Data Assimilation
158 views
Skip to first unread message
Paulo Marques
Jan 4, 2012, 5:48:19 PM1/4/12
to FDS and Smokeview Discussions
Hi,
I want to know if is possible some kind of data assimilation in FDS?
Best Regards,
Paulo Marques
I want to know if is possible some kind of data assimilation in FDS?
Best Regards,
Paulo Marques
dr_jfloyd
Jan 4, 2012, 5:57:36 PM1/4/12
to fds...@googlegroups.com
If you are referring to data assimilation as discussed in:
http://en.wikipedia.org/wiki/Data_assimilation
then the answer is no.
http://en.wikipedia.org/wiki/Data_assimilation
then the answer is no.
Rein
Jan 6, 2012, 7:59:27 AM1/6/12
to FDS and Smokeview Discussions, Wolfram Jahn, wolfram Jahn (QMUL)
Hi Paulo,
We have used FDS and data assimilation techniques in our forecasting
fire dynamics research. I quote from the thesis of Dr Jahn that I
supervised at the University of Edinburgh (2010) [1]:
"This thesis proposes and studies a method to use measurements of the
real event in order to steer and accelerate fire simulations. This
technology aims at providing forecasts of the fire development with a
positive lead time, i.e. the forecast of future events is ready before
those events take place. A simplified fire spread model is
implemented, and sensor data are assimilated into the model in order
to estimate the parameters that characterize the spread model and thus
recover information lost by approximations. The assimilation process
is posed as an inverse problem, which is solved minimizing a non
linear cost function that measures the distance between sensor data
and the forward model. In order to accelerate the optimization
procedure, the ‘tangent linear model’ is implemented, i.e. the forward
model is linearized around the initial guess of the governing
parameters that are to be estimated, thus approximating the cost
function by a quadratic function. The methodology was tested first
with a simple two-zone forward model, and then with a coarse grid
Computational Fluid Dynamics (CFD) fire model as forward model."
As explained in [2], this work and related technologies allow for the
first time to forecast fire dynamics assimilating sensor data and
solving an inverse problem. It would lead to the implementation of
infrastructure protection in smart buildings. This problem had been
deemed previously as "impossible" by experts.
The is a PhD thesis [1] completed last year. Now all chapters have
been published in the litterature. I think you will be most interested
in the two papers using FDS v5 [3, 4]:
*Forecasting Fire Dynamics Using Inverse Computational Fluid Dynamics
and Tangent Linearisation*
in Advances in Engineering Software, 2011. http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
(at the time of writing this email, the paper is accepted but in
press, and the corrected proofs are not online yet. They will be
available in a few days. Send me an email if you want a personal copy.
It is Chapter 5 in [1]).
and
*Forecasting Fire Growth using an Inverse CFD Modelling Approach in a
Real-Scale Fire Test*
in Fire Safety Science, 2011. http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
A recent presentation at the Institute of Physics - Combustion
Modelling for Challenging Application meeting provides an overview of
the work [5]. Other data assimilation work of the same methodology
using a zone model instead was published in [6]. A closely related PhD
thesis using zone model is [7].
We are most interested in feedback from the fire modelling commmunity.
Your comments are welcome.
G.
[1] Wolfram Jahn, "Inverse modelling to forecast enclosure fire
dynamics", PhD Thesis, University of Edinburgh, 2010. http://hdl.handle.net/1842/3418
[2] A Cowlard, W Jahn, CA Empis, G Rein, JL Torero, Sensor Assisted
Fire Fighting, Fire
Technology 46 (3), 2010. http://dx.doi.org/10.1007/s10694-008-0069-1
[3] W Jahn, G Rein, JL Torero, Forecasting fire dynamics using inverse
Computational Fluid Dynamics and Tangent Linearisation, Advances in
Engineering Software (in press), 2011. http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
[4] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
Inverse CFD Modelling Approach in a Real-Scale Fire Test, Fire Safety
Science, 2011, volume 10, pp 1349-1358, (Proceedings of the 10th
International Symposium on Fire Safety Science). http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
[5] G Rein, Inverse Modelling to Forecast Enclosure Fire Dynamics
(invited lecture), Combustion Modelling for Challenging Application,
Spring Technical Meeting, Institute of Physics, Combustion Group,
Southampton. http://www.scribd.com/doc/56141346/Forecasting-Fire-Dynamics-IOP-May-2011
[6] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
Inverse Zone Modelling Approach, in Fire Safety Journal 46, pp. 81–88,
2011. http://dx.doi.org/10.1016/j.firesaf.2010.10.001
[7] Sung-Han Koo, "Forecasting fire development with sensor-linked
simulation", PhD Thesis, University of Edinburgh, 2010.
