FI78367B - FOERFARANDE FOER ATT ERSAETTA PCB-HALTIGA ASKARELBLANDNINGAR I ELEKTRISKA INUKTIONSANORDNINGAR MED PCB-FRIA DIELEKTRISKA KYLMEDEL. - Google Patents

Description

1 78367

Method for replacing PCB-containing askarel mixtures used in electrical induction equipment with PCB-free dielectric refrigerants The method for replacing PCB-containing dielectric refrigerants with PCB-free dielectric refrigerators

The invention relates to electrical induction devices, for example electrical power transformers, in particular to dielectric coolants contained in these devices, and more particularly to those coolants which comprise and which contain polychlorinated biphenyl, PCB. More particularly, this invention relates to methods of converting PCB-containing electrical induction devices, such as transformers, to substantially PCB-free transformers to bring these transformers into compliance with "PCB-free transformers" according to U.S. regulations.

PCBrt has been found to be an excellent transformer coolant due to its non-combustibility, chemical stability and thermal stability as well as good dielectric properties. U.S. Patent No. 2,582,200 discloses the use of PCBm alone or a mixture thereof and compatible viscosity regulators such as trichlorobenzene, and such mixtures of trichlorobenzene and PCBs are commonly referred to as "ascarel mixtures." These askarel mixtures may also contain small amounts of additives such as ethyl silicate, epoxy compounds and other such substances used as cleaning additives due to the decomposition products of halogen compounds which may result from the arcing that may occur. ASTM D-2283-75 describes numerous types of askarel and considers their physical and chemical properties.

2 78367

However, in the U.S. Toxic Substances Control Act of 1976, PCBs are classified as environmentally and physiologically hazardous substances and, due to their good chemical stability, PCBs are non-degradable in nature. Thus, they persist in the environment and may even increase their concentration in biological terms (accumulation in higher organisms through the food chain). For this reason, transformers are no longer manufactured in the United States using PCBs or askarel fluids. As old units containing PCBs are still used in certain circumstances, it is necessary to take precautions such as ditches and to carry out regular inspections. A further disadvantage of transformers containing PCBs is that maintenance requiring heart emptying is prohibited and that the owner of the transformer is responsible for any contamination of the environment, including cleaning costs due to leaks, tank damage or other PCB emissions, or toxic side effects from fires. . To replace a transformer containing PCBs, it is necessary to (1) turn off the transformer, (2) drain the transformer from the PCB and flush the unit as instructed, (3) remove the unit and replace it with a new transformer, and (4) transport the old transformer to an approved landfill. area for burial (or solid waste incineration plant). Even after this, the owner of the transformer who agrees to bury it will continue to own this transformer and will remain responsible (legally responsible) for any future pollution problems caused by the transformer. Liquid waste generated during this compensation measure must be incinerated in facilities specifically approved for this purpose. Thus, replacement of a PCB transformer may be expensive, but more significant, as most transformers containing PCB alone or askarel transformers are indoors, in building foundations, or in special enclosed, potentially inaccessible spaces, removal or installation of the transformer may not be physically possible. nor is it desirable for capital to be exploited.

3 78367 A desirable approach to this problem would be to replace PCB oil with a harmless / compatible liquid substance. The new transformers use a number of different types of liquids, as described in Robert A. Westin, "Assessment of the Use of Selected Replacement Fluids for PCB's in Electrical Equipment," EPA, NTIS, PB-296377, March 1, 1979; J. Reason and W. Bloomquist, "PCB Replacements: Where the Transformer Industry Stands Now," Power, October, 1979, pp. 64 ... 65; Harry R. Sheppard, "PCB Replacement in Transformers," Proc. Of the Am. Power Conf., Ss. 1062 ... 68; Chem. Week, 130, 3, 24 (1/20/82); A. Kaufman, Chem. Week, 130, 9, 5 (3/3/82); CMR Chem. Bus., October 20, 1980, p. 26; Chem. Eng., July 18, 1977, p. 57; BE Patent Publication 893,389; Europ. Plastic News, June, 1978, p. 56. These include silicone oils such as polydimethylsiloxane oils, modified hydrocarbons (to provide a high flash point, such as RTEmp, a liquid substance protected by RTE Corporation), synthetic hydrocarbons (poly-alpha-olefins). ), high viscosity esters (such as dioctyl phthalate and PAO-13-C, a liquid substance patented by Uniroyal Corporation), and phosphate esters. Numerous halogenated alkyl and aryl compounds have been used. These include liquid trichloro- and tetrachlorobenzenes and toluenes, as well as patented mixtures thereof (e.g., liquid mixtures of isomers of tetrachlorodiarylmethane and trichlorotoluene). Liquid mixtures of isomers of trichloro- and tetrachlorobenzene are particularly suitable because they are non-flammable (i.e., have a high flash point) and because their physical and chemical properties are similar to those of ascarel mixtures to be removed. Other such proposed liquid substances are tetrachlorethylene (e.g. 4 78367

