The reports and certificates issued by Graham Hearn at Wolfson electrostatics have been quoted with vigor by opponents to conductive piping. New review show that the Wolfson tests has flaws. The flaws are of the character that the published articles and certificates issued by Wolfson Electrostatics no longer can be used to prove that non-conductive piping is electrostatically safe - even if installed correctly.
For more information, refer to Electrostatic ignition hazards with plastic pipes at petrol stations by Dr. Harold Walmsley, published in the Journal of Loss Prevention in the Process Industries, Volume 25, Issue 2, March 2012, Pages 263–273. See an abstract of or buy the complete published article at www.sciencedirect.com. Electrostatic ignition hazards arising from fuel flow in plastic pipelines by G. L. Hearn in Journal of Loss Prevention 15 (2002)
For a correct assessment of the risk of electrostatical ignition in a piping system, worst-case conditions must be replicated in test or the test must allow the worst-case conditions to be estimated. The important factors to include are:
Fuel charging properties like fuel conductivity, impurities, additives and specific fuel blend. Variances between fuel batches are large.
Fuel velocity (speed).
Flow time of the fuel in the pipe.
Pipe line geometry and configuration. Elbows, reductions and filters increase turbulence and increases charge generation. As installation practices varies and conditions at individual stations never are precisely the same it is hard to establish worst possible case.
Dissipation or relaxation conditions. These include factors like temperature, humidity and resistance to earth.
Fuel charging properties Hearn used a 50/50 toluene/iso-octane mix in his tests. The extra processing needed to isolate the pure components tends to remove polar impurities that are usually present in commercial fuels and that would increase charging if present. The streaming current of a synthesised test fuel is likely to be lower than that of a worst-case petrol by a factor of 2 to 5.
Flow velocity Hearn used flow velocities of 1.6 to 2.8 m/s. These are realistic flow velocities for the filling station.
Flow time in the pipe Voltages on non-conductive piping increases with the time fuel keeps flowing through the pipe. A typical gravity fuel drop takes about 15 minutes. Worst case is unknown but will be higher for larger tanks and if 3" fill pipe is used instead of 4" pipe. Flow time in Hearn's tests was limited by the available volume of test fuel. Each test run was 35-60 seconds. The maximum voltage 15 minutes into delivery is approximately double the voltage 1 minute into delivery. If the delivery continues for more than 15 minutes the voltages will be even higher.
Pipe line geometry and configuration The pipe configuration in Hearn's tests is realistic, but might not be worst case either.
Dissipation or relaxation conditions The resistances of dissipation paths through both the fuel and the pipe walls increase at low temperature and humidity. Humidity during Hearn's tests was 35-47% and temperature 10-18°C. This is not worst-case conditions.
Conclusions Hearn used a realistic pipe configuration but did not ensure that streaming currents, test duration or dissipation conditions were worst-case. Because of this his measured voltages are likely to be several times lower than the worst-case. When this is allowed for, his data suggests that brush discharges from insulation surfaces and sparks from unbonded conductors are both credible ignition sources.
A number of documented incidents supports this conclusion. Most incidents appear to be associated with sparks from isolated conductors but ignitions from incendive brush discharges cannot be ruled out.
Based on the above critique of Hearn's tests they can no longer be used to justify the use of non-conductive piping on filling stations.
International Seminar for Fuel Pipeline Safety: The Report
This is a report written on tests performed in April 2010. In the report is stated the importance of simulating worst case conditions in order to achieve reiable results. When reviewed, however, the report suggests that a number of deviations from worst-case conditions were made during the tests:
Worst-case fuels were not used
Worst-case temperature and humidity was not used
Worst-case pipe geometry was not used
Attempts were made during the test to record lower voltages than those measured.
Worst-case fuels were not used
Fuel with worst-case charging properties was not used in the tests. Instead the test fuels used are very "pure" and unlikely to contain the impurities and polar additives usually found in commercial fuels. Advice from an independent expert on which fuels to use for testing was disregarded.
Worst-case temperature and humidity were not used
Temperatures record during the tests range between 13° to 39.5°C (heated) which is too warm to provide worst-case conditions. Cold and dry winter conditions are most likely to provide the worst case.
Worst-case geometry was not used
The length of pipe exposed in the tests were very short and considerably longer exposed sections of pipe are often seen in tank chambers. The large amount of steel pipe and fittings in the chambers are not only unrepresentative of realistic installation, they also limit the electrostatic field to be measured. Thus the test geometry used for the tests was not worst-case. Calculation of possible consequesces for deviations from worst-case
To allow for worst case fuel charging a typical batch variation factor of x3 should be used.
Voltages are approximately proportional to exposed pipe length, so a factor of x2 is needed to allow for something close to worst-case.
The possible consequences of the presence of biocomponents is hard to quantify, so are ignored for a conservative estimate. The same goes for temperature and humidity, their influence is also ignored.
The maximum voltage recorded in the tests was 4.7 kV. Applying the x3 and x2 correction factors suggests that the worst-case equivalent voltage would be at least 28 kV (4.7 kV x 3 x 2 = 28.2 kV). This is above the suggested hazard threshold of 25 kV, even though several less-than worst-case factors were ignored due to the difficulties to estimate them.
The tests show that with highly charging fuels and worst-case conditions, hazardous voltages can occur in non-conductive piping systems.