http://hdl.handle.net/1842/4187
--
*Dr Guillermo Rein*
Senior Lecturer in Mechanical Engineering
University of Edinburgh
http://www.eng.ed.ac.uk/~grein
"so easy it seemed, Once found, which yet unfounded most would have
thought, Impossible!" J Milton
We have used FDS and data assimilation techniques in our forecasting
fire dynamics research. I quote from the thesis of Dr Jahn that I
supervised at the University of Edinburgh (2010) [1]:
"This thesis proposes and studies a method to use measurements of the
real event in order to steer and accelerate fire simulations. This
technology aims at providing forecasts of the fire development with a
positive lead time, i.e. the forecast of future events is ready before
those events take place. A simplified fire spread model is
implemented, and sensor data are assimilated into the model in order
to estimate the parameters that characterize the spread model and thus
recover information lost by approximations. The assimilation process
is posed as an inverse problem, which is solved minimizing a non
linear cost function that measures the distance between sensor data
and the forward model. In order to accelerate the optimization
procedure, the ‘tangent linear model’ is implemented, i.e. the forward
model is linearized around the initial guess of the governing
parameters that are to be estimated, thus approximating the cost
function by a quadratic function. The methodology was tested first
with a simple two-zone forward model, and then with a coarse grid
Computational Fluid Dynamics (CFD) fire model as forward model."
As explained in [2], this work and related technologies allow for the
first time to forecast fire dynamics assimilating sensor data and
solving an inverse problem. It would lead to the implementation of
infrastructure protection in smart buildings. This problem had been
deemed previously as "impossible" by experts.
The is a PhD thesis [1] completed last year. Now all chapters have
been published in the litterature. I think you will be most interested
in the two papers using FDS v5 [3, 4]:
*Forecasting Fire Dynamics Using Inverse Computational Fluid Dynamics
and Tangent Linearisation*
in Advances in Engineering Software, 2011. http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
(at the time of writing this email, the paper is accepted but in
press, and the corrected proofs are not online yet. They will be
available in a few days. Send me an email if you want a personal copy.
It is Chapter 5 in [1]).
and
*Forecasting Fire Growth using an Inverse CFD Modelling Approach in a
Real-Scale Fire Test*
in Fire Safety Science, 2011. http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
A recent presentation at the Institute of Physics - Combustion
Modelling for Challenging Application meeting provides an overview of
the work [5]. Other data assimilation work of the same methodology
using a zone model instead was published in [6]. A closely related PhD
thesis using zone model is [7].
We are most interested in feedback from the fire modelling commmunity.
Your comments are welcome.
G.
[1] Wolfram Jahn, "Inverse modelling to forecast enclosure fire
dynamics", PhD Thesis, University of Edinburgh, 2010. http://hdl.handle.net/1842/3418
[2] A Cowlard, W Jahn, CA Empis, G Rein, JL Torero, Sensor Assisted
Fire Fighting, Fire
Technology 46 (3), 2010. http://dx.doi.org/10.1007/s10694-008-0069-1
[3] W Jahn, G Rein, JL Torero, Forecasting fire dynamics using inverse
Computational Fluid Dynamics and Tangent Linearisation, Advances in
Engineering Software (in press), 2011. http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
[4] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
Inverse CFD Modelling Approach in a Real-Scale Fire Test, Fire Safety
Science, 2011, volume 10, pp 1349-1358, (Proceedings of the 10th
International Symposium on Fire Safety Science). http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
[5] G Rein, Inverse Modelling to Forecast Enclosure Fire Dynamics
(invited lecture), Combustion Modelling for Challenging Application,
Spring Technical Meeting, Institute of Physics, Combustion Group,
Southampton. http://www.scribd.com/doc/56141346/Forecasting-Fire-Dynamics-IOP-May-2011
[6] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
Inverse Zone Modelling Approach, in Fire Safety Journal 46, pp. 81–88,
2011. http://dx.doi.org/10.1016/j.firesaf.2010.10.001
[7] Sung-Han Koo, "Forecasting fire development with sensor-linked
simulation", PhD Thesis, University of Edinburgh, 2010.
