Explosion protection Old Truth Stands no Longer: How dangerous is splash filling really?
One of the basic mantras when handling flammable liquids is to avoid free falling and splash filling at all costs. Dangerous electrostatic charges could otherwise cause an explosion with devastating effects. But how substantiated are these concerns? Recent experiments in Sweden and France came up with surprising results.
Filling tanks or containers with liquids is frequently done in the chemical industry. In many cases, the liquid is filled via a pipe ending at a distance above the bottom of the container and therefore may splash to the bottom or liquid surface. This process has often been blamed to generate dangerous discharges that may ignite dust or gases in the receiving tank.
Little has been known about the electrostatic charge that is generated through the splash filling process. Some experts claim that the Lenard effect, which states that electrostatic charges are generated when drops of a polar liquid are torn, could provide for an ignition source. Some scientists assume that falling droplets are charged by their contact with air molecules and thus create electrostatic hazards during splash filling.
To overcome this issue, first of all some theory is required: It is commonly assumed that electrostatic charging is caused by friction. This, nevertheless, is an over-simplification: In fact, charging is actually generated by separating materials which were in close contact before, if at least one of them is electrically insulating. As these charges occur at the material boundaries, so that the formation of interfaces is an absolute.
While these are natural for solid materials, liquids form them at their free surfaces. Only gases show no interfaces. Therefore, charge may occur at interfaces of solid/solid, solid/liquid and liquid/liquid but not between gases on the one hand and liquids or solids on the other.
Experience seems to contradict this fact as it is given proof that liquids will be electrostatically charged when being atomized, for example in air. But is there really a conflict? Not at all because when atomizing a liquid it will be disaggregated and at the interfaces resulting electrostatic charging comes into being, regardless of its conductivity and not by friction with the surrounding air.
The Electric Waterfall
However, science is no matter of faith but of evidence, for example by experiments. In 2015 these were conducted at the University of Poitiers, France. With a falling height of 5 m, droplets of different liquids at diameters of about 2 mm were tested. Electrostatically charged droplets, accurately defined, were released out of a dispenser fell through cylindrical metal pipes and were collected in a metal cup, all surrounded with an earthed tube.
This measuring principle bases on electric induction and enables to record the extremely low amounts of charges (range of Pico Coulombs of single droplets), being transferred to an electrometer amplifier and indicated on a PC. The measured charge on a single falling droplet is depicted in the diagram on page 44 and in comparison therewith the theoretical course is shown.
The Amplitude Response
The decisive criterion for evaluation of the droplet charge is its amplitude response, whereby the small changes in view of the extremely small amounts of charge have to be tolerated. It is remarkable to notice, that the peak is getting narrower due to the increasing falling velocity.
It follows: The charge level indicated at the electrometer when the droplet passes the first cylinder is virtually the same when it passes the second, the third, the forth and finally arrives at the collecting cup.
There is a very low charge discrepancy (possibly caused by evaporation of the liquid) which was not detectable even by the extremely sensitive electrometer. In general, no change of the charge was noticed all along the fall on droplets that kept in shape.
It also was determined, that droplets which disassembled themselves in multiple parts during the fall showed a change of their electrostatic loading status. This statement corresponds with the experience of the so called “waterfall electricity”. With a water cascade falling down several hundred meters, electrostatic charges occasionally come into being within the wafts of mist.
It is proven that the assumption that droplets are charged by friction with air does not hold true, and that applies to a closed liquid jet as well. But this result gained by tests in the micro-liter scale does not help very much, however, the question is, are there any electrostatic ignition risks when pouring flammable liquids at liter-scale? That means: Single drops are one matter, larger volumes of liquid that splash are something else.
But as the old saying goes, “the proof of the pudding is in the eating” — in other words: A scientific examination has to be conducted. On the contrary of investigating the droplet charge, charge measurement is no longer sufficient. It is especially important to focus on the question whether the resulting amount of charge is capable of igniting flammable liquids.
Decisive for that is the amount of energy necessary for ignition, called Minimum Ignition Energy (MIE). In the “Electrostatic Hazards Guideline IEC 60079-32-1 Table C2” a larger number of flammable liquids with their MIE values are listed and, fortunately, correlated with the relevant Minimum Ignition Charge (MIQ).
The Missing Link
This makes it possible to find out which hazards arise from splash filling of flammable liquids. However, one link is still missing: How is it possible to determine without contact the amount of charge implemented in a gas discharge?
A knowledge gained a long time ago can help further: Gas discharges manifest themselves, e.g., by cracking noises in a radio receiver of older design. As far as they feature plasmas, in general necessary for ignition, they emit high frequency signals which may be received by appropriate antennas. To measure these signals antennas were tested while connected to an oscilloscope.
Programmed to Receive
Discharges from a plastic box to an earthed probe were used to control if signals appear. It soon turned out that the antenna should be circular and about a meter in diameter. Fortunately this did fit into a test tank suitable for trials. Two different antennas with minor different characteristics were used to obtain reliable results. The test tank (which was 3.56 m high with 2.08 m in diameter) made of stainless steel was borrowed from a nearby chemical factory run by Akzo Nobel.
To calibrate the antennas a coulomb meter was used, installed at the bottom of the test tank. Objects (insulating or conductive) were electrostatically precharged and then lowered down to the probe of the coulomb meter. The readings of the latter were compared with the signals from the antennas and forwarded to the oscilloscope where they could be evaluated.
Setting the Stage
From a small storage metal vessel which was lowered down into the inside of the large tank with a rope, different liquids were allowed to fall down and splash from different heights. The experimental setup was checked by charging the storage vessel filled with conductive tap water to about 10 Kilovolts.
Then the water was allowed to splash 3.5 meters downwards through the antenna assembly and into the receiving vessel. Both one and 10 liters of pre–charged tap water showed strong electrostatic discharges (about 50 times above the sensitivity level of roughly 1 Nano-Coulomb).
Three different liquids were tested, one of them with low conductivity (transformer oil), one with high conductivity (tap water) and one intermediate medium (regular diesel fuel). In all of the 90 experiments with different volumes (one or ten liters), no discharge from the splash filled liquid could be detected. Therefore, the static charge generated by splash filling with a splashing height 3.5 meter must be below approximately 1 Nano-Coulomb, independent of the conductivity of the liquid.
Coming Full Circle
This is where everything comes full circle in relation to the already mentioned IEC 60079-32-1 Table C2: Single splash filling by itself in the liter-range will not create any electrostatic ignition hazards due to the splashing process.
However, it has to be taken into consideration that precharged splashing liquids do in fact very well create ignition hazards, as they may build up charged slugs of conductive liquids — a phenomenon that was clearly established by the splash filling experiments. Consequently, for a risk evaluation, especially in the Chemical Industry the electrostatic charge generated in filters, pumps, and pipes must be taken into account and evaluated.
A Matter of Scales
Accordingly it can be readily deduced that all liquids in amounts up to at least ten liters, if not precharged by any previous action, will not generate any dangerous discharges when falling down up to three meters by means of gravity. But it is essential to note that the bucket is made of conductive material and reliably earthed.
So the investigations of electrostatic liquid charging by free fall through air are finished in the range of milliliters and liters. The researchers plan to continued them in the range of hectoliters but because of the elaborate process this cannot be scheduled at the moment.