Hi, I'm searching a CFd software able to compute fluid (water, charged water, gas) dynamic in confined domain.

I have to modelize the flow velocity, turbulence, pressure & head losses and cavitation inside pipes, valves, ...

So I need to draw my geometry and mesh it before computing!

Sure that all the commercial CFD (i.e. Fluent) are able to do it, but the problem is the price: they are too expensive to purchase them only for a specific task (computation in confined area)

I'm searching a meshing & solver tool which cost a reasonable price!

thank you for help

Have a look at EFD-Lab from nika

www.nika.biz

You can use opensource codes such as OPENFOAM, it covers main featurs of FLUENT.

i would go for openFOAM personally, the nika solver is unproven and all the results I have seen from it have been very poor

Do you know lot's of software editors that offer 90 days money back garantee? NIKA does it. Do you know that NIKA has more than 1000 industrial customers using EFD products to solve engineering task related to fluid flow and heat transfer. Do you know that EFD.Lab covers 80% of Fluent functionnality coverage, including compressible, transient, radiation, rotationnal domain, cavitation, etc. Ben, CFD is a science, so please be precise in your argumentation.

Which technology is being used by NIKA?

- mesh (structured, unstructured, cartesian, ...) - numerical schemes - turbulence modelling

The web-site gives the impression that it is a windows only application - how does it handle parallelism?

I'm not arguing about the functionality, more the accuracy, the meshing technology seems somewhat odd (partial cells and all) and it speaks volumes that none of the "big three" or any other comercial code to my knowledge uses it. I also know a lot of people who use floworks (not including myself) and the general complaint seems to be that the ceiling of the solvers capability is reached quickly and that it is somewhat inflexible (although of course I guess it seems to be designed as a pretty picture tool more than a hardcore industrial code such as Fluent, STAR-CD or CFX)

But saying that if you are confident in the accuracy, robustness and flexibility of the solver could you point us to some validation papers please....

Give me your email address and I'll send you the fundamental manual including technology details and more than 20 validation examples.

By the way our meshing technology is not unique (please get in touch with Fluent ...). The secret is in the resolution of partial cells.

The strength of EFD is in its unique level of integration in CAD: boundary conditions and mesh specification are associated to the CAD features thus allowing very short calculation loop and real CAD optimization (who has been able to efficiently parametrized a mesh in a standard CFD package without having to plan all the future changes before creating the first node ??!)

EFD solves Navier-Stokes equation using finite volume method: nothing brand new that can justify doubts in results quality. EFD technology was born in 1985.

Don't worry I believe you and I dont really have time to go trawling through manuals. My comments were just going off people I know who have used it and reports I have seen. I'm sure it is very good etc. Out of interest what is the pacakges limit? It's rare that I run under 1 million cells now serial runs are a thing of the past so how is the parallel scaling on linux clusters and what sort of cell counts can the solver handle?

Although you're point about CAD integration is somewhat miss informed as there are few other parametric packages around now that have it, i.e. fluent for catia, STAR-CAD....

Try Flo++... an excellent, reasonably-priced package, with very helpful technical support.

<www.softflo.com>

I've the impression you're mixing lots of things there: size of mesh means nothing. The more power you get from your hardware the less mesh optimization you run. Why 1 milion cells where 500k can do the job with a smart mesher and some autoadaptativity?

We run EFD products on laptops. Do you know lots of small and medium sized company that have access to linux clusters.

CAD integration means nothing if you don't have a mesher that can handle design changes automatically whitout manual reworking of newly created small gaps and features.

Fluentforcatia and Starxx are limited versions (no transient analysis in FluentforCatia !!) of well known and recognized CFD packages that have been integrated in CAD for marketting purpose with the target to sell more standalone version.

EFD solutions have been developped for profitability driven industrial company running what-if analysis during the whole design cycle (from quotation to manufacturing). It's a full package with no limitation answering the needs of all industry area and that can be used by non specialist (I know it hurts !! ;-)) with the right methodology and the right support (provided by NIKA itself).

All reasonable, i guess it boils down to whether you want a cheap and cheerfull solver that will produce you some pretty pictures or a proper full scale industrial code.

