a review of soil gas migration in natural gas...

A Review of Soil Gas Migration in Natural Gas Incidents

Presentation to the

Western Regional Gas Conference

Henderson, NV

August 29, 2018Mr. David W. Heldenbrand, PE CFI

Bison Engineering, Inc.

Special Thanks to Mr. Jeff Winship, Pixel Graphics Inc.

David Heldenbrand, PE CFI

IntroductionIn 1988, I participated in a soil gas migration study right here inHenderson, Nevada. It was conducted by others, but a great dealof data was produced and analyzed.

The testing in Henderson in 1988 was specifically designed to testthe theory that gas migrates through an annular space around thepipeline.

It was necessary in that case for gas to travel 700 feet along theoutside of a 16” pipeline in order for that theory to explain theexplosion. That theory was found to be incorrect.


Introduction

Summary of the Nevada Soil Gas Migration Test Results

– Minor concentrations of tracer gas were detected as far as 90 feetfrom the source.

– A positive gas flow rate through the soil was measured 70 feet fromthe source.

– Visual indications showed gas was pushed through asphalt as far as20 feet radially from the source and could be seen pushing dustthrough the asphalt.


IntroductionSince I measured a small gas flow rate 70 feet from the source, I begantesting whether gas explosions occur farther than a 50 foot radius from aleak.

I have found that after reviewing numerous cases that gasexplosions farther than 50 feet from a gas leak are very rare.


IntroductionWhy did I do this presentation?

NO meaningful studies regarding this subject have ever been conducted.

This presentation will:

• Examine the results of the Nevada soil gas migration testing program.

• Review a number of NTSB and state OPS reports.

• Summarize data from cases in which I have been personally involved over the past30 years.

• Propose improvements for future investigations.


Outline of Presentation

• Does Gas from a Leak “Follow the Pipeline”?– What Do Bar Hole Surveys Show?

• Does Natural Gas from a Leak Only Migrate Up?• What Gas Flow Rate into a Structure is Necessary to Cause an Explosion?• How Does Soil Type Affect Soil Gas Migration?• How Does Pipeline Pressure Affect Soil Gas Migration?• What do the Cases that Occur over 100 Feet from a Leak have in Common?• How Does the Leak Flow Rate Affect Soil Gas Migration?• How Much Time Does it Take for Gas to Migrate through the Soil and Cause

an Explosion?


Outline of Presentation

• Summary

• What Might These Cases Have in Common?

• What Does All This Mean?

• How Does a Soil Gas Migration Explosion Occur?

• How Do Sewer Lines or Other Utilities Affect Soil Gas Migration?

• Suggestions for Future Investigations.

• Practical Applications for First Responders

• Risk Assessments

• Odorant Scrub

• Conclusions


Does Gas from a Leak Follow the Pipeline?



• The soil conditions at the Nevada site were dry, well graded sandand gravel, so it was a highly permeable soil.

• The backfill for the pipeline was the original well-consolidated soil.

• After three weeks of testing, no evidence was generated toconclude that gas “followed the pipeline” to cause this explosion.

• As I encountered other gas explosions, I observed numerousdifferent conditions, but very few cases were found where the gasmigrated significantly greater than 50 feet to cause an explosion.

• Let us examine some specific examples.


Does Gas Follow the Pipeline?

The two theories of whether gas “follows the pipeline” are:

• An annular space theory

• A theory of loose backfill material

• Most cases, however, involve no loose or porous backfill



Theoretical “Annular Space” Caused by Temperature or Pressure Variations

• This is an Autocad drawing of a theoretical “annularspace”. Take note that there are no dimensions

• In addition, there are:• No photos of any annular space• No measurements of any annular space• No soil parameters for any annular space• No length parameters• No temperature variation parameters• No pressure variation parameters• No flow rate parameters• In fact, no studies whatsoever have ever

verified that this theoretical “annular space”actually exists

I contend that there is no annular space at all


Let’s examine the following cases:

• Ft. Worth, TX- October 4, 1971

• Odessa, WA - December 26, 2008

• Williamsport, PA - January 25, 1977

• Jackson, MS- December 24, 2008

• Rancho Cordova, CA- December 24, 2008

• Bowie, MD- June 23, 1973


Fort Worth, Texas - October 4, 1971

The gas service that fractured was to a residence on one side of the streetbut the explosion took place at the residence closest to the leak.

