aboutsummaryrefslogtreecommitdiffstats
path: root/fg21sim/extragalactic
diff options
context:
space:
mode:
authorAaron LI <aly@aaronly.me>2017-11-16 14:14:18 +0800
committerAaron LI <aly@aaronly.me>2017-11-16 14:14:18 +0800
commit229139f4c3d326828d6ce4d4b2497c67643e96f8 (patch)
tree5d6e91bcbc5ac5461ad56c12c90aaa0ee083b86d /fg21sim/extragalactic
parent91b0e2df43d6f64f7e0888c1cbc276d9c6b2922c (diff)
downloadfg21sim-229139f4c3d326828d6ce4d4b2497c67643e96f8.tar.bz2
clusters/halo: Tune the maximum acceleration timescale to 10 Gyr
Diffstat (limited to 'fg21sim/extragalactic')
-rw-r--r--fg21sim/extragalactic/clusters/halo.py23
1 files changed, 17 insertions, 6 deletions
diff --git a/fg21sim/extragalactic/clusters/halo.py b/fg21sim/extragalactic/clusters/halo.py
index f69c1cc..0b8446a 100644
--- a/fg21sim/extragalactic/clusters/halo.py
+++ b/fg21sim/extragalactic/clusters/halo.py
@@ -711,9 +711,21 @@ class RadioHalo:
the electrons during the merging period, i.e., the acceleration
timescale is set to be infinite after "t_merger + time_cross".
- However, a zero diffusion coefficient may lead to unstable/wrong
- results, so constrain the acceleration timescale to be a large
- enough but finite number (e.g., 10 Gyr).
+ WARNING
+ -------
+ A zero diffusion coefficient may lead to unstable/wrong results,
+ since it is not properly taken care of by the solver. Therefore
+ give the acceleration timescale a large enough but finite value
+ after turbulent acceleration.
+ Also note that a very large acceleration timescale (e.g., 1000 or
+ even 10000) will also cause problems (maybe due to some limitations
+ within the current calculation scheme), for example, the energy
+ losses don't seem to have effect in such cases, so the derived
+ initial electron spectrum is almost the same as the raw input one,
+ and the emissivity at medium/high frequencies even decreases when
+ the turbulence acceleration begins!
+ By carrying out some tests, the value of 10 [Gyr] is adopted for
+ the maximum acceleration timescale.
Parameters
----------
@@ -734,9 +746,8 @@ class RadioHalo:
Ref.[donnert2013],Eq.(15)
"""
if (t < self.age_merger) or (t > self.age_merger+self.time_crossing):
- # NO acceleration; use a large enough timescale to avoid
- # unstable results.
- tau_acc = 100 # [Gyr]
+ # NO acceleration (see also the above NOTE and WARNING!)
+ tau_acc = 10 # [Gyr]
else:
# Turbulence acceleration
tau_acc = self.tau_acceleration # [Gyr]