1. ## Fire Scenario

I was asked this scenario, and curious if anyone can provide an answer to it along with the formula used to come up with it.

If a two-story wood frame house totally 10,000 square feet (5,000 per floor) were on fire. How long would it take to burn through the wood structure. In addition, how long would it smolder for after collapse.

2. Originally Posted by shumza
I was asked this scenario, and curious if anyone can provide an answer to it along with the formula used to come up with it.

If a two-story wood frame house totally 10,000 square feet (5,000 per floor) were on fire. How long would it take to burn through the wood structure. In addition, how long would it smolder for after collapse.
There is absolutely no way, no way in the world to answer this question. There is no scientific basis for anyone to provide a quantifiable answer.

The reason for this is that there are a multitude of variables that must be considered-way too many to list here.

Computer fire modeling will help you give specific answers to specific situations with specific data. But this scenario is way too general and vague to give any type of intelligent answer.

3. Originally Posted by shumza
I was asked this scenario, and curious if anyone can provide an answer to it along with the formula used to come up with it.

If a two-story wood frame house totally 10,000 square feet (5,000 per floor) were on fire. How long would it take to burn through the wood structure. In addition, how long would it smolder for after collapse.
You could run a fire model. They only have an accuracy rate of 50%.
It could get you into the ball park provided you have all the necessary data points. Which unfortuately you most likely don't if the house burned to the ground.
I am interested as to why this question has been raised. Sounds like someone has a theory about a fire and they are trying to establish a timeline. If you can share more information I will see if I can point in the right direction.
You can send it as a private message if you don't want to post it for all to see.

4. Originally Posted by cubbie
You could run a fire model. They only have an accuracy rate of 50%.
It could get you into the ball park provided you have all the necessary data points. Which unfortuately you most likely don't if the house burned to the ground.
I am interested as to why this question has been raised. Sounds like someone has a theory about a fire and they are trying to establish a timeline. If you can share more information I will see if I can point in the right direction.
You can send it as a private message if you don't want to post it for all to see.
Do you have a citation for your statement that fire models are only accurate 50% of the time?

There are also a hundred ways to get the data points needed to complete a fire model on a black hole.

5. ## Fire Models

Originally Posted by GeorgeWendtCFI
Do you have a citation for your statement that fire models are only accurate 50% of the time?

There are also a hundred ways to get the data points needed to complete a fire model on a black hole.
It’s a quote from Dr. Jim Quintiere over dinner a while back. He was asked how accurate fire modeling really is. His reply was 50/50. Dr. DeHann was ask the same question and provided the same answer 50/50. The qualification provided was the inaccuracy of the data entered into most models. People run models without 100 percent of the data points. Unless you collect all your data points prior to the fire you will never have 100 percent accuracy in any fire model.

Fire models are wonderful tools to confirm a theory. However they are only as accurate as the data collected and correctly inputted. Models have problems accounting for changes in the environment during the fire. Such as opening of windows during the fire. Wind direction and speed changes during the fire. Introduction of suppression efforts and changes in ventilation during the fire by firefighters.

You can get data points at a black hole. However at a black hole there are still going to be hundreds of data points that aren't there anymore. Those missing data points will affect the accuracy of your model.

Personally I would never make a origin and cause determination solely based on a fire model.

6. Originally Posted by cubbie
It’s a quote from Dr. Jim Quintiere over dinner a while back. He was asked how accurate fire modeling really is. His reply was 50/50. Dr. DeHann was ask the same question and provided the same answer 50/50. The qualification provided was the inaccuracy of the data entered into most models. People run models without 100 percent of the data points. Unless you collect all your data points prior to the fire you will never have 100 percent accuracy in any fire model.

Fire models are wonderful tools to confirm a theory. However they are only as accurate as the data collected and correctly inputted. Models have problems accounting for changes in the environment during the fire. Such as opening of windows during the fire. Wind direction and speed changes during the fire. Introduction of suppression efforts and changes in ventilation during the fire by firefighters.

You can get data points at a black hole. However at a black hole there are still going to be hundreds of data points that aren't there anymore. Those missing data points will affect the accuracy of your model.

Personally I would never make a origin and cause determination solely based on a fire model.
I don't think that you understood what they were saying. It sounds like it was an off-the-cuff remark from the sounds of the qualification.

