# Conservative versus Non-conservative forms

 Register Blogs Members List Search Today's Posts Mark Forums Read

 April 30, 2010, 13:47 Conservative versus Non-conservative forms #1 New Member   Miguel Caro Join Date: Apr 2010 Posts: 26 Rep Power: 8 Hi, i readed in the Leveque's book and the Toro's book about the difference between the conervative and non conservative formulation. But it is not clear how they explained the difference between these two forms. Does anybody has a clear explanation of this? Thanks.

 April 30, 2010, 16:13 #2 Senior Member   Join Date: Feb 2010 Posts: 145 Rep Power: 9 I have been trying to understand this myself. Both forms are conservation laws but the basic idea is that the mathematical character of the equation may be suited better towards different physics or have other advantages/disadvantages. My comments won't be too helpful but here's some random information that I've gathered. See discussion at http://www.cfd-online.com/Forums/main/10994-conservative-non-conservative-form.html There may also be information at http://books.google.com/books?id=J5y...page&q&f=false http://ocw.mit.edu/NR/rdonlyres/Eart...A870/0/ch1.pdf For energy conservation, it is better to solve the Total Energy Equation than the Thermal Energy Equation if there are shocks. A seemingly common misconception is that only one of these forms of the equations includes friction. Both are fundamental conservation laws which are related to each other and which include friction. The thermal energy equation, for example, is the total energy equation minus the dot product of the velocity vector and conservation of momentum. I was also told by an ANSYS engineer that - Method is inherently conservative - Guarantees conservation of fluxes through the control volume - Overall solution will be conservative in nature but may not be the actual solution - Formulation sensitive to distorted elements which can prevent convergence if such elements are in critical flow regions Unfortunately, I do not know which form of the equation to which these apply. Maybe this will give a starting point for issues to look out for. I'm sorry that I do not have any clear information. Good luck.

 April 30, 2010, 19:51 #3 Senior Member   Join Date: Jul 2009 Posts: 230 Rep Power: 11 The conservative form derives from expressing the conservation equations in a divergence form, i.e. the equations have the form of local time rate of change + divergence of flux = source terms. The non-conservative form is a form of the equations that is not expressible in this form. One of the principle advantages to the the conservative form is that once the equations are discretized, the flux terms "telescope", that is if you sum the fluxes into and out of a row of cells, the intercell fluxes cancel and the net flux is just the flux out of one end of the row - flux in the other end. rachitsigh11 and sarodesr like this.

 May 1, 2010, 08:32 #4 Member   Join Date: Mar 2009 Posts: 33 Rep Power: 9 Conservative form has an advantage: its integral form allows discontinuous solutions (leading to Rankine-Hugoniot relation and correct shock speed). Non-conservative form does not have such a character (it allows only smooth differentiable solutions). Discrete conservation is important in computing shocks. If conservation is violated, a shock may travel at a wrong speed. Successful non-conservative schemes may turn out to satisfy some form of discrete conservation. Rather than conservation itself, it is more important to use a right conservation form (need to pick right quantities that should be conserved) since a wrong conservation form can lead to a wrong shock speed.

May 1, 2010, 17:32
Thank you all, it is now a little more clear
#5
New Member

Miguel Caro
Join Date: Apr 2010
Posts: 26
Rep Power: 8
Quote:
 Originally Posted by gory Conservative form has an advantage: its integral form allows discontinuous solutions (leading to Rankine-Hugoniot relation and correct shock speed). Non-conservative form does not have such a character (it allows only smooth differentiable solutions). Discrete conservation is important in computing shocks. If conservation is violated, a shock may travel at a wrong speed. Successful non-conservative schemes may turn out to satisfy some form of discrete conservation. Rather than conservation itself, it is more important to use a right conservation form (need to pick right quantities that should be conserved) since a wrong conservation form can lead to a wrong shock speed.
Thank you all!!

 May 3, 2010, 09:10 #6 Senior Member   Join Date: Feb 2010 Posts: 145 Rep Power: 9 Does CFX allow you to choose which form of the equations are solved? I see that CFX-Pre allows one to choose "Thermal Energy" or "Total Energy." Which one is the conservative form? I do not see where there is a choice for the momentum equations. Thanks so much for any information!

