Hi all, I have a really basic dumb question: what does yt do when it creates a profile? For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones") Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin? The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data? Thank you! Yuan
Hi Yuan, This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative. On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Matt, Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles). What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way. Yuan On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Yuan, You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting. Matt On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Matt, Ok. I was being lazy :P. Here is a 2d histogram (without labels but the range is the same as the yt plot). It still shows more points at density of 1e-23 and temperature of 1e5. These should be particles in and around cooling/cold clumps. It is 10 thousand years after the particles are injected (the first particle output). Maybe somehow the particles move very very quickly from the cell center to strange locations...? Yuan On Mon, Oct 5, 2015 at 2:36 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting.
Matt
On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Sorry! I lied. The color range is different. Though if I make the range similar to the yt plot, then I do not see any particles at high densities. I should probably figure out how to make the volume weighted plot for particles. Yuan On Mon, Oct 5, 2015 at 2:57 PM, Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Ok. I was being lazy :P. Here is a 2d histogram (without labels but the range is the same as the yt plot). It still shows more points at density of 1e-23 and temperature of 1e5. These should be particles in and around cooling/cold clumps.
It is 10 thousand years after the particles are injected (the first particle output). Maybe somehow the particles move very very quickly from the cell center to strange locations...?
Yuan
On Mon, Oct 5, 2015 at 2:36 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting.
Matt
On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Yuan — How are you getting the density/temperature values for the tracer particle plot? If it is from the Enzo tracer particle values, note that they are CIC interpolated from the grid, and so may differ from the cell values (in particular, the high-T, high-rho values you are seeing might be from interpolation between hot and cold cells). Greg
On Oct 5, 2015, at 3:23 PM, Yuan Li <bear0980@gmail.com <mailto:bear0980@gmail.com>> wrote:
Sorry! I lied. The color range is different. Though if I make the range similar to the yt plot, then I do not see any particles at high densities.
I should probably figure out how to make the volume weighted plot for particles.
Yuan
On Mon, Oct 5, 2015 at 2:57 PM, Yuan Li <bear0980@gmail.com <mailto:bear0980@gmail.com>> wrote: Hi Matt,
Ok. I was being lazy :P. Here is a 2d histogram (without labels but the range is the same as the yt plot). It still shows more points at density of 1e-23 and temperature of 1e5. These should be particles in and around cooling/cold clumps.
It is 10 thousand years after the particles are injected (the first particle output). Maybe somehow the particles move very very quickly from the cell center to strange locations...?
Yuan
On Mon, Oct 5, 2015 at 2:36 PM, Matthew Turk <matthewturk@gmail.com <mailto:matthewturk@gmail.com>> wrote: Hi Yuan,
You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting.
Matt
On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com <mailto:bear0980@gmail.com>> wrote: Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com <mailto:matthewturk@gmail.com>> wrote: Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu <mailto:yuan@astro.columbia.edu>> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org <mailto:yt-users@lists.spacepope.org> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org <http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org>
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org <mailto:yt-users@lists.spacepope.org> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org <http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org>
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org <mailto:yt-users@lists.spacepope.org> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org <http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org>
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org <mailto:yt-users@lists.spacepope.org> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org <http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org>
<0_phase_particles_color.png>_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org <mailto:yt-users@lists.spacepope.org> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
---------------------------------- Greg Bryan Professor of Astronomy Department of Astronomy, Columbia University 1325 Pupin Physics Laboratories, Mail Code 5246 550 West 120th Street New York, New York 10027 (USA) email: gbryan@astro.columbia.edu <mailto:gbryan@astro.columbia.edu> phone: +1(212)854-6837 fax : +1(212)854-8121
Hi Greg, Thank you! You were right! Yes, they are from the Enzo tracer particle values. I thought that they were just the values of their host cells, and did not know that they were CIC interpolated. If I want to compare the different heating/cooling rates (as a function of radius for example) calculated from the thermal history of the particles with the rates from the simulation itself, what is the best way of doing it so that I am comparing apples to apples? Yuan On Tue, Oct 6, 2015 at 8:30 AM, Greg Bryan <gbryan@astro.columbia.edu> wrote:
Yuan — How are you getting the density/temperature values for the tracer particle plot? If it is from the Enzo tracer particle values, note that they are CIC interpolated from the grid, and so may differ from the cell values (in particular, the high-T, high-rho values you are seeing might be from interpolation between hot and cold cells).
Greg
On Oct 5, 2015, at 3:23 PM, Yuan Li <bear0980@gmail.com> wrote:
Sorry! I lied. The color range is different. Though if I make the range similar to the yt plot, then I do not see any particles at high densities.
I should probably figure out how to make the volume weighted plot for particles.
Yuan
On Mon, Oct 5, 2015 at 2:57 PM, Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Ok. I was being lazy :P. Here is a 2d histogram (without labels but the range is the same as the yt plot). It still shows more points at density of 1e-23 and temperature of 1e5. These should be particles in and around cooling/cold clumps.
It is 10 thousand years after the particles are injected (the first particle output). Maybe somehow the particles move very very quickly from the cell center to strange locations...?
Yuan
On Mon, Oct 5, 2015 at 2:36 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting.
Matt
On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote:
Hi all,
I have a really basic dumb question: what does yt do when it creates a profile?
