Résumé

S.4.6 Ligthning prevention in a SEVESO installation

Métriques

19
4
810.47 Ko
 application/pdf
bitcache://ff7093413f8ea04e252229de17f31cb5d90b03db

Licence

Creative Commons Aucune (Tous droits réservés)

Sponsors

Co-organisateurs

ilpa_logo.png
<resource  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                xmlns="http://datacite.org/schema/kernel-4"
                xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
        <identifier identifierType="DOI">10.23723/9740/9957</identifier><creators><creator><creatorName>Robert Moutier</creatorName></creator><creator><creatorName>Ted Pierini</creatorName></creator></creators><titles>
            <title>S.4.6 Ligthning prevention in a SEVESO installation</title></titles>
        <publisher>SEE</publisher>
        <publicationYear>2014</publicationYear>
        <resourceType resourceTypeGeneral="Text">Text</resourceType><dates>
	    <date dateType="Created">Wed 26 Feb 2014</date>
	    <date dateType="Updated">Tue 13 Jun 2017</date>
            <date dateType="Submitted">Fri 20 Apr 2018</date>
	</dates>
        <alternateIdentifiers>
	    <alternateIdentifier alternateIdentifierType="bitstream">ff7093413f8ea04e252229de17f31cb5d90b03db</alternateIdentifier>
	</alternateIdentifiers>
        <formats>
	    <format>application/pdf</format>
	</formats>
	<version>32837</version>
        <descriptions>
            <description descriptionType="Abstract"></description>
        </descriptions>
    </resource>
.