http://hdl.handle.net/1842/4187
--
*Dr Guillermo Rein*
Senior Lecturer in Mechanical Engineering
University of Edinburgh
http://www.eng.ed.ac.uk/~grein
"so easy it seemed, Once found, which yet unfounded most would have
thought, Impossible!" J Milton
Rein
Jan 15, 2012, 5:50:04 AM1/15/12
to FDS and Smokeview Discussions, Guillermo Rein
Hi,
I have just been told that the last article (using FDS5) is now fully
published. The reference is:
Forecasting fire dynamics using inverse Computational Fluid Dynamics
and Tangent Linearisation
Advances in Engineering Software 47 (2012) 114–126, 2011
http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
Cheers
G.
> in Advances in Engineering Software, 2011.http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
I have just been told that the last article (using FDS5) is now fully
published. The reference is:
Forecasting fire dynamics using inverse Computational Fluid Dynamics
and Tangent Linearisation
http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
Cheers
G.
> (at the time of writing this email, the paper is accepted but in
> press, and the corrected proofs are not online yet. They will be
> available in a few days. Send me an email if you want a personal copy.
> It is Chapter 5 in [1]).
>
> and
>
> *Forecasting Fire Growth using an Inverse CFD Modelling Approach in a
> Real-Scale Fire Test*
> in Fire Safety Science, 2011.http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
> press, and the corrected proofs are not online yet. They will be
> available in a few days. Send me an email if you want a personal copy.
> It is Chapter 5 in [1]).
>
> and
>
> *Forecasting Fire Growth using an Inverse CFD Modelling Approach in a
> Real-Scale Fire Test*
>
> A recent presentation at the Institute of Physics - Combustion
> Modelling for Challenging Application meeting provides an overview of
> the work [5]. Other data assimilation work of the same methodology
> using a zone model instead was published in [6]. A closely related PhD
> thesis using zone model is [7].
>
> We are most interested in feedback from the fire modelling commmunity.
> Your comments are welcome.
> G.
>
> [1] Wolfram Jahn, "Inverse modelling to forecast enclosure fire
> dynamics", PhD Thesis, University of Edinburgh, 2010.http://hdl.handle.net/1842/3418
> A recent presentation at the Institute of Physics - Combustion
> Modelling for Challenging Application meeting provides an overview of
> the work [5]. Other data assimilation work of the same methodology
> using a zone model instead was published in [6]. A closely related PhD
> thesis using zone model is [7].
>
> We are most interested in feedback from the fire modelling commmunity.
> Your comments are welcome.
> G.
>
> [1] Wolfram Jahn, "Inverse modelling to forecast enclosure fire
>
> [2] A Cowlard, W Jahn, CA Empis, G Rein, JL Torero, Sensor Assisted
> Fire Fighting, Fire
> Technology 46 (3), 2010.http://dx.doi.org/10.1007/s10694-008-0069-1
> [2] A Cowlard, W Jahn, CA Empis, G Rein, JL Torero, Sensor Assisted
> Fire Fighting, Fire
>
> [3] W Jahn, G Rein, JL Torero, Forecasting fire dynamics using inverse
> Computational Fluid Dynamics and Tangent Linearisation, Advances in
> Engineering Software (in press), 2011.http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
> [3] W Jahn, G Rein, JL Torero, Forecasting fire dynamics using inverse
> Computational Fluid Dynamics and Tangent Linearisation, Advances in
>
> [4] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
> Inverse CFD Modelling Approach in a Real-Scale Fire Test, Fire Safety
> Science, 2011, volume 10, pp 1349-1358, (Proceedings of the 10th
> International Symposium on Fire Safety Science).http://dx.doi.org/10.3801/IAFSS.FSS.10-1349
> [4] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
> Inverse CFD Modelling Approach in a Real-Scale Fire Test, Fire Safety
> Science, 2011, volume 10, pp 1349-1358, (Proceedings of the 10th
>
> [5] G Rein, Inverse Modelling to Forecast Enclosure Fire Dynamics
> (invited lecture), Combustion Modelling for Challenging Application,
> Spring Technical Meeting, Institute of Physics, Combustion Group,
> Southampton.http://www.scribd.com/doc/56141346/Forecasting-Fire-Dynamics-IOP-May-...