It doesnt hurt me, i am all for greater integration at a design engineer level, I use STAR-Works a lot although mainly as a mesher and pre-processor, its all fully parametric etc

I guess floworks etc is a good feed for the larger industrial codes, it gives people a taste of CFD and when they want to move on up and study with greater accuracy and a wider range of situations they can buy one of the more established codes. In this way its a good thing as it gets more people interested in CFD and using it and give them some experience before moving on to bigger things.

I must preface this and admit that I know nothing about this NIKA software.

I was just reading through this thread and something struck me as odd. Ben, in a previous post the NIKA guy offered to send you copies of the code manual as well as some validation examples. You state that you are too busy to look at them, but say that you believe the code to be no good based on some other work that you have seen that had bad results. You later take the liberty to say that NIKA is a "cheap and cheerful solver" that will only give you "pretty pictures"

I have used several commercial codes, and I have managed to produce some terrible results at times. If I was to have shown these results to people, they would have thought that these "big" commercial codes are terrible. I have also obtained some excellent results from these commercial codes. There is nothing wrong with these "big" commercial codes, they just sometimes didn't include the best models for the flows that I wanted to predict or I had made some small errors in the set-up. Surely if you want to make a judgement on the fidelity of NIKA, you should look at the work of people using it that are experts in it's implementation.

I have written several of my own codes that could only be described as "cheap and cheerful". Some of these codes have given me the best CFD answers that I have ever obtained. That is because I wrote them to suit the specific problems that I wanted to study. I made a post on here a couple of weeks ago trying to explain that CFD is just a TOOL in the arsenal of a Fluid Dynamicist. Continuing this line of thought, one could think of CFX, Fluent etc. as a sledgehammer, but if one just wants to tap in a nail to hang up a picture, one does not need a sledgehammer, a regular hammer will do! A code does not have to be huge and contain all of the bells and whistles to be a good code, it depends to a large extent on the application for which you want to use it.

This is very true of course, you should always choose the right tool for the job. My comments were more based on the experiences of friends/co-workers who have purchased FloWorks believing that the codes capabilities were greater than they were (due to advertising promisses etc) and ending up having to swap codes anyway. I am certainly not doing down your coding efforts and everything has its place, blame the marketing and sales guys I say!

I agree with you there, in my experience, the advertising and sales guys for most companies (not just CFD) don't have a clue about the wares that they are selling.

I think that this is a function of our modern society though. Very few people take the time to understand things these days, it seems there is very little of substance out there. I guess this is the wrong forum for such a discussion though!

Hi Mohamed,

I'm from CD-adapco and we are developping upfront flow and thermal simulation software. In our offering we have a very useful and affordable product that includes geometry modelling and all physics you're specifying in your email.

http://www.cd-adapco.com/products/ST...AR-Design.html

Feel free to contact me if you will.

Jean-Claude Product Manager Cd-adapco

I agree with you, but not all.

If one need to know the friction loss in a straight pipe, it is not necessary to use CFD, 1D will do.

I only used Fluent, CFX, Flow3D, Flotherm... expensive CFD codes. Even though, these codes don't provide same quality answers for different problems. Such as Flow3D uses empirical or input heat transfer coefficient when solving turbulent thermal wall bounded problems, but FLUENT analog thermal boundary to flow boundary. The results are just so different.

There are so many difference between CFD softwares. I don't quite believe cheap CFD software. For example, can this NIKA study the tip leakage problem of a high-speed rotating impeller? If so, how much time I need to spend on generating mesh, and how it deal with interaction between impeller and diffuser. How about two phase problems? Even FLUENT is still hiring people to develop their two phase flow solver (VOF).

How much the code can do really depends on how much effort (money) has been spend on the code.

Maybe I wasn't clear L.e.i. Here is my philosophy ...

If you want a code to do as many different things as possible, then by all means go with one of the big players (CFX, Fluent etc ...)

If you have a specific application in mind, it may be better to go with one of the smaller codes. This is if you can find one whose code is particularly well suited to your problem.

If you want to have full control, and likely get the best solution to a given problem, write your own code (assuming you have a year or two to spare ...)

Hi Mohamed,

If you just want something cheap, it's worth looking at the major code vendors who produce their own "upfront" tools (CD-adapco, Fluent, CFX may have one, I'm not sure). The price of these tools in very close to the other upfront offerings (e.g. NIKA). They've usually got less functionality, but this doesn't sound like it is much of a problem.