<html><head><style type='text/css'>p { margin: 0; }</style></head><body><div style='font-family: arial,helvetica,sans-serif; font-size: 12pt; color: #000000'><BR>
<DIV style="FONT-STYLE: normal; FONT-FAMILY: Helvetica,Arial,sans-serif; COLOR: #000; FONT-SIZE: 12pt; FONT-WEIGHT: normal; TEXT-DECORATION: none">ACCELERATOR SEMINAR<BR><BR>Using the Grid-Free Fast Multipole Method in Collective Effect Simulations <BR><BR>He Zhang, <BR>Center for Advanced Studies of Accelerators, Jefferson Lab<BR><BR>Currently the mos popular algorithm for the simulations on collective effects in accelerators is the Particle-In_Cell (PIC) method. The PIC method is based on a grid, which makes it difficult to deal with highly non-uniform charge distribution and/or complex geometry. Also the information of the fine structure of the charge distribution is lost when distributing the charges on the grid points for field calculation. In this presentation, I will introduce the grid-free fast multipole method (FMM), which does not have the above problems, and its application in collective effect simulations. The FMM has an efficiency of O(N) for calculation of the Coulomb forces between N charged particles with any arbitrary distribution. Using FMM, the whole charged domain is decomposed into small boxes of various sizes according to the charge density, so that each box contains no more than a pre-specified maxium number of particles. Taking advantage of the fact that the long-rang Coulomb force decreases fast as the distance between the charges increases, the force contributed from a box to the region far away was represented by a multipole expansion, while the forces to the region nerby is calculated directly with Coulomb formula. the force on each particle is calculated as the summation of contributions from its near region and its ar region. The adaptive box structure makes the FMM suitable for any charge distribution and geometry, and it can simulate the fine structure in a high correlated system since the near region interaction is calculated accurately. Specifically the differential algrebra (DA) based FMM will be discussed and simulations on space charge dominated photoemissions process, in which the DA based FMM is implemented, will be presented. <BR><BR>Wednesday, August 13, 2014<BR>11:00 a.m. <BR>CEBAF Center Auditorium<BR><BR><BR><SPAN id=DWT1557 class=ZmSearchResult>Seminar</SPAN> archives are found here: <SPAN id=OBJ_PREFIX_DWT1558_com_zimbra_url class=Object><A href="http://wwwold.jlab.org/accel/seminars.html" target=_blank>http://wwwold.jlab.org/accel/seminars.html</A></SPAN><BR><BR>
<DIV id=c527d5bf-234f-450e-be05-77753629e03f><SPAN></SPAN>------<BR><B><FONT size=2>Audrey N. Barron</FONT></B><BR><FONT size=2>Administrative Assistant</FONT><BR><FONT size=2>Center for Advanced Studies of Accelerators</FONT><BR><FONT size=2>Jefferson Lab</FONT><BR><FONT size=2>757-269-7327</FONT><BR><FONT size=2><A>anichols@jlab.org</A></FONT> / <FONT size=2><A href="http://www.jlab.org/" target=_blank>www.jlab.org</A></FONT> <BR><BR><IMG style="BORDER-BOTTOM: 0px; BORDER-LEFT: 0px; BORDER-TOP: 0px; BORDER-RIGHT: 0px" src="cid:0608573aa3c13451c3e2e4835dbcea41b46fb153@zimbra"><BR><SPAN></SPAN><BR></DIV><BR>_______________________________________________<BR>Accelerator_staff mailing list<BR>Accelerator_staff@jlab.org<BR>https://mailman.jlab.org/mailman/listinfo/accelerator_staff</DIV><BR></div></body></html>