X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Received: from ispmxmta06-srv.windstream.net ([166.102.165.167] verified) by logan.com (CommuniGate Pro SMTP 5.2c1) with ESMTP id 2463599 for flyrotary@lancaironline.net; Fri, 09 Nov 2007 12:40:39 -0500 Received-SPF: pass receiver=logan.com; client-ip=166.102.165.167; envelope-from=montyr2157@alltel.net Received: from ispmxaamta05-gx.windstream.net ([151.213.143.21]) by ispmxmta06-srv.windstream.net with ESMTP id <20071109173959.TAUT29331.ispmxmta06-srv.windstream.net@ispmxaamta05-gx.windstream.net> for ; Fri, 9 Nov 2007 11:39:59 -0600 Received: from Thorstwin ([151.213.143.21]) by ispmxaamta05-gx.windstream.net with SMTP id <20071109173958.GYS9878.ispmxaamta05-gx.windstream.net@Thorstwin> for ; Fri, 9 Nov 2007 11:39:58 -0600 Message-ID: <002f01c822f7$8ee4b980$6501a8c0@Thorstwin> From: "M Roberts" To: "Rotary motors in aircraft" Subject: Re: Total,duct, Ambient or Velocity???? Date: Fri, 9 Nov 2007 11:40:02 -0600 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_002C_01C822C5.442905E0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.3138 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198 This is a multi-part message in MIME format. ------=_NextPart_000_002C_01C822C5.442905E0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Well.. I haven't spent $$$ on it but SolidWorks now comes w/ a version = of CosmosFloWorks in the package in 2008.. I've got the software so I may = give it=20 a whirl.. I was waiting to test it before I commented but.. since people = are talking directly about it.. :-) Jarrett Jarrett, I'm not implying that CFD isn't fun to play with, and if you have CFD = available go for it! Just don't expect to learn anything more than gross = trends for design purposes. I have been involved in several internal = flow designs where we used "expert" PhD types for analysis using high = end programs like Fluent. In the end you still wind up in the lab with = modeling clay to make things work right. Even if you know what you are = doing and have good data to plug into the models (heat exchanger data, = boundary layer conditions at the inlet and exit, pressure and = temperature conditions, prop wake effects, etc) you are still limited by = your fabrication capacities. Most of us don't have the ability to build = an optimum heat exchanger. We have to use what is available off the = shelf. So the system will be optimized for readily available exchangers, = not the optimum. Then what level of fabrication accuracy is available = for the duct work. Most of us carve some foam and slather some = fiberglass on it. We don't use a 5 axis router to make a mold from a = highly refined CFD model and vacuum bag some duct work. To get the 5% = extra you might get from accurate CFD analysis is going to take a lot of = testing to get data to "correct" the CFD model and then a lot of high = end fabrication. All for....5-10% gain.....maybe. CFD has it's place. Primarily in very specialized areas of design = engineering like turbo machinery. In these cases building and testing = prototypes is so expensive, you can afford to hire the brain/computing = power to develop the models and test/refine them. Each new real world = test results in tweaking all the fudge factors for the models specific = use....and there are a lot of fudge factors.=20 In our application, you would be better off with a pencil, paper, = calculator, and a copy of Kucheman and Webber, plus a thermo 1 text book = and a pile of old NACA reports. You are going to have to do this anyway = as a gut check for your CFD models. If you don't you will be off down = the rabbit hole with no trail of bread crumbs.......trust me. In the end you will still have to test to refine. Best to avoid analysis = paralysis from the get go. One test can invalidate years of mental = gymnastics. Our understanding of physics is still limited, especially non-steady = state, non linear things like turbulence. Of course the people who sell = the software will never admit this......but it is the truth. Models are = models and reality trumps all. FWIW Monty=20 ------=_NextPart_000_002C_01C822C5.442905E0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

Well.. I haven't spent $$$ on it but = SolidWorks now=20 comes w/ a version of

CosmosFloWorks in the package in = 2008..  I've=20 got the software so I may give it

a whirl.. I was waiting to test it before = I commented=20 but.. since people are talking

directly about it..  :-)

 

Jarrett

 

Jarrett,

I'm not implying = that CFD isn't=20 fun to play with, and if you have CFD available go for it! Just = don't=20 expect to learn anything more than gross trends for = design=20 purposes. I have been involved in several internal flow designs where we = used=20 "expert" PhD types for analysis using high end programs like = Fluent.=20 In the end you still wind up in the lab with modeling clay to make = things work=20 right. Even if you know what you are doing and have good data to = plug into=20 the models (heat exchanger data, boundary layer conditions at the inlet = and=20 exit, pressure and temperature conditions, prop wake = effects, etc) you=20 are still limited by your fabrication capacities. Most of us don't = have the=20 ability to build an optimum heat exchanger. We have to use what is=20 available off the shelf. So the system will be optimized for readily = available=20 exchangers, not the optimum. Then what level of fabrication = accuracy is=20 available for the duct work. Most of us carve some foam and slather some = fiberglass on it. We don't use a 5 axis router to make a mold from a = highly=20 refined CFD model and vacuum bag some duct work. To get = the 5%=20 extra you might get from accurate CFD analysis is going = to take a=20 lot of testing to get data to "correct" the CFD model and then = a lot=20 of high end fabrication. All for....5-10% gain.....maybe.

CFD has it's place. = Primarily in very=20 specialized areas of design engineering like turbo machinery. In = these=20 cases building and testing prototypes is so expensive, you can = afford to=20 hire the brain/computing power to develop the models and test/refine=20 them.  Each new real world test results in tweaking all the = fudge=20 factors for the models specific use....and there are a lot of fudge=20 factors. 

In our application, you = would be better=20 off with a pencil, paper, calculator, and a copy of Kucheman and Webber, = plus a=20 thermo 1 text book and a pile of old NACA reports. You are going to have = to do=20 this anyway as a gut check for your CFD models. If you don't you will be = off=20 down the rabbit hole with no trail of bread crumbs.......trust = me.

In the end you will still = have to test=20 to refine. Best to avoid analysis paralysis from the get go. One test = can=20 invalidate years of mental gymnastics.

Our understanding of = physics is still=20 limited, especially non-steady state, non linear things like turbulence. = Of=20 course the people who sell the software will never admit = this......but it=20 is the truth. Models are models and reality trumps all.

FWIW

Monty 

------=_NextPart_000_002C_01C822C5.442905E0--