Interestingly enough, a second explosion occurred in N. Richland Hills, Texas(which is near Ft. Worth) on the same day. Heavy rains occurred in the areathe day of both explosions.



• Odessa, WA- The structure was 60 feet from the leak, but there was no gasservice to the structure. The initial OPS investigation concluded that naturalgas was not the cause of the fire specifically because there was no gasservice to the structure.

• Williamsport, PA- The structure was 70 feet from the leak but, again, no gasservice to the structure.

• Jackson, MS- Similar to the case in Ft. Worth, TX. A leak was found on oneservice line but the fire occurred at the house closest to the leak. Also, thefire was on the opposite end of the house from the gas service line andmeter.

• Rancho Cordova, CA- Service line was on the opposite side of the house fromthe leak on the main.



Bowie, MD

The Bowie, MD case is interesting because the spread of the gasis both uphill and downhill in an elliptical shape. It is not relatedto the route of the pipelines and no new or select backfill wasreported.

The gas traveled 120 ft. along a drainage path to cause theexplosion.



Leak Location

Does Gas from a Leak Follow the Pipeline?Summary

The annular space theory is that the pipeline swells and shrinkswith temperature and/or pressure variations thus creating a voidalong the surface of the pipe for the gas to travel.

There are no photographs, no measurements, no researchpapers, or any other evidence of an “annular space”. So,contrary to common perceptions, there is, simply, no evidencethat this “annular space” actually exists and there is noevidence that gas “follows the pipeline”.


What Do Bar-hole Surveys Show with Regard to Gas Following the Pipeline?

Bar hole surveys are extremely useful at locating a leak in thepipeline.

So, where does one typically conduct the survey to find a leak?---On top of the route of the pipeline.

This tends to reinforce the theory that “gas follows the pipeline”.

Let’s look at some actual data.



We will examine several cases as follows:

• Portales, NM- May 23, 1982

• Queen Creek, AZ- Feb 2, 2004

• Crystal Springs, MS- February 4, 1986

• Bowie, MD- June 23, 1973


Portales, NM

Queen Creek, AZ


Minimal explosion damage- Gas came through the pipe chase in the floor two weeks after the gas leak was stopped



Watkins Furniture Store, Crystal Springs, MS

This explosion was caused when a contractor cut the PE gas line and repaired ithimself with a PVC coupling.

The gas line pulled out of the coupling several years later. The gas migrated about50 feet perpendicular to the water line ditch to cause the explosion.

The gas also migrated along the sandy backfill of the water line 400 feet downhilland 100 feet uphill in the ditch line. There was no gas service to the structure.



Crystal Springs, Mississippi


Bowie, MD

The migration pattern of the gas is clearly not related to theroute of the gas line.

The ground surface slopes downward from the house thatignited.

Some of the gas went uphill and some of the gas went downhilland at an angle to the gas lines.


Bowie, MD again. Migration pattern is:172 feet uphill318 feet downhill145 feet acrossThe house that exploded was 120 feetuphill and the closest house was 72 feetfrom the leak.

The elliptical migration pattern is notrelated to the pipeline or the asphalt, butmight be related to drainage pattern.

Elevated soil gas readings were recordeda year after the explosion.



Leak Location


• These four cases are representative of many bar-hole surveys.

• These bar-hole surveys show that:

• Gas spreads out in all directions for the most part. Yes, somefactors influence the pattern (Bowie, MD) but for most cases, thegas migration pattern is radial or elliptical from the leak.

• So, investigators cannot assume that the gas “follows thepipeline” to cause a fire or explosion.


Does Natural Gas from a Leak Only Migrate Up?



The specific gravity of natural gas is about 0.60. Natural gas inopen atmosphere is lighter than air and tends to float up.

But, we see numerous cases where natural gas is forceddownward through the soil. This must be caused by gaspressure.



What evidence is there on this subject?