In actuality, computer fire models are ALWAYS 100% accurate. It runs a series of mathematical calculations (always the same) based on the data entered.

Data collection can take an extremely long time. I had to collect data on a fire in a school auditorium. It took two 8 hour days to fully collect the data.

It is also important to know that every time a fire model is run, there are a series of assumptions made within the model. These assumptions can be used to as a substitute for exact data points. For example, I had a fire in a school bus modeled once. NIST had no info to plug into the model for the sheet metal in the bus body, so they made an assumption and used data from a boat fire that involved sheet metal that was determined to be a very close match to the school bus walls.

It is also important to remember that the models are TOOLS. They are not evidence. I am unaware of any court case in this country where a model was admitted as evidence. If anyone is aware of such a precedent, I would be interested in seeing it.

7. Originally Posted by cubbie
I am interested as to why this question has been raised. Sounds like someone has a theory about a fire and they are trying to establish a timeline.
I just had the same topic come up on a fire a few weeks ago. It was a total burn-down of a second home that had been closed up for the season. No FD response because nobody ever saw it (or if they did, they didn't call it in). The question came up as to the time line, especially since there was nearby lightning four days prior to discovery.

Ended up with more questions than answers.

Brian

8. Originally Posted by SWFD22
I just had the same topic come up on a fire a few weeks ago. It was a total burn-down of a second home that had been closed up for the season. No FD response because nobody ever saw it (or if they did, they didn't call it in). The question came up as to the time line, especially since there was nearby lightning four days prior to discovery.

Ended up with more questions than answers.

Brian
"Black Hole" fire investigations are among the most challenging investigations a fire investigator will face. The most important factor is not to jump to conclusions. At the end of the day, if the answer is still "I don't know", be honest and say "I don't know". There is no shame in an undetermined fire.

9. Originally Posted by GeorgeWendtCFI
I don't think that you understood what they were saying. It sounds like it was an off-the-cuff remark from the sounds of the qualification.

In actuality, computer fire models are ALWAYS 100% accurate. It runs a series of mathematical calculations (always the same) based on the data entered.

Data collection can take an extremely long time. I had to collect data on a fire in a school auditorium. It took two 8 hour days to fully collect the data.

It is also important to know that every time a fire model is run, there are a series of assumptions made within the model. These assumptions can be used to as a substitute for exact data points. For example, I had a fire in a school bus modeled once. NIST had no info to plug into the model for the sheet metal in the bus body, so they made an assumption and used data from a boat fire that involved sheet metal that was determined to be a very close match to the school bus walls.

It is also important to remember that the models are TOOLS. They are not evidence. I am unaware of any court case in this country where a model was admitted as evidence. If anyone is aware of such a precedent, I would be interested in seeing it.
Fire modeling programs calculate several characteristics that all lean towards fluid mechanics. Characteristics that are not computed include (1) the ignitability of objects from small flames. (2) The spread of fire over surfaces. (3) The actual size of the fire (heat release rate).
To overcome these problems a modeler gives it their best guess and manually enters the data before running the model.

Reference: Fire Modeling: An introduction for attorneys (Dr. Vytenis Babrauskas)
History Note: Dr. Babrauskas released the first fire modeling computer program in 1975.

If fire modeling has a zero error rate; as you have stated and I quote: “In actuality, computer fire models are ALWAYS 100% accurate. It runs a series of mathematical calculations (always the same) based on the data entered.” Why do anything else but models if they are 100% accurate? Why spend long hours collecting data for a model that you say can’t be use as evidence? Why would you not have your model admitted as evidence if its a tool that is 100% accurate?

“In actuality, computer fire models are ALWAYS 100% accurate. It runs a series of mathematical calculations (always the same) based on the data entered.
The data generally is reliable enough to use the model as a tool, but it would not be relaible enough to pass a Frye or Daubert test. The science is not there. A fire model can be a very useful tool. It is not evidence.

11. Originally Posted by GeorgeWendtCFI

The data generally is reliable enough to use the model as a tool, but it would not be relaible enough to pass a Frye or Daubert test. The science is not there. A fire model can be a very useful tool. It is not evidence.
Accuracy is important in a general context. The user must evaluate the absolute accuracy of the model in the context of "how accurate do I have to be to answer the technical issue at hand"? Courts have accepted computer fire modeling because it is generally relied upon in the fire protection engineering and fire investigation communities. It should also be noted that the most widely used fire models are products of the federal government and are admissible by default in federal courts.