Ligthning prevention in a SEVESO installation ANTARGAZ’s GPL storage facilities Robert MOUTIER TelComTec TAVERNY, France Ted PIERINI ANTARGAZ PARIS LA DEFENSE, France Abstract—Prevention in A GPL storage facility is made with a lightning local detector. This prevention stops dangerous process. The detector is an electric field meal giving three levels of alarm and lightning strike events “field step”. In one year of exploitation we have made statistics of exploitation an comparison with meteorological alerts. Keywords—ligthning prevention, electric field meal I. INTRODUCTION ANTARGAZ has developed a lightning prevention with a local storm detector for its Liquefied Petroleum Gas (LPG) storage facilities. Fig.1 LPG storages II. PREVENTION APPROACH A. Risk presentation The goal of a LPG storage facility is to stock and distribute LPG by tank trucks or by bottles. Storage of LPG has no permanent hazardous area. But loading and unloading of trucks, wagons... and LPG manipulation present such risks. LPG can explode under certain conditions when it is vapor and because it can evaporate under atmospheric pressure. We can easily understand how big the lightning risk is if the LPG is in open air. The most effective way to eliminate this risk is to forbid any dangerous operation in case of instant lightning. For the incident inventoried by ANTARGAZ on its LPG storage facilities we have selected three critical situations after a lightning strike. They have been close to cause an explosion (in the separation of a truck loading LPG pipe) an accident on an operator (near an impacted lighting mast). B. Lightning prevention instruction On ANTARGAZ’s GPL storage facilities for this prevention, an instruction is applied under the responsibility of the operator. In case of stormy presence it is prohibited to: • Purge storage tanks, • Realize operation of loading or unloading wagon or truck, • Remove the loading LGP pipe of a truck, • Make maintenance visits at the top of the tanks, • Carry out work at height. These operations all depend on human operation. C. Instruction application Two mains difficulties result from the application of this instruction: • The subjective personality of the operator and its interpretation of local weather, • The desire to maintain exploitation operations. For a long time, to the first point, the application was under the responsibility of the operator who performed the meteorological situation of the moment. This interpretation could be questionable if only for just does not generate productivity losses, or simply because of a lack of lightning watch. An operator is able to interpret multiple signals and make a good synthesis in meteorology. He is able to identify a storm depending on the nature of the sky, the wind, the thunder. However, human reliability and objectivity are not good. It therefore seems difficult to envisage in advance where and when the first strike falls: off or on the site? Note that on this point the electromagnetic pulses detectors are in the same issue. For the second point, it is clear that a stop pumping causes a pause on the production with productivity losses. When the stop times are too long, they become incompatible with industrial objectives and therefore not applied correctly. The operator would be tempted to delay pauses of operations when loading and unloading trucks or wagons are being or stop filling bottles. But the responsibility of the operator is engaged in the application of such instruction. Any non-application could result in a crash, could be sanctioned. We understand that this application, without being given the necessary means, is not satisfactory from the point of view of safety of people and the environment. D. ANTARGAZ policy The LPG business is based on a logistic programmed in advance, as a LPG storage facility doesn’t work 24 hours a day, so it is quite possible, if necessary for safety, to, delay 30 minutes to an hour a loading operation in the case of occurrence of a lightning risk. The safety of its staff, the truck drivers and its neighborhood are priorities for ANTARGAZ. One can also add compliance with regulations in force and the preservation of its production facilities. For all these reasons, ANTARGAZ decides to put enforcement prevention lightning in place on all of its facilities. III. TRIGGER DEVICE A. Choice of a trigger device To adopt anticipated temporary preventive measures, the trigger cannot be based on the perception of a thunderbolt from a human way (time count perception of lightning and thunder) or electromagnetically. The preventive effect can only be effective if the thunderbolt seen or detected in the near environment of the site is early enough. This may be satisfactory in the case of an approaching storm front, but we cannot rule out the fact that the first lightning strike can take place on the LPG storage facility. That’s why we chose to use the measurement of the electrostatic field on the site with a local storm detector which allows the analysis of lightning perturbations in a given around the site perimeter. Recall that for a site with a risk of explosion using a local storm detector field strength measurement (FSM) class I is recommended by the NF EN 50536 "Protection against lightning - Thunderstorm warning systems." B. STORMDETEC The storm detector measures the electrostatic field via a rotary sensor type electrometer by capacitive bridge. It is commonly called "field mill". On ANTARGAZ facilities, this detector is the STORMDETEC©. Fig.2 Sensor of the STORMDETEC© The local electrostatic field is influenced by the presence of the ground charges, image that are above the load sensor in load cloud position. Since the presence of cloud charges will produce lightning process, the measurement of this particular field is perfectly correlated with the risk of lightning strikes. The establishment of thresholds of perception of the amount of charges in the presence allows us to estimate the lightning strike probability. The crossing of these thresholds can trigger