Emerging Trends in Kitchen Fire Investigation

Emerging Trends in Kitchen Fire Investigation

Stovetop Fires and the Ignition of Cooking Oils

Peter D. Layson
Wesley C. Richardson
Orion P. Keifer
Applications Engineering Group, Inc.
1200 Mayport Road, Atlantic Beach, Florida 32233, (800) 777-7668 www.aegiforensics.com

K. David Cheers
Jack A. Ward
Jack Ward Fire Consultants, LLC
P. O. Box 16467, Jacksonville, Florida 32245
(866) 559-3473 jackwardfire.com

Correspondence author:  David Cheers, Phone 866-559-3473, Fax 904-278-7292, E-mail davidcheers@jackwardfire.com

INTRODUCTION

In October 2004, the United States Fire Administration (USFA)/National Fire Data Center (NFDC) published the results of a study on kitchen fires based on the 2002 statistics from the National Fire Incident Reporting System. The paper identified that fires originating in kitchens were responsible for thirty percent of all reported structure fires. These kitchen fires were responsible for three hundred thirty two fatalities, or twelve percent, of all fatalities in structures, and resulted in $876,000,000 in property loss. Not surprisingly, cooking operations were responsible for nearly ninety percent of all kitchen fires, with unattended cooking operations as the leading cause. The USFA/NFDC study found that in over half of all kitchen fires, the cooking material was the first fuel ignited, and in nearly forty percent of those cases the fuel was oil, fat or grease.

The sizable impact caused by cooking fires warrants investigation from both the public and private sectors. These investigations require the application of fire science in developing hypotheses to identify how the range or stovetop came into contact with a fuel for a sufficient duration to cause ignition.  A detailed evaluation of the sequence of events, from the heating equipment used to the time involved, can provide valuable information towards identifying accidental cooking fires or debunking witness testimony like the often heard “the pan of oil burst into flames two minutes after I turned the burner on”.

The purpose of this paper is to help fire investigators understand the characteristics and time tables of oil/grease fires, and to use as benchmarks in evaluating if the scenario given is consistent with an accidental fire.

 

TESTING

In 2003, testing funded by Jack Ward Fire Consultants, State Farm Insurance Company, The Florida Advisory Committee on Arson Prevention (FACAP), and Applications Engineering Group, Inc. (AEGI), was performed in a controlled laboratory environment at AEGI in accordance with NFPA 921. The purpose was to determine the parameters pertinent to oil/grease fires and their effects, including burner settings, size, oil type and oil quantity.  A range with Calrod (open coil) burners was used.  The results were presented at the FACAP annual training seminar in Ocala, Florida.

As newer flattop/ceramic glass top ranges and cooktops were utilized, a need arose to identify the time and temperatures involved to attain ignition compared to the older Calrod burners. These tests, funded by Jack Ward Fire Consultants and AEGI, were conducted in 2010 and 2011.

Testing was conducted on oils commonly used in cooking.  Chart 1 lists the types of oils tested.  In the instrumentation, several Type K thermocouples were used.  For the oil temperatures, presented, the thermocouple was completely immersed in the oil, but not touching the bottom of the pan.  Surprisingly, while the smoke temperatures of the different oils varied, the ignition temperatures for the different oils were very close.

 

Another consideration in determining ignition times is fuel age. A series of tests were performed to determine the flashpoints of vegetable oil which had been aged up to 8 hours at 204 Celsius (°C) [(400 Fahrenheit (°F)].  Every hour a sample was tested for its ignition temperature by placing a drop of it on a stainless steel/aluminum clad 25 cm (10”) pan over a hot plate, to give finer temperature control. A K type thermocouple was placed on the bottom of the pan, to measure the temperature. Drops of oil were placed on the pan, some of which autoignited and some of which did not, until the autoignition temperature was determined.  The new vegetable oil had an autoignition temperature of 446°C (835°F), and the vegetable oil aged for 8 hours ignited at 371 °C (700°F), which is a decrease of approximately 10°C (17°F) per hour.  Chart 2 shows the correlation between the autoignition temperatures of oil samples versus the total age time that they were heated.  After testing for the autoignition temperature of aged fuel, 100 mL quantities of oil aged for 0, 2, 4, and 8 hours were then heated on a large Calrod burner in a 25 cm (10”) aluminum pan on high until ignition was achieved. Chart 3 shows the correlation between the time until autoignition of the oil samples versus the time aged.