• Sepulpa, Oklahoma- About 1987

• Saratoga Springs, Utah- 2/07/2007

• Holly Springs, Mississippi- 4/03/2008

• Jackson, Mississippi- 12/24/2008

• Queen Creek, Arizona- 2/2/2004



• I first had questions from an explosion in Sepulpa, Oklahomaabout 1987.

• A building exploded after a plastic gas line had cracked.

• The gas line was removed about a month prior to my siteinvestigation.

• Elevated bar hole readings were recorded a month after thepipeline was removed.

• If gas only migrated up, there should have been no gas in theground.



• Saratoga Springs, Utah, a gas line was struck by a directional drilling crewabout 40 feet from a home.

• The house was constructed with poured concrete basement walls and apoured concrete floor slab.

• A simulated leak from the leak location with an air compressor showedflow into the basement all along an expansion joint seam at the bottom ofthe basement wall.

• This shows not only that gas can be pushed downward but gas pressurewas forcing gas into the basement.

• The simulated leak was witnessed immediately upon the startup of thecompressor.



Holly Springs, Mississippi

• A cracked cast iron line leaked for an unknown period of time.

• Bar-hole surveys several weeks later showed elevated LEL readingsbetween the gas line and the home.

• No natural gas service line to the home

• The ground sloped significantly downward to a creek behind the home.

• The home had a basement constructed of concrete block walls.

• The leaking gas was forced over 40 feet laterally and 3 to 5 feet below thelevel of the pipeline. These results are similar to the previous case in Utah.



• Another similar case occurred in Jackson, Mississippi on December24, 2008.

• A service line pulled out of a coupling to a house located uphillfrom the leak.

• If gas traveled uphill and followed the gas line, that would havebeen the house that would have burned.

• Instead, the house located downhill but, again, closer to the leakignited.

• The gas service line was on the opposite end of the house from thesmall fire.



Jackson, Mississippi


• Queen Creek, Arizona• On January 15, 2004, an odor complaint resulted in the location of

a pipeline leak.• Two weeks after the leak was repaired, a very small explosion

occurred (as shown previously) in a kitchen island.• Elevated bar-hole readings were found 150 feet away.• Vapor extraction continued for six weeks prior to ordering an

extensive geotechnical program to determine why standard vaporextraction techniques had not eliminated the remaining vapors.


Minimal explosion damage- Gas entered the pipe chasetwo weeks after the gas leak stopped



Queen Creek, Arizona

• The pipeline was installed at a 3 foot depth and numerous vaporextraction wells were drilled to 5-6 feet.

• The geotechnical firm drilled 23 borings in the area to design asolution. They determined detailed soil parameters and measuredthe LEL in the soil that was being extracted. The results wereremarkable.

• Six weeks after the leak had been stopped, elevated LEL readingswere being recorded at depths up to 35 feet.



Queen Creek, AZ


An Approximate Flow Net Diagram of the Arizona Site

Stream lines : RedIsobars perpendicular to stream lines

Top View of Flow Net Diagram



• The soil borings identified a less porous layer of soil at a 5-6foot depth.

• The geotechnical company concluded that the reason for theextent of migration and the depth of migration was the lesspermeable layer, but the less permeable layer was below thepipeline.

• The pattern and extent of lateral migration in this case wassimilar to other cases.



• No evidence the pipeline influenced the migration pattern

• No evidence the asphalt influenced the migration pattern

• No utilities at depths where the gas was detected

• Migration pattern is similar to Portales, New Mexico so, no evidence that soil conditions influenced migration patterns



• Natural gas can be forced by gas pressure to levels well belowthe elevation of the pipeline.

• Bar-hole testing only measures to a 3 foot depth, so it is notcorrect to assume that gas cannot be pressured downward.

• We can conclude that gas pressure from a leak forces gas in agenerally hemispherical pattern deep into the ground fromthe leak.


What Gas Flow Rate is Necessary to Cause an Explosion?


What Flow Rate is Necessary to Cause an Explosion?

First, we need to establish a benchmark of gas flow rate into a structure to cause anexplosion.• Annandale, VA- NBS stated 720 CFH was needed to cause that explosion. This

was calculated by using 10% of the volume of the home with an infiltration rateof one air change per hour.

• San Juan, Puerto Rico- NTSB recorded an in situ flow rate of 102 CFH adjacent tothe structure that exploded.

