Processing math: 100%

Labels

Thursday, October 11, 2018

The solution of cosmic ray propagation equation



transport equation

(DNe)+E(dEdtNe)+Qe=Net


energy change of particle
E(t+dt)=E(t)+dEdtdt

dt=dEdE/dt

t2t1dt=E2E1dEdE/dt


if dE/dtb(E)=aE2  and t2>t1
t2t1=1a[1E21E1]

E(t1)=E(t2)1a(t2t1)E(t2)

and
E(t2)=E(t1)1+a(t2t1)E(t1)


green function
Ge(r,r,E,E,t,t)=1[π4f(E,E)]3/2b(E)exp[(rr)24f(E,E)]δ(tt+h(E,E))


f(E,E)=EED(u)dudu/dt and h(E,E)=EEdudu/dt

δ function
δ(g(x))=iδ(xxi)|g(xi)|

t>t    the time axis  ------t'---------t----->
δ(tt+h)
h=EEdEdE/dt=1a[1E1E]

g(E)=tt+h=tt+1a[1E1E]
g(E)=tt+1a[1E1E]=0
g(E)=1aE2

δ(tt+h)=δ(tt+[1a(1E1E)])=δ(EE1a(tt)E)[a(E1a(tt)E)2]


The solution of transport equation is given by
Ne(r,E,t)=drdEdtGe(r,r,E,E,t,t)Qe(r,E,t)=dtdrb(E)b(E)Qe(r,E,t)(4πf(E,E))3/2exp[(rr)24f(E,E)]

In this equation, the E is given by E=E1a(tt)E.


In fact, the solution N(r,E,t)=dtdrb(E)b(E)Q(r,E,t)(4πf(E,E))3/2exp[(rr)24f(E,E)]

is general. It doesn't  depend on the assumption of the form of energy loss.
if we expand δ(tt+EEdEdE/dt) directly, we can get the general solution. In this case E is a function of E and tt.




The integration of spatial part
To simplify our mark, we let the r=0. It means we take the position of observer as original point.
if the spatial part of source term Q is q(r), the integration for spatial part is
SI=drq(r)exp(r24f(E,E))=q(r)exp(r24f(E,E))r2drdΩ


Assume the source locates at rs and the position with respect to source is r0 .
For the geometry, one can refer the figure 1 in the paper https://arxiv.org/pdf/1009.0894.pdf.

Then r=rs+r0 and r2=r2s+r202rsr0sin(θ)cos(ϕ).

Note that the rs is fixed,  f(r)=f(r0) and dr=dr0.

If the particles are homogeneous injected by the sphere surface with radius R, the source term becomes q(r)=14πR2δ(r0R).
I mark 4f(E,E) as L.
SI4πR2=exp([r2s+r202rsr0sin(θ)cos(ϕ)]/L)r20δ(r0R)dr0sin(θ)dθdϕ=R2exp([r2s+R2]/L)exp(2rsRsin(θ)cos(ϕ)/L)sin(θ)dθdϕ

let 2rsR/L=c.

how to compute the integration 2π0π0exp(csin(θ)cos(ϕ))sin(θ)dθdϕ  ?

The author of paper On the point-source approximation of nearby cosmic ray sources says the integration is 4π/csinh(c). I checked this result by numerical integration with python.


#for example

import scipy.integrate as integrate
from math import *
def f(x,y):
    return exp(-10*sin(x)*cos(y))*sin(x)
integrate.nquad(f,[[0,pi],[0,2*pi]]) 
#output (13839.636596576718, 4.0309916130354395e-05)
4*pi/10*sinh(10)
#output 13839.63659657671

2π0π0exp(csin(θ)cos(ϕ)=4πcsinh(c)

SI=2rsRexp(r2s+R24f(E,E))sinh(rsR2f(E,E)

N(r,E,t)=dtb(E)b(E)Q(t,E)14π3/2f1/2rsRexp(r2s+R24f(E,E))sinh(rsR2f(E,E))

ttdt=EEdEdE/dt


REFERENCE
solution
On the point-source approximation of nearby cosmic ray sources 
numerical integration
https://docs.scipy.org/doc/scipy/reference/tutorial/integrate.html
Dirac function
https://en.wikipedia.org/wiki/Dirac_delta_function

Others
On the contribution of nearby sources to the observed cosmic-raynuclei


No comments:

Post a Comment