Example of a fire model admitted in court as evidence would be in Rhoad Island. When more than 100 people died in a night club fire. The NIST model was accepted in court.
There has to be an acceptable rate of error or the model can be excluded.

12. Did the Rhode Island case go to trial? It's my understanding that the criminal litigation resulted in plea bargains and the civil litigations were all settled. If there was no trial, there was no evidence. If it had gone to trial, I would bet that there would have been a Frye or Daubert challenge to the model. Something can also be admitted as an exhibit in court, but not as evidence.

I agree that there would be an "acceptable" error rate. But that error rate will not be in the model itself. The model is designed to run a complex series of mathematical equations. The model is incapable of getting a wrong answer.

The error rate would be in the overall process, including data collection, data integrity and data input. That is where humans take over and that is where any error would occur.

In addition, establishing the error rate would be quite difficult.

13. I think we have maybe reached the same conclusion but we are each looking at the problem from a different prospective.
If we start with base answer of 30 as the correct answer to the problem. If I collect data points that lead me to enter 3 X 5 in my model. I will get an answer of 15. The model is 100 % accurate based on the data entered. The answer will remain 15 no matter how many times I enter 3 X 5 into the model and run it.
It’s still the wrong answer because the data entered was flawed. To get the correct answer I would have needed to enter 6 X 5 into the model.
Your interruption is the model is 100 % accurate because 3 X 5 does equal 15.
My interruption is the model is only 50 % accurate because the correct answer to the problem was 30 not 15. This is what Dr. Quintiere and Dr. DeHann were referring to when they said most models are only 50% accurate.

I will do some more research on fire models being admitted as evidence. I have checked several sources who tell me that there are cases.

14. Originally Posted by cubbie
I think we have maybe reached the same conclusion but we are each looking at the problem from a different prospective.
If we start with base answer of 30 as the correct answer to the problem. If I collect data points that lead me to enter 3 X 5 in my model. I will get an answer of 15. The model is 100 % accurate based on the data entered. The answer will remain 15 no matter how many times I enter 3 X 5 into the model and run it.
It’s still the wrong answer because the data entered was flawed. To get the correct answer I would have needed to enter 6 X 5 into the model.
Your interruption is the model is 100 % accurate because 3 X 5 does equal 15.
My interruption is the model is only 50 % accurate because the correct answer to the problem was 30 not 15. This is what Dr. Quintiere and Dr. DeHann were referring to when they said most models are only 50% accurate.

I will do some more research on fire models being admitted as evidence. I have checked several sources who tell me that there are cases.
I didn't have an "interruption". I had an interpretation.

I'm done with this discussion.

15. Originally Posted by GeorgeWendtCFI
I didn't have an "interruption". I had an interpretation.

I'm done with this discussion.
Good enough.

My interpretation is very simple garbage in garbage out.

16. I hate to get involved in your ****ing match but I also have a few comments.
As to the original question, good luck, I believe anyone would just be guessing as to the answer.
As for the Fire Modeling, I haven't heard yet where this has been introduced and stayed admissible as evidence, educate me if it has. If it has, that scares me to death because I do not feel that these models are good enough yet to be court room evidence that may put someone in prison or ruin their life. Interesting yes, evidence, no.
I was at the Fire Academy in the Modeling class, I ran the exact same numbers on the same scenario twice and did get different results. They where close, but not exact, in this instance, the model was not 100% accurate.
I think the models are beneficial for checking out your hypothesis, but shouldn't be relied upon over good investigations, and are not yet ready for court.
My two cents, let the beatings begin..
Richard Howard
CFEI, CFI, K-9 Handler
Evansville Fire Department (Also formerly with Unified, Indianapolis)
Evansville, Indiana

17. Originally Posted by ussmertzdd691
I hate to get involved in your ****ing match but I also have a few comments.
As to the original question, good luck, I believe anyone would just be guessing as to the answer.
As for the Fire Modeling, I haven't heard yet where this has been introduced and stayed admissible as evidence, educate me if it has. If it has, that scares me to death because I do not feel that these models are good enough yet to be court room evidence that may put someone in prison or ruin their life. Interesting yes, evidence, no.
I was at the Fire Academy in the Modeling class, I ran the exact same numbers on the same scenario twice and did get different results. They where close, but not exact, in this instance, the model was not 100% accurate.
I think the models are beneficial for checking out your hypothesis, but shouldn't be relied upon over good investigations, and are not yet ready for court.
My two cents, let the beatings begin..
Richard Howard
CFEI, CFI, K-9 Handler
Evansville Fire Department (Also formerly with Unified, Indianapolis)
Evansville, Indiana

Thank you, I agree completely.
Fire models are only as accurate as the data used to create the model.