alarms. These alarms are used as decision support to trigger the desired prevention. Because it can detect a lightning strike before its arrival, this sensor meets the preventive expectations we were looking for. Furthermore, it also provides another essential information: the end of the stormy period. However, it is important to note that the detector is unable to know when will intervene cloud to ground discharge or how far is the thundercloud. C. ANTARAGAZ Alarm levels The following levels have been set for an adaptation of dangerous LPG depot operations: • Level 1: threshold 2 kV / m, thunderstorm, informs the birth of thunderstorm activity, be vigilant about evolution, • Level 2: threshold 4 kV / m, close storm, informed of the impending storm, do not start risk operations • Level 3: threshold 8 kV / m, active local storm, inform a high probability of lightning strikes, interruption of risk operations IV. STORM DETECTOR MEASUREMENT A. Values ot the electrostatic field The concentrations of charges and their movement, see Fig. 3, lead to significant variations of the electrostatic field on the ground, see Fig.4 and 5. We can consider the whole as an atmospheric capacitor. Fig. 3 Global atmospheric capacitor Fig. 4 Electrostatic field under a thunderstorm Fig. 5 Field recording, February 4, 2014, LORIOL This record shows a stormy episode (Level 3 reaches) with heavy rain but not lightning. Different strike, see fig. 6, can happen during a stormy episode, lead to "field step." These are rapid variations of the electrostatic field at ground. They are detectable due to the time constant of the atmospheric load capacitor, see fig. 7. Fig. 6 Lightning phenomena: (1) intra-cloud, (2) in the air, (3) between clouds, (4) cloud to ground Fig. 7 Field recording, August 30, 2012, CALMONT This record shows a stormy episode with many "field step” which we cannot given or the type of discharge or its distance from the site. B. Storm detector setting The position in space of the sensor influences its measurement. On the production line, it is calibrated on a stand adjustment; accuracy is a few V / m in the range 0 to 400 kV / m. This calibration is done according to its shape and an ideal position on a mat, 50 cm above a flat floor space. Any peak effect strengthens the local field measurement and thus changes the measured relative to the ideal value. The alarm is linked to threshold crossing. It seemed preferable to present comparable threshold values to the different ANTARGAZ’s operators. The sensor adjustment is performed initially, after a month of operation, by comparing the average value of the field in nice weather (not stormy periods) with a theoretical value. This is not sufficient because this field can vary throughout the seasons, the region and the daytime or the night time. Then a second time with respect to maximum values periods of storms, which should not exceed 20 kV / m. From experience this procedure is satisfactory. Setting uses a coefficient called "environmental coefficient" of multiplying the measured value to present the operator with a corrected value. V. DETECTION VALIDATION During one year we compared alarms storm detector STORMDETEC© with those issued by the French monitoring network of lightning METEORAGE on LPG storage facility LA GARDE located near Toulon. An area of 20 km around the site was chosen for this comparison, see Fig. 8. Each second ground impact in this area initiates a starting METEORAGE alert. The end is determined by a timeout of 30 minutes after the last detected impact. Fig. 8 METEORAGE impacts detected, July 18, 2013, LA GARDE Fig. 9 Field recording, July 18, 2013, LA GARDE, zoom level 3 The alarm level 3 of the storm detector was triggered between 5:10 p.m. 4:05 p.m. The alert was triggered by METEORAGE between 12:23 5:43 p.m. and between 7:58 p.m. 7:28 p.m. Fig.10 “télé-compteur” METEORAGE, July 18, 2013, LA GARDE This Tele-counting shows that the impacts in a 2 km radius around the facility of LA GARDE occurred between 4:32 p.m. 4:03 p.m. 24 strikes including 20 cloud to ground were counted. The number of "field steps" detected by the storm detector is approximately 10 major and more than 30 in the same period of time, see Fig. 9. A summary of events compared between alarms Level 3 storm detector STORMDETEC© and METEORAGE alerts is given in Table 1 for the LA GARDE facility between April to September 2013: During this period, we found 34 days with alarms that: 30% with METEORAGE alerts coinciding with alarms level 3 STORMDETEC©, see the day of 18 July 2013 Fig. 9. 35% with METEORAGE alerts coinciding with "field steps" detections by STORMDETEC© but without alarm level 3, see the case of the day on June 9 fig.11. That day, "field steps" were detected with less than 2 kV/m level coinciding with the warning METEORAGE sequence. The lightning was far from the site. Fig. 11 Field recording, June 9, 2013, LA GARDE, zoom after 10 p.m. The remaining 35% is with Level 3 alarms triggered by the STORMDETEC© coinciding with threshold crossings but without "field step". No lightning was seen by METEORAGE, see if the day of May 18, fig. 12. Fig. 12 Field recording, May 18, 2013, LA GARDE VI. CONCLUSION On the facility of LA GARDE the year 2013 taking given the working periods we calculated that lightning prevention entailed less than 5% of working time pauses. Knowing that most of the loading of trucks has been postponed, not canceled or relocated, the loss is less. Prevention lightning dangerous operations in ANTARGAZ’s GPL storage facilities with a storm detector types field mill approach is a perfectly answer for the safety protection of operators and infrastructures while allowing to better manage operating business. It has led to acceptable operating losses. ACKNOWLEDGMENT METEORAGE for the authorization of data publication DUVAL MESSIEN for STORMDETEC© data REFERENCES [1] C. GARY, La foudre, Dunod [2] C. BOUQUEGNEAU, Doit-on craindre la foudre, EDP Sciences [3] NF EN 50536 Protection against lightning - Thunderstorm warning systems.