The ABINIT Users Mailing List ( CLOSED )

[Anglade Pierre-Matthieu <anglade@gmail.com>]

Hi,

In case you suspect the raise of kinetic energy is due to integration
errors, a  solution would be to use a higher order symplectic
algorithm. At the usual time step used in MD, an algorithm like the
Blanes&Moan srkna14  [1] usually outperform verlet in total energy
conservation by 4 to 6 order of magnitude. Moreover it is pretty much
straight forward to implement...

regards,

PMA


[1] Journal of Computational and Applied Mathematics
Volume 142 ,  Issue 2  (May 2002)


On Fri, Jul 17, 2009 at 11:04 AM, marco mancini<man74cio@gmail.com> wrote:
> Hi Manuel,
> to perform plasma simulations I used ionmov=12 which conserve kinetic
> energy.
> bye
> Marco
>
>
>
> Marco Mancini
> ______________________________
> CEA-DAM Ile de France
> Bruyerès-le-Chatel
> 91297 Arpajon CEDEX
> ______________________________
> Tel:  +33(0)6.62.03.47.93
>        +33(0)1.69.26.45.56
> ______________________________
> Email:  man74cio@gmail.com
> ______________________________
> http://man74cio.kelio.org
>
>
> 2009/7/17 <mcotelo@gmail.com>
>>
>> Hi abiniters:
>>
>> I'm trying to use ABINIT to perfom some simulations for dense plasmas. I
>> take
>> the idea from these articles:
>>
>> Very-high-temperature molecular dynamics
>> F. Lambert, J. Clerouin, G. Zerah
>> Physical Review E 73, 1 (2006)
>>
>> First Principles Molecular Dynamics of Dense Plasmas
>> M.P. Surh and TW Barbee and LH Yang
>> Physical Review Letters 86, 5958--5961 (2001)
>>
>> Ab Initio Determination of Thermal Conductivity of Dense Hydrogen Plasmas
>> V. Recoules and F. Lambert and A. Decoster and B. Canaud and J.
>> Cl{\'e}rouin
>> Physical Review Letters (2009)
>>
>> I'm using options ionmov=6 and ionmov=8 to run MD simulations with time
>> steps
>> of about 2.5 fs but the results are not good. After a few time steps, the
>> kinetic energy suddenly rises orders of magnitude and the energy is not
>> conserved..
>>
>> Can anybody give me some advice? Do I have to use some special options or
>> take
>> care about some parameters?
>>
>> I found few information about MD simulations for plasmas. Could you tell
>> me
>> where I can found more information? Could you recommend me some papers?
>>
>> ********************************************************************
>> Manuel Cotelo Ferreiro
>>
>> Universidad Politécnica de Madrid
>> ETSII - Instituto de Fusión Nuclear
>> calle José Gutiérrez Abascal 2
>> 28006 MADRID
>> ESPAÑA
>>
>> tlf : +34 91 336 31 08
>> fax : +34 91 336 30 02
>> ********************************************************************
>>
>
>



-- 
Pierre-Matthieu Anglade