 

 

In preparation for the range/stovetop testing, three free standing electric ranges were purchased. It should be noted that no propane or natural gas fired ranges were utilized. The ranges chosen were a General Electric flattop range (model JBS55DM2WW), a Whirlpool flattop range (model WFE324LWQ0), and a General Electric open coil “Calrod” range (model JBP26WY6WW).  The flattop ranges tested have glass-ceramic tops of the most common, infra-red/radiative heating design.  It should be noted that some very old flattop models heated by conduction and some models heat by induction; however, they are relatively rare.  Several different variables were evaluated during the testing, including:

• Burner type (flattop/Calrod)
• Burner diameter [14 cm 5.5”), 16.5 cm (6.5”), 19 cm (7.5”), 20 cm (8”), and 25 cm (10”)]
• Fuel volume [100 mL (3/8 cup), 300 mL (1 ¼ cup), 650 mL (2 ¾ cup)]
• Fuel type (vegetable oil, new canola oil, used canola oil, olive oil, gasoline)
• Container type (aluminum Teflon pan, cast iron pan, aluminum pot)

An aluminum plate, instrumented with thermocouples imbedded in the plate was used to test the thermal characteristic of the burners.  The burners on the flattop ranges set on “high” cycled on and off, never exceeding 572°C (1061°F), nor did they get hot enough to melt the calibrated aluminum plate.  However, the burners on the Calrod range on “high” did not cycle off, allowing temperatures to reach in excess of 593°c (1100°F), melting the calibrated aluminum plate.

A matrix of tests was performed using various burners, oil levels, and cooking containers.  Additional testing was conducted with mixtures of oil and gasoline.  The parameters of interest include the time for the oil to begin to smoke and time the oil reaches its autoignition temperature, bursting into flames.  A tabular summary of the testing is included in Appendix A.  Several important trends were observed and will be discussed below.

The testing revealed that significant oil vapor was produced and became easily noticeable as the oil ignition temperature was approached. This vapor produced an acrid odor irritating the eyes and nasal passages of the testing participants as it spread throughout the testing area.  The vapor was readily detectable by the olfactory system well in advance of the oil ignition. Once ignition occurred, darker smoke was generated which was also readily detectable even though the fire remained confined to the pot or pan. The production of oil vapor increased as the volume of oil was increased. Charts 4 and 5 show the difference between the smoke and ignition points for a few of the tests performed on the GE flattop and GE Calrod ranges.

 

 

Ignition did not occur when 3000 mL [approximately 12.5 cups or 6.5 cm (2.5”) in a 25 cm (10” pot)] of oil was heated on the largest available burner, although vapor production was significant and occurred over a long period of time.

As noted previously, the Calrod burners remain on when in the high position.  As expected, when the same container and volume of oil was tested on high, the time until vapor production and ignition was significantly shorter for the Calrod burners than the flattop burners.

During four of the tests, a known amount of gasoline was mixed with the cooking oil in order to establish if this additional fuel would alter the ignition time.  It was observed that the gasoline vaporized prior to reaching its ignition temperature adding no noticeable fuel once ignition occurred. The time until ignition with the gasoline/oil mixtures was consistent with the tests performed using the same respective oil volume. The only noticeable change with the addition of the gasoline was that piloted ignition of the gasoline, and subsequently the cooking oil, could be achieved.

Additional tests were performed to measure the flame height of various oil quantities in a pan. Oil volumes of 100, 75, 50 and 25 mL were heated in a 25 cm (10”) stainless steel pan on a 19 cm (7.5”) Calrod burner on high until ignition was achieved. During the test, wood was placed 45.7 cm (18”) above the pan to simulate a cabinet. The fire was allowed to progress until either the cabinet caught fire or the fire self-extinguished. Only the 25 mL volume test did not cause ignition of the cabinet. This indicates that oil quantities less than 25 mL are unlikely to communicate into the above cabinets causing catastrophic fire damage.

The next series of tests involved heating 100 mL of oil in 3 different containers to record their thermal characteristics via a forward looking infrared camera (FLIR). The three containers were: a 25 cm (10”) diameter and 4.4 cm (1 ¾”) high Teflon coated aluminum pan, a 25 cm (10”) diameter cast iron pan, and a 3 liter (3 quart) aluminum pot.  A greater temperature gradient could be seen with both the cast iron pan and the aluminum pot, while the temperature distribution in the aluminum pan was minor. However, in each case it is clearly seen that only the bottom of a pan needs to reach the ignition temperature to allow ignition of the oil. This phenomenon is shown in photographs 1, 2 and 3.

 

 

 

The final range test was designed to determine whether a 19 cm (7.5”) Calrod burner set to medium-high could ignite 100 mL of vegetable oil in a 25 cm (10”) Teflon aluminum pan.  Despite heavy smoke and an acrid odor, ignition did not occur and the oil was mostly vaporized after 45 minutes.  The residual oil in the pan resembled burnt honey, and it reached a maximum temperature of 338°C (640°F).  This test confirms that not all range burners are capable of igniting cooking oils when set to the medium-high temperature.

From this testing, several conclusions can be made:

1. Oils and fats used in cooking have different temperatures at which they smoke, but their autoignition temperatures are similar [around 371°C (700°F)]).

2.  When the oil is smoking, it releases an acrid smelling and easily detected odor.

3.  As oils age at high temperatures, the ignition temperature decreases.  The effect is normally negligible for most residential cooking, but may be significant in a restaurant where the oil may stay at high temperatures for extended periods of time between changes.

