise from the tire and the road also affect the comfort of the driver and car passenger, thus it is important issue for the car manufacture when selling the vehicle. No one like to buy a very noisy car. And I think, the acoustic in the car structure also is an important field, not to mention the acoustic in the engine compartment. (9). To limit the noise level by law (as presented by Heinz)has been tried before at and around the airports. But this requires the airplane to fly in more difficult ways to avoid the residential area. It can be very dangerous especially with sudden wind change and heavy rains or snow. So, instead of finding an optimum condition to land an airplane and at the same time reduce the impact of noise to a minimum on people living around the airport (through further study of acoustic CFD, as was implied by Heinz), my solution of relocating the airport to a more remote area is a better solution. In this way, the safety of the airplane and the noise pollution problem will be solved at the same time. (10). To further study the noise generation problem, Heinz proposed a possible noise generation mechanism, he said " often noise generation is closely connected to vortex generation". This simply could not be true. I have lived in large cities such as Washington D.C., Philadelphia, Los Angeles, and the methods they used to reduce the noise pollution has been to build more walls, around the residential areas, along the express ways, roads. So, it simply says that the source of the noise pollution in big city comes from the interaction between the tire, the aerodynamics, and the road surfaces. The impact on the tire treads on the road create the noise, not the vortex generation. (11). As for the noise source from the aircrafts, it is mainly from the jet engines of the aircraft. And the modern jet engine has sound absorbing material designed into the cowl, duct, nozzle, etc. The sound generation and propagation from the exhaust jets has been studied for many years by NASA. I don't think, we have to use the CFD to simulate the turbulence from the first principle again. (12). Even the noise pollution problem created by the motocycles, which are popular in many big citys in Asia, is not generated by vortex generation. Anyone who is familiar with 2-stroke IC engine knows that a pressure wave is generated through the exhaust pipe in every revolution of the engine. The muffle is used to even out the pressure wave, not the vortex. (13). By the way, I studied German from a German lady at college, so, if you have trouble to understand French, you need to find a French lady. As I said, my time is very expensive, and I don't think I can change anyone in the way he or she thinks. By the time you finished reading this posted message, it is already to late to erase it from your memory bank. So, one needs to learn how to forget. (forgive and forget are about the same, maybe forgive is easier on ones memory and will make it easier to forget.) I am sorry to say that my posted message has occupied part of your brain, and I wish it will be gone soon.

Juin Kim January 18, 2000 00:57

Re: Noise of John C. Chien
Mr. Chien

It is good to hear that you are retiring from this forum. I think I can expect more technically oriented helps from this forum. You were one of major contributors to the forum, but you can not replace all of the other contributors.

I think you and I have a little different point of view about the forum. I do not want to either push you to accept my ideas or pursuade me to accept your opinions.

Actually I enjoyed some of your previous postings though. I really agree with the item (3) in the posting. I hope all of college professors have that kind of opinion. They usually want 'RESULTS !, RESULTS !' May I say that they attempt to ignore so called bird's eye view on CFD ? I guess you were in the same situation(when you were a student) like most of us in the forum who push themselve to get simulation results as early as possible so that there would be no trouble to get a degree.

I am very sorry to say that I read only first and last 5~6 lines of your posting. It is too long to read. I was also scared about that you might say "Go to church and listen to what old guys say" in the posting like "Find a French girl" in one of your previous postings.

John C. Chien January 18, 2000 03:15

Re: Noise of John C. Chien
(1). As I have said before, I don't keep track of people on Internet. I only try to use the question to present the answers in certain ways which I think would benefit more readers. (2). I had my BS, MS,and PhD from different schools, and in between, I worked to support myself and my family. It was not easy, and I used to work through mid-night at the school's computer center, between the window and the key-punch room. (3). I have received e-mails at my free Yahoo address from many friends at the forum. You seem to have no problem in communication, so, you are free to send me e-mail if you have the need to do so. In terms of the actual CFD experience, I am one of the most experienced engineer in the whole world. (4). I have just reviewed my three rolls of Provia/f-100 slide films, taken at the rose garden over the weekend. The saturated colors are just beautiful. (5). I must say that the messages I presented were structured artificially for the purpose stated in (1), there is no direct relationship between the answer and the question, or the person who presented the question. ( I think, this is not a place to review the CFD for the final exam.) (6). So, Please use the e-mail, for me, it is free from Yahoo. You can also post the message here to me, as long as I am not very busy, I will be very happy to help you. Take it easy, I don't go to church. I will be spending more time checking out this new Provia/f-100 slide film on weekends.

Dr. Hrvoje Jasak January 18, 2000 07:00

Re: Noise

I have recently seen a presentation (from a German University) about noise predction with RANS. It says that you have to split the problem into two parts: noise creation and noise propagation). The second bit is straightforward enough, but a bit of a nightmare to solve (the equation has a third derivative in time!).

As for the first bit, when you use LES you get the noise source right away, as the velocity oscillation in LES (it can be augumented to sub-scale frequencies, see below). From the u' you can directly calculate the noise source terms (don't know what the equation is called, but they said it was standard stuff). However, RANS calculation does not give you this kind of data. What these people do is to approximately "reconstruct" the oscillating velocity field from k and epsilon, with Driest-Kennedy spectra and other bits from isotropic homogenous turbulence theory. Once You've got the u' reconstructed, the calculation becomes TRANSIENT: parameters of u' are varied and the equation is intergrated. So, you basically end up doing a transient calculation (after RANS), which is not much cheaper than doing LES directly. Also the guy said that he would prefer using LES over reconstructed RANS if he can afford it.

The same procedure of reconstructing u' from the (subgrid) spectrum can be used to extend the range of frequencies for LES calculations.

I'm afraid I haven't got any references to this work at the moment, but if you need some I might be able to dig out some names.


Victor Serov January 18, 2000 10:19

Re: Noise

It would be quite helpful if you could look for some names and universities.



Dr. Hrvoje Jasak January 19, 2000 06:22

Re: Noise

The name is: dr. K. Debatin, University of Karlsruhe, Institute for fluid dynamics

(I didn't check the web site)


All times are GMT -4. The time now is 19:29.