Dear all,
I’m wondering if conducting an MDPDE simulation in LAMMPS is possible.
LAMMPS supports EDPD and MDPD in its MESO-DPD package. Looking at the literature, it seems that MDPDE is just EDPD with the density-dependent force term of MDPDE, whose equations are almost identical to EDPD’s. So, I want to know if it’s possible to use the MDPD pair style with the temperature enforcement of the EDPD style to obtain MDPDE.
I’m posting this question because I’m not very familiar with the DPD method. If anyone who knows it better knows of a special detail I’m missing, I would appreciate it if you could tell me or confirm that the methods can be combined.
Best regards.
Well, then you obviously need to either familiarize yourself or find a collaborator that is sufficiently familiar.
I don’t think that what you want to do is as easy as you suggest it could be. The problem is already at the technical level since you have multiple atom styles, multiple pair styles and multiple fix styles. Each of those have different (strict) requirements and I doubt that there is a way to satisfy all of them. While for the atom styles, you may be able to use a hybrid style, that route is closed for pair and fix styles. As a minimum, I would suspect you would need to implement a new pair style that combines the features of multiple existing pair styles.
But don’t take my word for it. I am only guessing here. Study the source and the corresponding literature and come to an informed conclusion.
Hi @akohlmey, thank you for your answer.
I’ve been reading about the DPD method, and the answer to my question is a bit of a yes/kind of.
The main differences between EDPD and MDPD in LAMMPS are that EDPD has a temperature-dependent exponent (s), mesoscopic heat friction coefficient (st), and the DPD thermostat, while MDPD assumes those values as constant, does not include the thermostat, and has an extra force term that depends on density.
So, if one is in a regime where s and st can be constant, then MDPD can be used as MDPDE using LAMMPS NVT or NPT thermostats. However, the MDPDE model has some details that the LAMMPS implementation of MDPD can’t capture. The aforementioned constant values and the repulsive force coefficient can be temperature-dependent. For example, this is important for modeling the correct temperature-dependent surface tension and viscosity.
I hope this thread helps anyone confused by the similarity of both methods.
The DPD-MESO package probably needs to be expanded to include proper MDPDE. Depending on the results of my research, I will probably do it myself.
Best regards.