4. The size and type of burner affects the speed at which the oil will reach autoignition temperature.

5.  The setting on the burner affects the speed at which autoignition temperature is reached.  Setting at medium and below is unlikely to cause the oil to reach the autoignition temperature; however, realize that burner controllers can vary, therefore the setting (for example medium) on one burner/range may not and probably does not correspond to the same setting on another burner/range.

6.  The amount of oil affects the time necessary to reach autoignition temperature.  Volumes of 25 ml [approximately 1/8 cup or 1 mm (one thirty-second of an inch) in a 20cm (8”) pan] and less are unlikely to cause a fire.  Likewise, very large volumes are also unlikely to cause a fire.  At volumes in the intermediate range, more oil takes longer to reach the autoignition temperature.

INVESTIGATION

The fire investigator must preserve the details of the range fire, most importantly, the volume of oil, the pan/pot used and the burner type and setting in order to use the data presented.  It is best practice to collect the pan, the range and the container of oil used in the cooking operation. These items should be placed into evidence for detailed evaluation in a laboratory setting where the cooking scenario can potentially be reproduced. If this is not possible, carefully document the evidence photographically and/or detail in the field notes.  Many times after a cooking fire has occurred, evidence of the operation still exists such as food particles or oil residue in the pan. To compare the amount of oil used to known data, one must look carefully at the pan as well as the surroundings. Statements of those who have placed oil in the pan are valuable, and those statements can be compared to empty/partially full cooking oil containers recovered from the room or trash receptacles. A “new” container of oil that was recently used can give clues as to how much oil was placed in the pan.  The damaged versus undamaged areas on the interior of the pan can be measured and provide clues as to the amount of oil in the pan at the time of the fire event (Photograph 4).

 

The range or cooktop must be carefully examined to determine the pre-fire position of the control knobs. In fires with minimal damage, this can be accomplished by viewing the orientation of the knob in relation to the index mark on the control panel.  In cases where the knob has been turned to the ‘off’ position after the fire, thermal damage is often present that will indicate its position at the time of the fire.  In the event the range or flattop suffered heavy heat damage, a forensic examination of the control modules’ internal components can reveal the orientation of the knob. Comparison to exemplar modules in the ‘off’ position can yield valuable information regarding a burner’s involvement in a fire (Photographs 5-7).

 

 

The above information can then be used to estimate the time that the oil was heating on the range and an estimate of when the oil will begin to smoke and finally autoignite.  Statements of the person cooking can be compared with the known testing data to evaluate if the presented scenario is consistent with the physical evidence found by the fire investigator.  Nonetheless, when encountering kitchen fires reported to have occurred as the result of an unattended cooking operation, caution should be used before attaching a label as to the cause. NFPA 921, Chapter 19, 2011, states the guidelines that may be used to classify the cause of the fire.  Section 19.5.1 addresses the issue of assigning responsibility and discusses the ‘Nature of Responsibility’.  Sometimes there is an inability to distinguish between an occupant who became distracted during a cooking operation and accidentally started a fire, versus an occupant who intentionally placed cooking oil in a pan on an energized burner and intentionally allowed the fire to develop. The Nature of Responsibility discusses that a fire may be related to an act or an omission, and can be done accidentally or intentionally. Simply stated, the act of leaving a pan of oil on an energized burner will result in the same fire patterns and heat damage regardless of whether it was accidental or intentional.  A prudent step in cooking fires, where the intent of the occupant cannot be established, is to opt out of the classification of a cause and state the facts. A human act, either accidentally or intentionally, resulted in the energized burner providing the heat required to ignite the oil or cooking material within the pan.  Although NFPA 921 provides that an undetermined cause could be utilized in a case where intent is not known, it is a disservice to those involved to use this classification when the complete ignition sequence can be readily identified yet the intent of the occupant is in question.

 

SUMMARY

The testing provided scientific data and parameters that can be considered when examining fires related to cooking operations.  The data will also prove useful in establishing minimum time frames required from the initial application of heat via a cooking surface to a pan of oil placed on top of it. The ability to establish time frames based on known amounts of oil and known sizes of burners and pans can prove useful when attempting to verify the reported ignition times from a witness statement. The testing further highlighted the need for both additional detailed information as well as a systematic forensic examination of the physical evidence related to the cooking fire.  A proper investigation paired with the formal study provides the investigator a more accurate assessment of whether a fire was accidentally or intentionally set.

 

REFERENCES

1. United States Fire Administration. National Fire Data Center, “Kitchen Fires”, Topical Fire Research Series, Volume 4, Issue 4, October 2004

2. National Fire Protection Association, “NFPA921 Guide for Fire and Explosion Investigations”

2002 and 2011 Editions

3.  Institute of Shortening and Edible Oils, “Food Fats and Oils”, Ninth Edition, 1999

4. Babrauskas, Vytenis, “Ignition Handbook, Principles and applications to fire safety engineering, fire investigation, risk management and forensic science”, Fire Science Publishers, Division of Fire Science and Technology, Inc.  Issaquah, WA, 2003