Encyclopedia of Environmental Science and Engineering, Volume I and II

(Ben Green) #1

PCBs AND ASSOCIATED AROMATICS 909


use of less flammable fluids, however, points to the fact that
both hydrocarbon and silicone oils do perform satisfactorily
and are suitable as askarel replacement fluids.
The decision to retrofill rather than replace a contami-
nated transformer depends, as discussed above, upon a
number of different factors. For example, the cost of new
oil and the cost of labor have to be weighed against the ben-
efits of reduced maintenance, service and disposal costs, the
impact of public perception, and reductions in the possible
costs of leaks, spills or fires. Once a decision has been made
to retrofill, the usefulness of the action is governed by the
effectiveness of the process to mitigate the problem of PCB
leaching from the impregnated coil.
A cleaning process for the decontamination of PCB
contaminated mineral oil in electrical equipment has been
described in the form of guidelines by the Ontario Ministry of
the Environment (1978). The stages involved are as follows:


  • The transformer is emptied of bulk PCB by open-
    ing the drain valve. The unit is allowed to drain
    into an appropriate container for at least 24 h.

  • The drain valve is then closed or the unit sealed and
    filled with solvent and left filled for at least 24 h.

  • The solvent should then be drained and the trans-
    former rinsed two more times by filling and drain-
    ing with clean solvent.

  • The second and third rinses may be retained for
    use as the first rinse in decontaminating other
    equipment.


The PCB solute in the mineral oil is distributed through-
out the liquid and solid phases of the transformer. The PCB
concentration remaining in the oil at equilibrium can be calcu-
lated from the partition coefficients for PCB between the trans-
former solid materials and the oil and the residual amounts of
flushing liquid remaining in the tank between each stage.
Considerable work has been done on the solvent decon-
tamination of electrical equipment and results have been
reported by S.H. Hawthorne^91. It was found that draining
alone removed 91–96% of the PCB while a combination
of draining and cleaning with trichloroethylene reduced the
PCB concentration by 99.76%. Soaking the drained trans-
former with hot solvent followed by vapor phase degreas-
ing enabled 99.10% of the PCB to be removed. When an
extended soak of 36 days was given to the transformer after
the cleaning procedure had been completed the amount of
PCB removed increased from the above figures to 99.04%,
99.96%, and 99.72% respectively, that is, the remaining PCB
concentrations were 600 ppm, 400 ppm and 2,800 ppm.
Hawthorne’s data allow some important conclusions:


  • A simple transformer flushing procedure is effec-
    tive in reducing the initial PCB concentration of an
    askarel filled transformer to less than 500 ppm.

  • The PCB remaining after each cleaning pro-
    cess resided almost exclusively in the interstices
    between the steel core plates and in the paper
    and insulation of the copper coils, with little or


no PCB remaining on the internal surfaces of the
case and radiators.


  • After the cleaning procedure had been completed
    the transformers were soaked for 36 days with the
    solvent from the third soak to remove the most
    rapidly leached PCBs. During this time it was
    found that the PCB concentration levelled out to
    a pseudo-plateau at 3500 ppm PCB in trichloro-
    ethylene after 10 days. In other words, 1000 ppm
    PCB had been leached out of the core/coil assem-
    bly in 10 days (100 ppm PCB/day). Also, the PCB
    concentration did not increase more than about
    140 ppm PCB in the 26 days following the start
    of the plateau. This means that the leaching rate
    during this period is about 5 ppm PCB/day.

  • The use of hot solvent was found to be more
    effective than cold solvent. Thus, when the retro-
    fill had been completed and the transformer is put
    back under load, the heated oil can be expected to
    leach PCBs at a greater rate than if the transformer
    were not in operation. According to Hawthorne’s
    results the initial leaching rate should be at least
    5 ppm PCB/day.

  • The lowest PCB concentration attained before
    the transformer was put back into service was
    400 ppm PCB. If the leaching rate remained
    steady at 5 ppm PCB/day for the 90 days required
    by EPA, then the concentration would climb
    above 500 ppm PCB and the unit would remain
    a PCB transformer, by definition. This assumes
    that the leaching rate is 5 ppm PCB/day. It is also
    assumed that the leaching rate remains constant
    at its initial, and therefore probably highest, rate.
    Even so, the results illustrate that a simple solvent
    flushing technique is able to reduce the PCB con-
    centration into the range where the transformer, in
    the absence of leaching from the core/coil, could
    be reclassified as a PCB contaminated unit.


An alternative method of cleaning involves vapor degreasing
the unit by spraying the cold internal surfaces with a solvent
vapor such as trichloroethylene generated by boiling the sol-
vent in an external vessel. The principle of operation is not
unlike the soxhlet extraction of materials.
The process offers numerous advantages:


  • the volume of contaminated solvent is kept to a
    minimum;

  • the solubility of PCBs in the hot solvent is greater
    than in cold liquid and therefore the extraction
    process is less time consuming;

  • the low viscosity solvent is able to penetrate the
    capillaries of the paper and core laminations and
    remove the PCBs from all the surfaces within the
    transformer.


The action of the vapor phase degreasing process is to ini-
tially dissolve the PCB on surfaces. Beyond this, capillary

C016_003_r03.indd 909C016_003_r03.indd 909 11/18/2005 1:12:39 PM11/18/2005 1:12:39 PM

Free download pdf