• IAAI study-House Full of Gas- used a value of over 400 CFH

This is just to have a comparative number, not to establish a minimum figure for anyexplosion.


What Flow Rate is Necessary to Cause an Explosion?

Infiltration rate in a structure is a very important factor to thedetermination of a necessary gas flow rate into any structure tocause an explosion.

This subject matter is beyond the scope of this presentation.

The above values are independent of any specific incident exceptPuerto Rico.


How Does Soil Type Affect Soil Gas Migration?



• Soil type must be a factor as to how far gas would travel tocause an explosion, right?

• Soil porosity is the percentage of pore space in soil.

• Permeability is the ability of a fluid to move through soil.

• More porous soils are usually more permeable if the pores areconnected and allow flow through the soil.

• Henry Darcy empirically derived a set of (basically) constantvalues to relate flow through various types of soils.


Michigan Environmental Education CurriculumGroundwater Supply

PermeabilityJust as the porosity of a soil affects how much water it can hold, it also affects how quickly water can flow through the soil.The ability of water to flow through a soil is referred to as the soil's permeability. As you can probably guess, the permeability of gravel is higher

than that of clay. But can you guess how much higher?Click the button below to see how long it might take water to travel 1 meter in different soil types...



• If gas migration to cause an explosion was related to soil type,then:

• There should be dramatic differences in the distance from agas leak to an explosion and/or

• In the frequency of incidents in clay soil compared to othersoils.

• What I have found is that standard permeability equations arenot applicable to soil gas migration explosions.


Chart of Darcy Constant for Various Soil Types(From Zimmerman)


If a 100 cfh flow rate can cause an explosion, how many square feet of surface area would be required to supply

that flow rate?


How many square feet would be required to cause an explosion at 50 feet and 50 psi?


How many square feet would be required to cause an explosion at 50 feet and 0.5 psi?



So, the permeability factor between gravel and clay is a billion,but neither the frequency of incidents nor the distance that thegas migrates appears to have any significant relationship to thepermeability of the soil.



Let’s Summarize• Queen Creek, AZ Clayey sand and sandy silt 50 feet to explosion

• Jackson, MS Dense wet clay 70 feet to explosion

• Odessa, WA Rock and volcanic ash 60 feet to explosion

• Saratoga Springs, UT Frozen silty clay 40 feet to explosion

• Ft. Worth, TX Dense clay 40 feet to explosion

• Rancho Cordova, CA Silty clay loam 50 feet to explosion

• Union, NJ Gravely sandy loam 50 feet to explosion

• Bowie, MD Not reported 120 feet to explosion


Intermission



How Does Pipeline Pressure Affect the Migration of Natural Gas through Soil?

How Does Pipeline Pressure Affect the Migration of Natural Gas?

Pressure in the gas pipeline

• The pressure range in this study is between 0.25 psi to 60 psi.

• In Wilmington, DE on July 2, 2003 a gas line at 0.25 psi wassnagged by a contractor and created a line separation directlyinside the adjacent building.

• So, with no soil gas migration involved, an explosion occurredwithin 15 minutes.



• In Philadelphia, PA on May 11, 1979, a cast iron gas lineoperating at 0.25 psi cracked and caused a gas leak.

• An explosion occurred 20-30 feet from the gas leak about 30minutes later.



An incident at 0.25 psi occurred in San Juan, Puerto Rico on November 21,1996. The gas was a propane/air mixture in this case.

A crack leaking at an in situ flow rate of 102 CFH migrated into a structure 15feet away and caused an explosion in which 33 people died.

If gas is forced through the ground under pressure, then the specific gravityof the gas would have little to do with whether it travels up or down.



A review of soil gas migration explosions from pipelines atpressures between 0.25-60 psi, reveals no relationship betweenpipeline pressure and the distance gas migrates to cause anexplosion.


How Does Pipeline Pressure Affect the Migration of Natural Gas through Soil?