> [5] G Rein, Inverse Modelling to Forecast Enclosure Fire Dynamics
> (invited lecture), Combustion Modelling for Challenging Application,
> Spring Technical Meeting, Institute of Physics, Combustion Group,
>
> [6] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
> Inverse Zone Modelling Approach, in Fire Safety Journal 46, pp. 81–88,
> 2011.http://dx.doi.org/10.1016/j.firesaf.2010.10.001
> [6] W Jahn, G Rein, JL Torero, Forecasting Fire Growth using an
> Inverse Zone Modelling Approach, in Fire Safety Journal 46, pp. 81–88,
>
> [7] Sung-Han Koo, "Forecasting fire development with sensor-linked
> simulation", PhD Thesis, University of Edinburgh, 2010.http://hdl.handle.net/1842/4187
> [7] Sung-Han Koo, "Forecasting fire development with sensor-linked
>
> --
> *Dr Guillermo Rein*
> Senior Lecturer in Mechanical Engineering
> University of Edinburghhttp://www.eng.ed.ac.uk/~grein
> --
> *Dr Guillermo Rein*
> Senior Lecturer in Mechanical Engineering
Paulo Marques
May 8, 2012, 8:13:56 PM5/8/12
to Rein, fds...@googlegroups.com
Hi,
I want to know how i can do the interaction of the inverse modeling
(TLM) with FDS proposed by Janh Wolfram?
Best Regards
Paulo Marques
On 15 Jan, 11:50, Rein <rei...@gmail.com> wrote:
> Hi,
>
> I have just been told that the last article (using FDS5) is now fully
> published. The reference is:
>
> Forecasting fire dynamics using inverse Computational Fluid Dynamics
> and Tangent Linearisation
> Advances in Engineering Software 47 (2012) 114–126, 2011http://dx.doi.org/10.1016/j.advengsoft.2011.12.005
> > the work [5]. Otherdataassimilationwork of the same methodology
> > > I want to know if is possible some kind ofdataassimilationin FDS?
>
> > > Best Regards,
>
> > > Paulo Marques
I want to know how i can do the interaction of the inverse modeling
(TLM) with FDS proposed by Janh Wolfram?
Best Regards
Paulo Marques
On 15 Jan, 11:50, Rein <rei...@gmail.com> wrote:
> Hi,
>
> I have just been told that the last article (using FDS5) is now fully
> published. The reference is:
>
> Forecasting fire dynamics using inverse Computational Fluid Dynamics
> and Tangent Linearisation
>
> Cheers
> G.
>
> On Jan 6, 12:59 pm, Rein <rei...@gmail.com> wrote:
>
>
>
>
>
>
>
> > Hi Paulo,
>
> > We have used FDS anddataassimilationtechniques in our forecasting
> Cheers
> G.
>
> On Jan 6, 12:59 pm, Rein <rei...@gmail.com> wrote:
>
>
>
>
>
>
>
> > Hi Paulo,
>
> > fire dynamics research. I quote from the thesis of Dr Jahn that I
> > supervised at the University of Edinburgh (2010) [1]:
>
> > "This thesis proposes and studies a method to use measurements of the
> > real event in order to steer and accelerate fire simulations. This
> > technology aims at providing forecasts of the fire development with a
> > positive lead time, i.e. the forecast of future events is ready before
> > those events take place. A simplified fire spread model is
> > implemented, and sensordataare assimilated into the model in order
> > supervised at the University of Edinburgh (2010) [1]:
>
> > "This thesis proposes and studies a method to use measurements of the
> > real event in order to steer and accelerate fire simulations. This
> > technology aims at providing forecasts of the fire development with a
> > positive lead time, i.e. the forecast of future events is ready before
> > those events take place. A simplified fire spread model is
>
> > > Best Regards,
>
> > > Paulo Marques
Paulo Marques
May 9, 2012, 12:18:27 PM5/9/12
to FDS and Smokeview Discussions
For better understand:
I will try to explain with some steps. I think that this is the best
methodology to follow, but if I’m wrong please say.
1º I want to understand the methodology and learn how it works for
implementing in my master degree thesis.
2.º After understanding, i will create a simulation with a fine grid
that can give-me the inputs for your TLM/FDS model, based on the
geometry of a maritime container where I will do some real data
assimilation tests. The objective is to use that values to input as a
real simulation.