For example, if I do this: prof=yt.create_profile(sp,"radius",["temperature"], accumulation=False,n_bins=n_bins,weight_field="ones")
Is prof["temperature"].value the arithmetic average of the temperatures of all the cells within each radius bin?
The reason why I ask is because when I create a profile plot of "CoolingRate" (my derived field) using yt, and compare that with the one I create with tracer particles, they look different at low temperatures. Since I put one particle in the center of each cell, I assume that the two methods should give the same results. Did I make a silly mistake somewhere or does yt do some sort of smoothing to the data?
Thank you! Yuan
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
<0_phase_particles_color.png> _______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
---------------------------------- Greg Bryan Professor of Astronomy Department of Astronomy, Columbia University 1325 Pupin Physics Laboratories, Mail Code 5246 550 West 120th Street New York, New York 10027 (USA)
email: gbryan@astro.columbia.edu phone: +1(212)854-6837 fax : +1(212)854-8121
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi all, Sorry for spamming the yt list again because this really is not a yt problem. After removing the CIC interpolation for the particles in Enzo (reading information from the local cells instead of doing any interpolation), the plot made with the particles agrees with the one from the simulation well. Thank you! Yuan On Tue, Oct 6, 2015 at 1:33 PM, Yuan Li <bear0980@gmail.com> wrote:
Hi Greg,
Thank you! You were right! Yes, they are from the Enzo tracer particle values. I thought that they were just the values of their host cells, and did not know that they were CIC interpolated.
If I want to compare the different heating/cooling rates (as a function of radius for example) calculated from the thermal history of the particles with the rates from the simulation itself, what is the best way of doing it so that I am comparing apples to apples?
Yuan
On Tue, Oct 6, 2015 at 8:30 AM, Greg Bryan <gbryan@astro.columbia.edu> wrote:
Yuan — How are you getting the density/temperature values for the tracer particle plot? If it is from the Enzo tracer particle values, note that they are CIC interpolated from the grid, and so may differ from the cell values (in particular, the high-T, high-rho values you are seeing might be from interpolation between hot and cold cells).
Greg
On Oct 5, 2015, at 3:23 PM, Yuan Li <bear0980@gmail.com> wrote:
Sorry! I lied. The color range is different. Though if I make the range similar to the yt plot, then I do not see any particles at high densities.
I should probably figure out how to make the volume weighted plot for particles.
Yuan
On Mon, Oct 5, 2015 at 2:57 PM, Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Ok. I was being lazy :P. Here is a 2d histogram (without labels but the range is the same as the yt plot). It still shows more points at density of 1e-23 and temperature of 1e5. These should be particles in and around cooling/cold clumps.
It is 10 thousand years after the particles are injected (the first particle output). Maybe somehow the particles move very very quickly from the cell center to strange locations...?
Yuan
On Mon, Oct 5, 2015 at 2:36 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
You're not applying any averaging when scattering, so you can really just get the shape, not the colors, from the particle plotting.
Matt
On Mon, Oct 5, 2015, 11:23 AM Yuan Li <bear0980@gmail.com> wrote:
Hi Matt,
Arithmetic average is what I want. It would be more meaningful to do a volume weighted average, but am trying to compare the profile I get from AMR simulation itself and the one I get from tracer particles, and tracer particles do not know their cell volume (I guess I can use find_field_value_at_point but I am worried about it taking too long for 1 million particles).
What confuses me is that the particles give me a different profile. If I make the phase plot of temperature vs density, it seems that there are a lot of particles at high density and low-ish temperatures (the plot with the blue points) that do not show up on the other plot that is made with the simulation output. The particles are injected to the center of each cell (the original scrip is from you I think), so I thought that every cell should be represented on both plots the same way.
Yuan
On Mon, Oct 5, 2015 at 12:51 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi Yuan,
This does do the average temperature as a function of radius, although note that you're weighting by ones, so it is the arithmetic average -- i.e., unweighted by any volume or mass. So for any AMR simulation, this favors the higher resolution data in a way that is non-conservative.
On Fri, Oct 2, 2015 at 3:03 PM, Yuan Li <yuan@astro.columbia.edu> wrote: > Hi all, > > I have a really basic dumb question: what does yt do when it creates a > profile? > > For example, if I do this: > prof=yt.create_profile(sp,"radius",["temperature"], > accumulation=False,n_bins=n_bins,weight_field="ones") > > Is prof["temperature"].value the arithmetic average of the temperatures of > all the cells within each radius bin? > > The reason why I ask is because when I create a profile plot of > "CoolingRate" (my derived field) using yt, and compare that with the one I > create with tracer particles, they look different at low temperatures. Since > I put one particle in the center of each cell, I assume that the two methods > should give the same results. Did I make a silly mistake somewhere or does > yt do some sort of smoothing to the data? > > > Thank you! > Yuan > > > > _______________________________________________ > yt-users mailing list > yt-users@lists.spacepope.org > http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org > _______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
<0_phase_particles_color.png> _______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
---------------------------------- Greg Bryan Professor of Astronomy Department of Astronomy, Columbia University 1325 Pupin Physics Laboratories, Mail Code 5246 550 West 120th Street New York, New York 10027 (USA)
email: gbryan@astro.columbia.edu phone: +1(212)854-6837 fax : +1(212)854-8121
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
participants (4)
-
Greg Bryan -
Matthew Turk -
Yuan Li -
Yuan Li