• Philadelphia, PA 5/11/1979 0.25 psi and 20-30 feet• Williamsport, PA 1/25/1977 10 psi and 70 feet• Pittsburgh, PA 1/17/2003 18 psi and 100 feet• Bowie, MD 6/23/1973 20 psi and 120 feet• Annandale, VA 3/24/1972 22 psi and 240 feet• Portales, NM 5/23/1982 25 psi and 60 feet• Odessa, WA 12/26/2008 40 psi and 60 feet• Jackson, MS 12/24/2008 45 psi and 70 feet• Rancho Cordova, CA 12/24/2008 55 psi and 45 feet



So, pipeline pressure is not a primary factor in the distance thatgas could flow through soil to cause an explosion.

We have looked at the most likely significant factors:

• Soil type

• Pipeline pressure

• No apparent relationship between pipeline pressure anddistance to an explosion.


What Might the Incidents over 100 feet from a Leak Have in Common?

• Although there are few incidents that are over 100 feet, what might they have in common?

• I first noticed on the case in Bowie, MD. that the migration pattern was along a drainage path towards a creek.


Bowie, MD again. Migration pattern is:172 feet uphill318 feet downhill145 feet acrossThe house that exploded was 120 feetuphill and the closest house was 72 feetfrom the leak.

The elliptical migration pattern is notrelated to the pipeline or the asphalt, butmight be related to drainage pattern.

Elevated soil gas readings were recordeda year after the explosion.



Leak Location

Annandale, VA

• Gas traveled 240 feet to cause this explosion

• NTSB concluded that the gas followed utility trench with rocky material.

• I personally examined the site in June 2018. I saw a distinct drainage pattern between the two houses that blew up.

• The cul-de-sac is the lowest point of the surrounding area and natural drainage is between the two houses that blew up.

• After examination of this case and others, the natural drainage path was the most porous route for the gas.


Annandale, VA


Annandale, VA

Let’s examine the Annandale case:• Backfill material in the water and sewer line trenches was rocky material, so the

soil had maximum permeability. The gas line ditch was a less permeablebecause select fill material was used with six inches of soil on top.

• Bar holes were about 3 feet and around the utility trenches.• The three houses that burned were closest to the leak and the trenches.• Gas was smelled in the house within 30 minutes of the fracture.• The explosion occurred within 60 minutes of the contractor-induced fracture.• NBS noted that tracer gas penetrated sealed block walls in the basement.• Tracer gas arrived at the house farthest from the leak within 30 minutes.• The NTSB concluded that the gas did not follow the gas line but followed the

rocky backfill in the adjacent ditch lines.


Annandale, VA


Leak Location

4909 Magdelene Ct

4911 Magdelene Ct

Annandale, VA


Annandale, VA


4909 Magdelene Ct

Annandale, VA


Annandale, Va


Annandale, VA


Steep Drainage Channel

Annandale, VA

• NBS conducted tracer gas tests.

• Drilled shallow test holes mainly along utility trenches.

• Found no elevated gas readings at 4908 and 4910.

• Used tracer gas mixed with compressed air at about 2200 CFH

• Obtained a “cavity pressure” of about 3 psi.

• The first tracer gas reading at the foundation was taken 30 minutes after startup and tracer gas was already there.

• They noted specifically that there was no gas along the gas line, and no gas in sewer line.


Annandale, VA

• NBS used an air compressor and refrigerant as a tracer gas

• Detected tracer gas within 30 minutes at the foundation wall 240 feet away.

• Pumped air/tracer gas at a rate of 2000 CFH

• If they pumped too much, it lifted up the asphalt

• No gas readings on North side of street


Annandale, VA

• It appears that most of the gas followed the natural drainage

• Steep natural drainage path between the two buildings that blew up

• Utility trenches were all rock. Soil samples only to the bottom of the trenches.

• Investigators followed only the expected outcome.


Rhine Street, Pittsburgh, PA

• Cracked 12 inch diameter cast iron line at 18 psi

• Leak reported at 2151 Rhine Street

• Looked for leak for 7 1/2 hours before the apartment building across the street at 2146 Rhine Street blew up

• Nobody was looking there, but gas was smelled and no tenants remained in the structure

• Flow test after valve was removed showed flow rate of 20,000 CFH in open air

• Temperature was 0-5 degrees F at the time of incident


Sites between 40-100 feet

• Odessa Washington

• Jackson, Mississippi

• Holly Springs, Mississippi


Odessa, WashingtonNo gas line to structure- 60 feet and 40 psi


Home between steep slopes and drainage paths on both sides

Jackson, MS


Holly Springs, MS


Common Features of Long Distance SGM

• Distinct natural drainage features

– Bowie, MD- 120 feet

– Annandale, VA- 240 feet

– Pittsburgh, PA- 100 feet

– Odessa, Washington- 60 feet

• Creek behind the house

– Jackson, MS- Downhill from gas line-70 feet

– Holly Springs, MS- Downhill from gas line- 40 feet



How Does the Leak Flow Rate Affect Natural Gas Soil Gas Migration?

How Does Flow Rate of the Gas Affect Natural Gas Soil Gas Migration?

• The biggest problem with evaluating this factor is lack of data. Typically,Immediately after a gas explosion, the gas line is repaired. To measure gas flow insitu, one would need to cut the line some distance away and measure the in situflow rate. That would require a severe disruption of service to other customers at aminimum. It is also extremely impractical and possibly unsafe.

• The case in San Juan, Puerto Rico provided in situ gas migration and explosion dataof 102 cfh.

• A case in Sublett, KS showed an in situ flow rate of 3200 cfh at 9 psi where thedistance to the explosion was about 20 feet. Pipeline pressure was 30 psi, so fullpipeline pressure and, thus, full flow rate were not attainable.



The testing in Nevada provided the only available data on this subject.On all gas leak situations, there is only one flow rate, so we cannottypically test the affects of different flow rates on soil gas migration atactual incident sites even if we measured in situ flow rates.

Let’s start with the testing setup used in order to properly examine thedata retrieved from the testing in Henderson, Nevada 30 years ago.




Note: Readings were recorded every 20 ft. after the closest test port at 8 ft.

How Does Flow Rate of the Gas AffectNatural Gas Soil Gas Migration?




• So, as the injection flow rate increased from 2400 CFH to up to 11,400 CFH, an air flow rate along the pipeline was detected at 68 feet.

• As the injection flow rate increased to 19,200 CFH, however, the flow out of all the ports along the pipeline stopped and, a large ground fracture occurred about 9 feet perpendicular to the pipeline.

• This is the only test of which I am aware where actual testing occurred to attemptto document gas flow along an “annular space”.

• While a tracer gas was detected 88 feet from the source and a gas flow rate wasrecorded as far as 68 feet from the source, neither indicate that a “higher thanexpected flow rate” of gas existed along the annular space of the pipeline.


Summarizing some of the results, the following conclusions can be reached:

• A very small percentage of the total flow was ever recovered from thetheoretical “annular space”.

• An input flow rate 40 times greater than the initial flow rate still produceda very small percentage of the total flow along the pipeline.

• Much less than 1% of the injected gas flow rate was ever recovered fromthe pipeline annulus and never farther than 68 feet regardless of theinjection flow rate.



• As the flow rate into the injection point increased from 300CFH to as much as 19,200 CFH, the flow rate out of the testpoints increased only marginally to a distance of 68 feet.

• However, as the flow rate increased to close to 19,200 CFH,the distance that flow was detected, then, collapsed.

• As I reviewed all the data, I saw clear evidence of the groundfracturing particularly as the flow rate increased.

• Fracturing was more noticeable closer to the injection point.



So, as the flow rate (of a leak) increased, neither the flow rate out ofthe test points nor the concentration of the tracer gas showedevidence of gas following the pipeline or any pattern other than thegas is simply pressured through the soil in a spherical pattern.

The gas pressure drops off rapidly as the distance from the injectionpoint increases, but some pressure did exist in the soil for at least 68feet from the injection point and, in other cases, as much as 150feet, but it appears that insufficient pressure exists past about 50feet from a leak to cause an explosion.


How Long Does it Take for Gas to Migrate to Cause an Explosion?



Let’s look at some cases where we do know the time factor.

• Annandale, VA- 240 feet in 30 minutes at 22 psi

• Philadelphia, PA- 30 feet in 15 minutes at 0.25 psi

• Chicago Heights, IL- 40 feet in 30 minutes at 28 psi



So, in cases where we know when the leak started because acontractor hit the line, the time from the initiation of the leakto the time of the explosion can be remarkably short.


Summary

• Gas from a leak can travel 50 feet or more to cause an explosion ina broad range of known and unknown conditions.

• Gas can be forced by pressure downward at least as far a 35 feet.• The lateral distance gas flows to cause a fire cannot be related to

soil type.• The distance gas flows to cause a fire is not proportional to

pipeline pressure.• This data shows that none of these conditions are particularly

unusual in a broad range of known or unknown conditions.• So, how do we explain these facts that all seem so unusual?


What Might These Cases Have in Common?

During the testing program in Henderson, Nevada in 1988, several interesting facts became apparent:

1. Gas flow out of the ground was measured as far as 68 feet from the injection point.

2. Flow into the ground was 11,400 CFH but the flow out at one location was only 0.34 CFH, so only 0.003% went along the pipeline and 99.997% migrated somewhere other than along the pipeline…So, there was no annular space.


What Might These Cases Have in Common?

3. When the injection point was dug up, a 1-foot diameter void had beencreated all around the injection orifice.

4. The soil that was displaced from around the injection point must havebeen displaced somewhere. Where did it go?

5. The gas must have opened up fractures in the soil and some of the dusthad been forced down the widened channels. This physically displacedthe soil particles that were originally around the injection orifice.

6. If the gas pressure or flow rate separates soil particles and allows gasflow through these micro-channels, then gas pressure can extend muchfarther from the leak than if fracturing did not occur.


What Does This All Mean?

• If the gas from a gas leak radiates from the leak in a spherical pattern and the leakrate is fixed, there is a limited distance that the gas is concentrated enough orunder enough pressure to cause an explosion.

• From existing data, gas explosions within 50 feet of a leak are not exceptional.• Numerous flow paths or no specific flow path at all could explain explosions

within this distance.• There is no relationship between the distance from a leak to an explosion and the

possible affect of other utilities.• There are quite a number of factors that affect that distance, but the research here

has revealed that natural gas explosions at a distance of over 100 feet areextremely rare.

• The data from the testing in Henderson, Nevada is consistent with theindependently observed data collected from numerous NTSB reports.


What Does This All Mean?

Since the gas is forced through the soil by pressure, “caps” suchas asphalt or snow or rain or ice are extremely ineffective.Additionally, there was no documented case where any of these“caps” extended the gas migration farther than would normallybe expected anyway.


How Does an Explosion Occur?

If the gas from a leak spreads out so uniformly, how does anexplosion take place?

If the gas does not follow the gas line or get trapped by ice orasphalt, and migrates uphill and downhill, how is there enoughgas concentration and flow to overcome room infiltration tocause a house fire or explosion?

Maybe the following diagrams will help envision how this mayoccur.


How Does an Explosion Occur?

So, after 30 years of review and analysis, comparing numerousfactors, and examining actual data from personal testing and othersources, I have concluded that house explosions caused by gasmigration occur as a result of gas being forced under pressure toform fissures in the soil. This condition allows gas to migrate muchfarther than “expected”.

The marginal gas pressure remaining at 50-100 feet, for example, isvery small but is enough to force the gas through pipe chases, seamsor other voids that exist in all structures at a sufficient cumulativeflow rate to cause an explosion or fire.


How Do Sewer Lines Affect Gas Migration Explosions?


How Do Sewer Lines Affect Gas Migration Explosions?

• So, as the gas spreads out from the leak, the pressure decreasesrapidly from the leak source, but low pressure extends longdistances. The Pepcon testing clearly shows this fact.

• So, unless the sewer line is within a foot or two of the gas leakitself, there is very little pressure or flow at any point in the “cloud”to drive enough gas into a vented sewer line to cause an explosion.

• Bar hole tests confirm that there is very little pressure in the soileven short distances from the leak. If this were not he case, gaswould be blowing out of bar holes relatively close to the leak.


Suggestions for Future Investigations

The testing that was done in 1988 raised many questions that have been very difficult to address since that time.

Assembling this basic information has been extremely laborious. It should not bethat way. To refine or even refute my conclusions, basic data should be routinelyacquired such as:• Pipeline pressure• Distance from the gas leak to the structure• Closest structure to the leak• Gas pipeline route• Utility line routes



• Soil conditions (even approximate would help)• Is the site uphill or downhill from the pipeline?• Special conditions- other utility routes• Identify Natural Drainage Channels• In situ flow rate from the pipeline, if possible (very difficult to obtain)• Mode of entry into home-Even if it cannot be determined, state it on a form.

Sometimes entry mode or modes are very clear and that would be helpful.• So, assuming that the gas “followed a pipeline” can be very misleading in the

investigation of these incidents and the above analysis leads me to conclude that it is usually, simply, erroneous.



• Possibly, in the future with better data, this can all be refinedand/or corrected. However, if we keep perpetuating the samemisinformation, we will never be able to improve ourinvestigations. That is what I hope to correct in the future.


Practical Applications for First Responders

• This study has shown that a majority of explosions occur within a 50 foot radius of an underground gas leak. We can, then, possibly develop safer ways FIRST RESPONDERS can investigate gas leaks.

• For example,– If the leak location is known (contractor hit), then consider evacuating

within a 50 foot radius. If that is done, the chances of injuries from an explosion could be significantly reduced.

– If the leak location is not known, park emergency vehicles at least 100 feet from the reported leak. This will reduce ignition sources until better information can be obtained.


Practical Applications for First Responders

– If sewer manhole covers are readily accessible, check the sewer for gas with the CGI to make a quick and easy assessment of any spread of an underground gas leak in the area by this means. If there is gas in the storm or sanitary sewer line, a much broader approach to the incident is needed.

– Focus on protecting people and equipment within a 50 foot radius first.

– Look for Drainage Channels and extend the range, if necessary.


Practical Applications of First Responders

• A 50 feet radius from the leak is a potential danger zone. Gas farther than 50 feet from the leak usually does not ignite.

• If critical choices need to be made, concentrate on a 50 foot radius.

• If the exact leak location is not known, a 100 foot radius from the caller might be prudent.


Practical Applications for First RespondersUnknown Leak Location from Call In


Risk Assessment

• Structures farther than 50 feet from a pipeline are much less likely to be affected by a large gas leak.

• If a gas line is on one side of the street, the houses on the other side (more than 50 feet away) are much less likely to be affected by even a large gas leak.

• If a large gas leak is in a drainage path, look for a possible risk to building farther than 100 feet in that drainage path.


Soil Scrub of Odorant

• This review has shown that it is very likely that gas travels very fast from gas pipelines to cause an explosion.

• Odorant scrub is caused by the adsorption of odorant molecules onto the surface area of dry soil particles.

• Odorant scrub is, therefore, extremely “surface area” dependent.

• This theoretical mechanism for soil scrub is, then, vastly reduced if the soil fractures to form numerous channels. People smell gas from an underground gas leak frequently. This is, at least, anecdotal evidence that soil scrub of odorants is very inefficient in a gas explosion scenario. Simply accepting the premise that odorant is scrubbed from the gas may be an incorrect assumption.


Conclusions

• Traditional ways of conducting soil gas migration investigations need to be reassessed.

• Most incidents occur within 50 feet of even a large leak.

• Incidents farther than 100 feet can occur along drainage channels uphill or downhill from the gas line.

• Pipeline pressure and soil type are not major influences to gas migration.

• Pipeline and utility routes may not be as significant as once thought.


Conclusions

• Gas is forced under pressure and through fissures in the soil to great depths and long distances.

• The distance from a leak to a building explosion is limited by the flow rate of the gas.

– Too small of a flow rate does not produce sufficient cumulative flow and too great of a flow rate creates a blow out condition.

– The maximum flow rate might be about 19,000-20,000 CFH

– The minimum flow rate has not been determined.


Conclusions

• This information is useful to first responders.

• This information is useful for risk management.


Thank You

Thank you for your time and attention. There has been a greatdeal of information shared and extremely varied concepts havebeen presented. Hopefully, I have simplified them andcategorized them in a relatively concise and understandablemanner. If you have any questions or comments, please feel freeto contact me at any time.


Bison Engineering, Inc.


a review of soil gas migration in natural gas...

Documents