TY - JOUR
T1 - Dissipative particle dynamics simulation of flow generated by two rotating concentric cylinders
T2 - II. Lateral dissipative and random forces
AU - Filipovic, N.
AU - Haber, S.
AU - Kojic, M.
AU - Tsuda, A.
N1 - Copyright:
Copyright 2008 Elsevier B.V., All rights reserved.
PY - 2008/2/7
Y1 - 2008/2/7
N2 - Traditional DPD methods address dissipative and random forces exerted along the line connecting neighbouring particles. Espanol (1998 Phys. Rev. E 57 2930-48) suggested adding dissipative and random force components in a direction perpendicular to this line. This paper focuses on the advantages and disadvantages of such an addition as compared with the traditional DPD method. Our benchmark system comprises fluid initially at rest occupying the space between two concentric cylinders rotating with various angular velocities. The effect of the lateral force components on the time evolution of the simulated velocity profile was also compared with that of the known analytical solution. The results show that (i) the solution accuracy at steady state has improved and the error has been reduced by at least 30% (in one case by 75%), (ii) the DPD time to reach steady state has been halved, (iii) the CPU time has increased by only 30%, and (iv) no significant differences exist in density and temperature distributions.
AB - Traditional DPD methods address dissipative and random forces exerted along the line connecting neighbouring particles. Espanol (1998 Phys. Rev. E 57 2930-48) suggested adding dissipative and random force components in a direction perpendicular to this line. This paper focuses on the advantages and disadvantages of such an addition as compared with the traditional DPD method. Our benchmark system comprises fluid initially at rest occupying the space between two concentric cylinders rotating with various angular velocities. The effect of the lateral force components on the time evolution of the simulated velocity profile was also compared with that of the known analytical solution. The results show that (i) the solution accuracy at steady state has improved and the error has been reduced by at least 30% (in one case by 75%), (ii) the DPD time to reach steady state has been halved, (iii) the CPU time has increased by only 30%, and (iv) no significant differences exist in density and temperature distributions.
UR - http://www.scopus.com/inward/record.url?scp=43049102759&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=43049102759&partnerID=8YFLogxK
U2 - 10.1088/0022-3727/41/3/035504
DO - 10.1088/0022-3727/41/3/035504
M3 - Article
AN - SCOPUS:43049102759
SN - 0022-3727
VL - 41
JO - Journal of Physics D: Applied Physics
JF - Journal of Physics D: Applied Physics
IS - 3
M1 - 035504
ER -