From a8caad1ade24f21b9de039150ec4c75419d8eedc Mon Sep 17 00:00:00 2001 From: Miikka Vaisala Date: Thu, 18 Jul 2019 17:34:09 +0800 Subject: [PATCH] A draft of the sink particle plan. --- doc/manual/convert.sh | 1 + doc/manual/sink_particle.md | 44 +++++++++++++++++++++++++++++++++++-- 2 files changed, 43 insertions(+), 2 deletions(-) diff --git a/doc/manual/convert.sh b/doc/manual/convert.sh index ad47094..39d997c 100755 --- a/doc/manual/convert.sh +++ b/doc/manual/convert.sh @@ -1 +1,2 @@ pandoc sink_particle.md -o sink_particle.html --mathml +pandoc manual.md -o manual.html --mathml diff --git a/doc/manual/sink_particle.md b/doc/manual/sink_particle.md index 4e8eabf..eb86610 100644 --- a/doc/manual/sink_particle.md +++ b/doc/manual/sink_particle.md @@ -27,6 +27,10 @@ Based on `sink` gravitational force effect is added to the momentun equation (vi ## Stage 4: Accretion +This part of the process is most complicated, and will require special +attention. The starting point should be the Truelove criterion, but we migh +need additional subgrid model assumptions. + Jeans length \begin{equation} \lambda_\mathrm{J} = \Big( \frac{\pi c_s^2}{G \rho} \Big)^{1/2} @@ -36,8 +40,44 @@ Truelove-Jeans density. \begin{equation} \rho_\mathrm{TJ} = \frac{\pi J^2 c_s^2}{G \Delta x^2} \end{equation} -where $J = \Delta x / \lambda_\mathrm{J}$ +where $J = \Delta x / \lambda_\mathrm{J}$. Lee et al. (2014) set $J=1/8$. + +Magnetic Truelove criterion +\begin{equation} +\rho_\mathrm{TJ, mag} = \rho_\mathrm{TJ} (1 + 0.74/\beta) +\end{equation} + +Accreted mass +\begin{equation} + (\rho - \rho_\mathrm{TJ, mag})\Delta x^3 +\end{equation} ## Stage 5: Data gathering (device to host) -We need to combine accretion calculated in multiple GPUs. +After accretion we will need to gather all data to `sink.mass`. This might need gathering from multiple GPUs, i.e. like in + +Depends on the stencils size and shape used for at the accretion stage. Currently the method is not clear. + +## Plan + +### 1. Add gravitating particle + +Add a particle with specific mass and location. No accretion included. + +### 2. Add simple accretion + +Add an accretion property of the particle. Use just a basic form. + +### 3. Add Complicated accretion + +Add accretion properties which might be useful for the sake of physical correctness and/or numerical stability. + +### 4. Add particle movement. (OPTIONAL) + +Make is possible for the particle to accrete momentum in addition to mass, and therefore influence its movement. + +### 5. Multiple particles. (VERY OPTIONAL) + +Create sink particles dynamically and allow the presence of multiple sinks. + +