18. Originally Posted by ussmertzdd691
I hate to get involved in your ****ing match but I also have a few comments.
As to the original question, good luck, I believe anyone would just be guessing as to the answer.
As for the Fire Modeling, I haven't heard yet where this has been introduced and stayed admissible as evidence, educate me if it has. If it has, that scares me to death because I do not feel that these models are good enough yet to be court room evidence that may put someone in prison or ruin their life. Interesting yes, evidence, no.
I was at the Fire Academy in the Modeling class, I ran the exact same numbers on the same scenario twice and did get different results. They where close, but not exact, in this instance, the model was not 100% accurate.
I think the models are beneficial for checking out your hypothesis, but shouldn't be relied upon over good investigations, and are not yet ready for court.
My two cents, let the beatings begin..
Richard Howard
CFEI, CFI, K-9 Handler
Evansville Fire Department (Also formerly with Unified, Indianapolis)
Evansville, Indiana
How did you like working for Unified?

19. I like Unified, I think they are a really good outfit. Problem was, Mike Reynolds hired me, (a good but goofy guy), when he left to open up the Florida office, the new guy was not giving me any work. After a year of begging for work, I finally left. I signed a 1 year no compete so I left so that I could start my year. If mike was still here, I would still be working with them.
They spent a lot of money on me in the beginning to waste me like the new guy did. But even with that, the top brass are good guys, Eddie is a trip, I would never bad mouth the company, only the new boss of Indianapolis.
R. Howard

20. Originally Posted by ussmertzdd691
I hate to get involved in your ****ing match but I also have a few comments.
As to the original question, good luck, I believe anyone would just be guessing as to the answer.
As for the Fire Modeling, I haven't heard yet where this has been introduced and stayed admissible as evidence, educate me if it has. If it has, that scares me to death because I do not feel that these models are good enough yet to be court room evidence that may put someone in prison or ruin their life. Interesting yes, evidence, no.
I was at the Fire Academy in the Modeling class, I ran the exact same numbers on the same scenario twice and did get different results. They where close, but not exact, in this instance, the model was not 100% accurate.
I think the models are beneficial for checking out your hypothesis, but shouldn't be relied upon over good investigations, and are not yet ready for court.
My two cents, let the beatings begin..
Richard Howard
CFEI, CFI, K-9 Handler
Evansville Fire Department (Also formerly with Unified, Indianapolis)
Evansville, Indiana
It may have been unintentional, but if you ran the same numbers twice and got a slightly different result, there must have been a subtle difference in the data submitted.

Use a calculator and run any equation using the same numbers 10, 100 or a 1000 times. You will get the same answer. That is what the model does thousands of times per second. It is scientifically impossible to run the same equations with the same numbers and come up with different answers. It cannot be done.

Side note; ussmertzdd691, can you email me at wendtcfi@optonline.et so I can email you a question? Thanks.

21. Of course an error in transposing numbers could have occurred. I was diligent in imputing the numbers but who knows? I am still not ready to rely on them to prove someones guilt.
Richard Howard

22. Originally Posted by ussmertzdd691
I am still not ready to rely on them to prove someones guilt.
Richard Howard
And neither am I. Under any circumstances. One should look at fire models as a tool. Fire modelling has a number of valid uses in the fire investigation process, but using it as a proof of fact is a very big mistake.

23. George, we agree.
I tried using your link to e-mail and it did not go through.
You can e-mail me at rhoward@evansvillefiredepartment.com or arsonk9@insightbb.com
Make any questions easy one's.

24. AMEN I could'nt agree more

Originally Posted by GeorgeWendtCFI
And neither am I. Under any circumstances. One should look at fire models as a tool. Fire modelling has a number of valid uses in the fire investigation process, but using it as a proof of fact is a very big mistake.

There are currently 1 users browsing this thread. (0 members and 1 guests)

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts