The problem : On the occasion of the boring of a sewage tube (diameter 3.5 m) in Aubervilliers, close to Paris, an area which had been contaminated with liquid kerosene had to be passed through for about 300 metres. An extensive decontamination programme had been completed; however some traces of free kerosene were suspected in the ground and confirmed by soil sampling and analysis. The TBM is a compressed air full face road header with a working pressure of 1.9 10⁵ Pa (0.9 b₍g)). To cope with the possible consequences from kerosene in the ground excavated by the TBM, a risk evaluation was awarded to JCLP HYPERBARIE.

2 - possible contamination of the tunnel atmosphere by kerosene from the extracted ground,

3 - eventual contamination of the environment via compressed air leaking through surface ground.

Two types of risks had to be evaluated : Fire / Explosion risks and toxicity of potentially contaminated atmospheres.

Risk evaluation : From the data of soil analysis kerosene could be evaluated and the contaminated zones delineated. From kerosene vapour partial pressures at the mean soil temperature, the possible concentration of kerosene in the various atmospheres concerned was calculated for comparison both with the explosive limits and the toxic threshold values. Conclusions of the report demonstrated that there was no risks of fire / explosions at any of the places even when welding in compressed air; however the same report showed that the toxic thresholds could be reached in the worst case situations.

Safety precaution : Specific safety rules have been established and implemented : monitoring of kerosene concentration in the tunnel, in the compressed air zone; information for the population concerning possible odours in the boring TBM zone; mobilization of a mobile team ready for environmental analysis of kerosene vapour. Specifically for hyperbaric work, the known air working chamber ventilation from leaks into the ground was estimated sufficient to prevent any kerosene vapour to reach the toxic threshold value. Therefore should hyperbaric manned intervention be required during the crossing of the contaminated ground, then the only specific safety rule would be kerosene partial pressure monitoring with alarm in the compressed air environment.

Results : Kerosene monitoring showed the expected contamination values in the tunnel and in the compressed air zone. No odours where detected in the environment nor were traces of kerosene. Routine maintenance manned interventions in compressed air could have been performed without restrictions during the excavation of the contaminated ground.

Introduction

During the construction of a very large sewer network to collect all storm water in the Paris area, several tunnels are being bored with Tunnelling Boring Machines (TBM). One of these tunnels in the town of Aubervilliers (3 km from Paris) had to cross through a zone known to have been previously contaminated with hydrocarbons which had leaked from various hydrocarbons storage tanks in an old metal treatment factory.

Upon moving away from this site that company had the legal duty to clean up the underground. This was done by pumping out all liquid phase (5 tons !). The major contaminant had been identified as kerosene. Despite the underlying ground restoration operation some kerosene was still adsorbed in the soil layer which was to be bored through by the TBM.

An evaluation of the risks associated with the presence of this contaminant has been carried out in preparation for the passage of the TBM. It was of major importance that the possible necessary actions could be decided early enough not to stop the boring operation by unexpected dangerous situations.

Levels of hydrocarbons contained in the ground as determined by soil investigation after the restoration and cleaning showed a maximum concentration of total hydrocarbons of 3750 mg/kg of ground in the most contaminated zone which has been controlled. The value selected for calculations supporting the risk evaluation had been 4000 mg/kg. Those contaminants did not contain volatile hydrocarbons, nor aromatics and could be classified as kerosene.

The TBM and boring technique

The TBM is a CSM BESSAC machine. It is a road header machine fitted with a bulkhead to form a compressed air chamber to support the ground and hold back the water. The cutter boom operator is located at atmospheric pressure and he can see the ground face and the excavator machine through an acrylic view port. The excavated ground is extracted from the cutting chamber, backward into the tunnel, via a screw decompressing the ground directly into the muck skips for removal.

Fig 1 - Principle of CSM Bessac TBM Fig 2 - Hyperbaric cutting of an anchoring line

The compressed air bubble is maintained at the pressure which is just sufficient to remove all water from the cutting chamber. The selected pressure is automatically adjusted by high flow regulators under close monitoring from the cutter arm operator (figure 1). To prevent large leaks of air from the cutting chamber into the ground, a special foam is thrown onto the free face of the ground with a directional foam jet. The foam is absorbed by the ground and thanks to fast polymerisation reduces the ground gas permeability.

Risk evaluation

The possible dangerous consequences of boring through the contaminated soil had to be evaluated in all areas which could be contaminated by kerosene vapour. Risks eventually associated with the kerosene vapour from the ground are related either to fire or toxicity, both conditions need to be evaluated carefully.

In the TBM

The ground may lead to evaporation of kerosene in two different places : in the cutting chamber at pressure (mark A in figure 3) and in the tunnel at atmospheric pressure during transportation of the excavated material towards the shaft (mark B in figure 3). Both situations need to be considered.

In the surface environment

Some compressed air is lost in the ground, although partly controlled by the foam system. It will be released freely in the environment. Going through the ground, this air will evaporate kerosene and may be contaminated when it comes out of the ground at surface in the adjacent streets (mark C on graph 3).

Properties of kerosene (Cas n° 808-20-6)

SOURCE	NIOSH [3]	I C S C [3]	FFTS [4]	INRS [5]	MSDS [6]
Hydrocarbon	C₉ to C₁₆		C₁₀ to C₁₈	C₉ to C₁₆
Exposure limit	100 mg/m³	N E	5 mg/m³ (Oil mist)	1500 mg/m³	N E
Vapour pressure	5 mmHg (37°C)		> 8 mmHg (38 °C)		1 mmHg (20°C)
Molecular weight (mean)		130		> 145	170
Vapour density (air = 1)		4.5		> 5
Flash point		37-65 °C	> 40 °C	> 38 °C	38 °C
Explosivity limits	0.7 – 5 % in air	0.7 – 5 % in air	1 – 6 %	0.5 – 6 %	0.7 – 5 %
Boiling point (distillation)			205-290 °C	140-300 °C	175-325 °C

For the appreciation of atmospheric contamination the most important data is the pressure of saturated kerosene vapour at 20 ° C. This vapour pressure, shown in table 4, will be the maximum partial pressure of kerosene vapour at 20°C, when free evaporation takes place in a confined space and when a steady state situation is reached. This partial pressure will be the maximum pollutant content of an atmosphere, even without any ventilation, at any pressure at the temperature of reference.

The mean partial pressure of kerosene vapour selected for the risk evaluation has been 1 mmHg (MSDS value) or 1.33 hPa. Temperature in the ground is close to 11 °C, therefore evaporation in the ground is much lower than at 20°C, however for the risk evaluation a rough estimate is acceptable provided the values are maximised. Since the relationship between partial pressure and temperature was not available, the 20°C value has been used in all cases (worst possible situation).

Kerosene being a mixture of various hydrocarbons from C₉ to C₁₆, without any indications of the ratios for various components, the mean molecular weight chosen for calculation is : M = 182 which corresponds to the molecular weight of C₁₃.

The Threshold Limit Value (TLV) selected for short exposure to kerosene vapour in a breathable atmosphere has been 1500 mg/m³, because INRS (Institut National de Recherche et de Sécurité) produces a large body of evidence to support the published value. This value must be calculated in partial pressure for use at any pressure: it .can also be expressed in ppm (using the mean Molecular weight 182 g, and a molecular volume of 24 litres for 20 °C) result is 197 ppm. This ppm value is a Fraction which then needs to be translated into partial pressure (hPa). Since this value is defined for 1 bar absolute, 197 ppm correspond to 197 10^-6 bar or 0.197 hPa maximum kerosene partial pressure (TLV in partial pressure – hPa-).

From the direct comparison of the saturated kerosene vapour partial pressure resulting from evaporation as described above (1.33 hPa) and the TLV (0.197 hPa) it can be concluded that an atmosphere saturated with kerosene at about 20°C would be beyond (7 times more) the limit of acceptability at any pressure.

When an atmosphere is in contact with the liquid phase of a volatile organic compound (VOC), without ventilation, after a while the level of VOC will reach its saturated vapour pressure value at that temperature. Note: Kerosene is not a "VOC" because the partial pressure at 20 °C is very low, however the vapour behaviour model is equivalent).

However when some rate of ventilation exists, after a while a different steady state level is reached. The resulting partial pressure of the chemical is equal to the ratio of the off-gassing rate (l/min.) of that chemical divided by the ventilation rate in the same unit, provided that the ventilation is efficient and that complete mixing takes place in the ventilated volume[1].

Using the worst possible case situation, it can be assumed that during decompression of the ground in the screw, the small volume of air included in the ground will be released into the tunnel and that it could be saturated with kerosene vapour at the pressure of the cutting chamber (between 0.5 and 1 bar_(g) in this case). Upon delivery in the tunnel this air is returned to atmospheric pressure and the resulting partial pressures of components are reduced by 1.5 or 2. It has been concluded that the air delivered at the extremity of the screw at atmospheric pressure cannot contain more that 1.33/1.5 hPa of kerosene or 0.88 hPa at atmospheric pressure. This small gas volume, even with a minimal dilution by the ventilation of the tunnel will not lead to any risk of toxicity for the workers in the tunnel (a few litres/sec against at least 1 m³/sec).

Should a need for a hyperbaric inspection or repair arise during the boring in the contaminated soil, the partial pressure of kerosene in the cutting chamber atmosphere is reduced as explained above by the high rate of ventilation of the cutting chamber to match the air leaks in the ground through the face. In addition, before excavation, traces of kerosene remaining in the ground are ventilated away by air leaks in situ. The mean flow considered for the risk evaluation is 60 m³/min. To reach a toxic partial pressure of kerosene in the cutting chamber it would be necessary to evaporate 0.118 m³/min of pure vapour of kerosene…or 118 x 182/24 = 89.5 g of kerosene per minute…. This was estimated as very unrealistic when one takes into account the possibility of kerosene adsorbed onto the bulk ground to evaporate into the ambient atmosphere only from the free surface of ground sitting in the cutting chamber (mark B on figure 3). It was concluded that kerosene cannot be a toxic risk for compressed air workers in that situation.

Air escaping from the ground in the environment can be assumed to be saturated with kerosene at a pressure close to atmospheric pressure (worst possible case) potential toxicity reduction effect of decompression is not taken into consideration because very little information is available concerning the exact pressure of evaporation…. In that case air escaping freely from the ground will contain anyhow a maximum partial pressure of kerosene 7 times more than the TLV.

When that air is reaching the surface in open air, immediate dilution will reduce the toxicity far below the TLV.

When that air is released in a semi-confined zone, like cellars below buildings it may accumulate and create a “toxic zone” until dilution will reduce the potential danger.

In addition the odour threshold of kerosene is below the TLV. Which means that it can be detected by human nose at partial pressures lower than TLV.

Some action must be taken to protect population in non ventilated areas and to make sure every one concerned is informed that some smell may be noticed, however it would not be dangerous for their health, however it must be properly reported.

The excavated ground from this zone is considered contaminated and is dumped under controlled conditions in a specially selected discharge plant.

Temperature of self-inflammation of kerosene in air is very high : above 220 °C and Flash point in air is above 38 °C

This means that an atmosphere saturated at 20°C with kerosene cannot flash spontaneously even with a spark.

The Lower Limit of Explosivity is 0.5 to 0.7 % corresponds to 5000 to 7000 ppm at atmospheric pressure. We have shown that at 20 °C this concentration would be higher than the kerosene saturated vapour pressure (1.33 hPa) : kerosene cannot stay in gas phase at concentration of 5000 ppm at 20°C, it will either not evaporate enough or would condense.

In the tunnel and in the environment, conditions of Lower Explosive Limit can never be attained.

While working under pressure one should consider the effect of the increased partial pressure of oxygen on flammability of hydrocarbons. It has been clearly established by Cleuet et al. (1994) [2] that in the range of pressure concerned (0 to 10 bar₍g)), in compressed air, the explosive limits of hydrocarbons (expressed in Fraction -ppm or %-) are not significantly modified from those measured at atmospheric pressure.

Therefore is can be considered that also in compressed air, in the conditions of tunnelling described, there is no risk of explosion even in case of open flame exposure.

Safety precautions

In order to cope with any unexpected condition like a high concentration of kerosene in the ground which would have gone undetected during the ground control campaign or an insufficient dilution in the cutting chamber while hyperbaric work would be in progress… several extra safety precautions have been implemented before crossing through the suspected zone.

Controls for kerosene contamination levels have been organised with two independent systems:

Samples of the extracted ground were collected in the shaft, at regular intervals, each sample was quickly controlled on total hydrocarbons content to monitor the penetration of the TBM into the contaminated area.