3.º Create a real simulation as the step 2 with a container, and try
to make real data assimilation with sensors. ( In this moment I’m
thinking the possibility to use the HPC of windows azure servers to
compute the cfd data.) (possibility to use matlab to read sensors and
work TLM? Or develop a new app?)
I’m a little lost, especially in the integration of TLM with FDS. I
know that I’ve to make some changes in FDS, so I’m studding the
development of it. But witch files to change? How to apply TLM in real
data assimilation and solve cfd with FDS? (P.S. I’ve a problem how to
read the binary files from FDS)
Tank’s for help me,
Cheers,
I will try to explain with some steps. I think that this is the best
methodology to follow, but if I’m wrong please say.
1º I want to understand the methodology and learn how it works for
implementing in my master degree thesis.
2.º After understanding, i will create a simulation with a fine grid
that can give-me the inputs for your TLM/FDS model, based on the
geometry of a maritime container where I will do some real data
assimilation tests. The objective is to use that values to input as a
real simulation.
3.º Create a real simulation as the step 2 with a container, and try
to make real data assimilation with sensors. ( In this moment I’m
thinking the possibility to use the HPC of windows azure servers to
compute the cfd data.) (possibility to use matlab to read sensors and
work TLM? Or develop a new app?)
I’m a little lost, especially in the integration of TLM with FDS. I
know that I’ve to make some changes in FDS, so I’m studding the
development of it. But witch files to change? How to apply TLM in real
data assimilation and solve cfd with FDS? (P.S. I’ve a problem how to
read the binary files from FDS)
Tank’s for help me,
Cheers,
Rein
May 9, 2012, 1:44:23 PM5/9/12
to fds...@googlegroups.com, wolfram Jahn (QMUL)
Hi Paulo,
Great to hear you are working on forecasting dynamics. First I would recomend that you reduce the scope of your Msc thesis becuase it sounds more like a PhD thesis and I think you might run out of time to do all that. Specially time consuming will be the assimilation of real data from the container into FDS.
Before you engage into the full coupling of the TLM (or other optimization) with FDS, I recomnd you use a very simple model as forward model (eg, a plume analytical expression, or a zone model as Wolfram did first) to assimilate data from the litterature or synthetic data from FDS. If this works, little by little move to FDS as the forward model.
Cheers
G.
Great to hear you are working on forecasting dynamics. First I would recomend that you reduce the scope of your Msc thesis becuase it sounds more like a PhD thesis and I think you might run out of time to do all that. Specially time consuming will be the assimilation of real data from the container into FDS.
Before you engage into the full coupling of the TLM (or other optimization) with FDS, I recomnd you use a very simple model as forward model (eg, a plume analytical expression, or a zone model as Wolfram did first) to assimilate data from the litterature or synthetic data from FDS. If this works, little by little move to FDS as the forward model.
Cheers
G.
Rein
May 9, 2012, 1:55:57 PM5/9/12
to fds...@googlegroups.com, wolfram Jahn (QMUL), G.R...@imperial.ac.uk
NOTE: Two weeks ago I gave this plenary at ECCOMASS, Aveiro on the topic
of forecasting. It could be of interest to some.
*Numerical forecasting of fire dynamics*
Plenary Keynote at Young Investigators Conference of the European Community on Computational Methods in Applied Sciences, Aveiro. April 2012.
http://www.scribd.com/doc/91845104
Cheers
G.
*Dr Guillermo Rein*
Senior Lecturer in Mechanical Engineering
University of Edinburgh
http://www.eng.ed.ac.uk/~grein
"The Titanic complied with all codes. Lawyers can make any device legal,
only engineers can make them safe" VM Brannigan
*Numerical forecasting of fire dynamics*
Plenary Keynote at Young Investigators Conference of the European Community on Computational Methods in Applied Sciences, Aveiro. April 2012.
http://www.scribd.com/doc/91845104
Cheers
G.
*Dr Guillermo Rein*
Senior Lecturer in Mechanical Engineering
University of Edinburgh
http://www.eng.ed.ac.uk/~grein
Paulo Marques
May 9, 2012, 2:05:31 PM5/9/12
to fds...@googlegroups.com
I just knew a few days ago. L
--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/ssqafx-UDugJ.
To post to this group, send email to fds...@googlegroups.com.
To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.
Reply all
Reply to author
Forward
0 new messages