 May 5, 2010, 09:12 #7 New Member   Nicholas F Camus Join Date: Sep 2009 Location: London, England. Posts: 21 Rep Power: 9 Jade the manual explains the difference between the "Thermal Energy" or "Total Energy" options. If I remember correctly the Total Energy option is just applied as a power [kg m^2 s^-3] to be applied to a domain, subdomain or boundary - the solver then calculates the heat flux accross each cell according to volume/surface ratios of the cells to the domain selected. Using the Thermal Energy option, you are required to supply the power per unit volume. Both of these options supply the solver with an enery source term (a addtional term in the energy equation), thus the solver is working with the conservative variables. The CFX solver always works with conservatives, although transformations to primatives are bound to occur during the soultion process in order to calculate various quantities such as fluxes etc. There is options to apply momentum source terms; again this is clearly described in the manuals. All the best.

 May 5, 2010, 09:34 #8 New Member   Nicholas F Camus Join Date: Sep 2009 Location: London, England. Posts: 21 Rep Power: 9 I have also just read your confused posts on variable conservation. OK. The books by R. LeVeque and F. Toro are primaraly (almost completely) concerned with hyperbolic systems. These are systems which can be cast into purely conservative form (i.e. in the form of a system of conservation laws (mass, momentum and energy)). There are man many system that can be written in this form: ideal gas dynamics, relativistic hydrodynamics, magnetohydrodynamics, and relativistic hydrodynamics (including multi-fluid systems) to name the major ones. The physics of such hyperbolic systems is different to those parabolic systems of viscous unsteady flow that codes such as CFX are designed to deal with. The wave structure of the hyperbolic flows is crucial, as shock formation, and strong rarefraction wave can seriously effect flow dynamics (typically only very fast flows). CFX and the like are not concerned with such wave probagation as at slow speed the wave strengths are very weak indeed - hence their used of a very feable (but robust) upwind scheme. Where the variable conversion comes into such hyperbolic code is in the calculation of the intercell fluxes, the calculation of which often needs sound speed and primative variables - density, pressure ect. The solution for these intercell fluxes for the hyperbolic systems is not streight forward, and it is not possible to use the physical fluxes F = Variable*velocity. This is what the Riemann problem is based around, and this is well documented in the above books. The type of variable used has no bairing on the shock capturing ability of the scheme in question. The ability to resolve strong shocks in such system is wholly dependent on the Riemann solver used. For Riemann solvers there are two main types Exact solutions (only for hydrodynamics and relativistic HD), and approxiamte solutions (the cleaver stuff). The second of these can be split into linear (Roe-type (not quite right)) Riemann solvers using eigenvalue formulations, which can use primative varibles as well as conservatives, and basic approxiamte solvers which use conservatives and are based around conservative formulations (see HLL Riemann solver for a popular example). Please feel free to ask any more questions. I think this should be enough for you to read up on for now! All the best, and good luck.

 May 6, 2010, 08:36 #9 Senior Member   Join Date: Feb 2010 Posts: 145 Rep Power: 9 Thanks Killercam. Actually, I am simply asking which equations, "Thermal Energy" or "Total Energy" is considered the conservative form. DO you know? Also I do not see where there is a choice for the momentum equations in the software. Thanks for any information.

 May 6, 2010, 09:01 #10 New Member   Nicholas F Camus Join Date: Sep 2009 Location: London, England. Posts: 21 Rep Power: 9 In CFX these are just terms used for specifing energy sources. If you are asking about equations concerned with purely thermal energy, and total energy... They are both energy equations and energy is a conservative quantity. As for their form, both when dervived correctly should be fully conservative, for any finite volume method. In CFX - if I can remember, here goes. For any boundary (Inlet, outlet etc.) you can specify boundary source terms for continuity, energy and turbulence. You can also specify momentum sources by introducing a sub-domain into you fluid domain. Sources - Boundary Source - Sources - and I think then use sources for continuity. Good luck.

 May 6, 2010, 09:33 #11 Senior Member   Join Date: Feb 2010 Posts: 145 Rep Power: 9 In that case, all the conservation equations are conservative? What is meant by the conservative and non-conservative forms then?

 July 20, 2010, 11:13 #12 New Member   aNaN Join Date: Jul 2010 Posts: 1 Rep Power: 0 hey anyone please explain about the non conservative form also ..

 Tags conservative form

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post Saturn CFX 45 February 8, 2016 05:42 Pankaj CFX 9 November 23, 2009 05:05 vijesh joshi Main CFD Forum 4 March 16, 2006 00:10 Cesare CFX 1 September 27, 2004 11:00 lego CFX 3 November 5, 2002 21:09

All times are GMT -4. The time now is 08:53.