Gas sampling with continuous monitoring (DRAEGER type Polytron IR Ex) calibrated with air saturated with kerosene at the temperature of the tunnel (1.33 hPa or close to) with an audible and visual alarm point set at 30 % of this saturated value.

There were 2 sampling lines, manually selected: one sample could be pumped from near by the delivery point of the screw, the other one was drawn from the working chamber to monitor the atmosphere in case of manned hyperbaric intervention.

Information

An information bulletin was distributed to the near-by population and displayed in the buildings around the zone to be excavated by the TBM. This bulletin explained how some kerosene vapour could be smelled when the machine will pass below but that there is no specific danger. A phone number for help was also indicated in case of smell in confined areas.

Information was also discussed with the fire brigade in a special meeting where all details of the study were explained (and accepted).

Operators

All tunnel workers have been informed on risk evaluation and safety measures. The team leader of each shift underwent the necessary training concerning the gas monitoring system and actions to be taken in case the alarm would be triggered.

A portable total hydrocarbon analyser calibrated in the same manner was made available to a mobile team, ready to go on sites in case an information would be received from the population. A small number of exposed cellars were selected for actual controls during the passage of the machine.

Results

When the machine reached the contaminated zone soil analysis confirmed the presence of kerosene and odour of kerosene was acknowledged in the tunnel. The gas detector showed not more than 1/3 of the saturated maximum value on the gas close to the decompressed ground.

No neighbour ever complained or called the special phone number, no traces of kerosene vapour were ever detected in the streets or cellars.

Few routine hyperbaric interventions for maintenance of the cutter arm were carried out, no personnel have reported kerosene odour during those interventions. Should a repair requiring welding or arc-air burning had become necessary, it would have been carried based on the evidence that there was no risk of fire whatsoever.

Conclusions

Facing a situation which had not been met in previous excavations with Bessac type TBMs, it was decided to thoroughly evaluate the risks, in the worst possible conditions. Based on the information obtained from cleaning campaigns and on the pollutant physico-chemical data it was possible to demonstrate that the risks of toxicity were minimal and that the risk fire was non-existent.

For extra safety precautions several control systems were implemented on site and in the tunnel which confirmed that the evaporation of kerosene was very mild and has never reached values of concern.

This approach of atmospheric contamination by vapours, based on maximum partial pressure of saturated vapour can be used for organic compounds even when their volatility is very low.

Incidentally, the day following the relieving of safety precautions in relation with kerosene, a burst of Hydrogen Sulphide (H₂S) was detected (from smell by the operator) which led to tunnel evacuation until an extra ventilation system could be installed (3 hours). The contaminated ground zone was probably only a few meters long. All detectable hydrogen sulphide was gone within a few hours of further boring.

The training and preparation of the crew to cope with an expected atmospheric contamination in the tunnel proved to have been useful since an early alert was given for an unexpected one.

1) CHOUTEAU J., BIANCO V., ORIOL P. et all., Expérimentation animale et humaine de vie prolongée sous pression en atmosphère oxygène-hélium. Technologie et résultats biologiques. 1967, Annales de l'Anesthésiologie Française, T VIII, nE spécial 1, 1-45.

2) CLEUET A. and P. GROS (Mise à jour 1994 par J. M PETIT) – 1994 Les mélanges explosifs, Gaz, Vapeurs, Poussières, Liquides, Solides – INRS éditeur – ED 335. Paris

3) NIOSH : Pocket guide to chemical hazards. Translated by International Safety Card. Programme International sur la sécurité des substances chimiques – Commission Européenne 1993

4) FTSS – Centre Canadien d'hygiène et de sécurité au travail. Fiche FTSS 1094432 – Kerosene – 2000-3

5) INRS Fiche toxicologique n° 140 Pétroles Lampants. INRS 40 rue Noyer 75014 – Paris FT 140 pp 4 – 1996

6) MSDS – material safety data sheet : http://www.jtbaker.com:80/mdss/k2175.ht from Mallinckrodt Baker, Inc. 222 Red School Lane, Phillipsburg, NJ 08865 USA.

Acknowledgement : SIAAP and Montcocol SA have authorised the publication of those results for benefit of the tunnelling community.

J-C. Le Péchon