Diamond Shamrock product Perclene TG) and polyol it as well as other Ester it.

Of all these PCB-free liquids, the most widely accepted are silicone oils. Their chemical, physical and electrical properties are excellent. They have high flash points (above 300 ° C) and no known toxic or environmental problems. These oils are poly (dime-style siloxanes) terminated with trimethylsilyl:

(CH3) 3S1O ((CH3) 2S1O) nSi (CH3) 3 Formula A

wherein the value of n is sufficient to provide the desired viscosity, such as a viscosity of about 50 centistokes (mm2 / s) at 25 ° C. Suitable commercial silicone oils are available from Union Carbide (L-305) and the like. In addition, U.S. Patent No. 4,146,491, GB Patent No. 1,540,138, and GB Patent No. 1,589,433 disclose mixtures of silicone oils and a variety of additives, these additives improving the electrical performance of silicone oil in capacitors, transformers, and other similar electrical devices. and aryl groups.

Replacing the PCB-containing askarel mixtures used in old transformers with silicone oils or some other replacement liquid substance may seem like a simple solution, which it is not. A typical transformer contains a lot of insulating cellulosic material that prevents improper contacts of electrical windings, etc., as well as the formation of arcs in them. Of course, this material is impregnated with the askarel mixture and may comprise 3 to 12% of the total volume of the liquid in the transformer. This absorbed ascarel mixture cannot be drained out or flushed out of the transformer by any known method, regardless of the efficiency of the method. When the original loose ascarel mixture is replaced with a new PCB-free liquid, the slow diffusion process gradually causes the old absorbed ascarel mixture to be gradually released, and the PCB content of this new liquid substance increases. As a result, this new coolant becomes contaminated.

In the U.S. classification of transformers, liquid substances with a PCB content of more than 500 ppm are referred to as "PCB transformers", substances with a PCB content of 50 to 500 ppm are referred to as "PCB-contaminated transformers", and PCBs with a PCB content of less than 50 ppm as "PCB-free transformers". While substances classified in the first two ways may incur the most significant costs in the event of a potential release or decommissioning, there are no U.S. regulations for the latter category. To reach this last category, the PCB content must remain below 50 ppm for at least 90 days with the transformer in operation and with an energy level sufficient to produce temperatures of at least 50 ° C. This requires an average elution rate of about 0.56 ppm / day over those 90 days. It is assumed that most, if not all, of the US states will introduce regulations equivalent to or more stringent than those issued by the US government. Milder regulations may be possible elsewhere.

There are numerous commercial emptying methods on the market, including those described in Dow Corning Corporation's advertising publication No. 10-205-82 (1982), "The RetroSil PCB Removal System", which discusses a PCB removal system, and a commercial publication by Positive Technologies, Inc. on a process called Zero / PC / Porty. These methods apply the most efficient measures for pre-cleaning, during which 6 78367 electrical equipment is not in operation. Most use flushing kits with liquids such as fuel oil, ethylene glycol, or a variety of chlorinated aliphatic or aromatic compounds. Trichlorethylene is preferably used as the rinsing liquid. Some purge processes, such as the Zero / PC / Forty method of Positive Technologies, Inc., alternate between gas scrubbing of fluorinated hydrocarbons and liquid purging. Once this first cleaning step has been completed, the transformer is filled with silicone fluid. Although these thorough rinsing operations are expected to be effective, they are not capable of removing PCB adsorbed on the media of the cellulosic material. As a result, the PCB content of the silicone refrigerant gradually increases as the PCB remaining in the transformer is extracted into the liquid when the transformer is used. Therefore, if the goal is to achieve a PCB-free state (the "PCB-free transformer" category as defined in U.S. regulations), it is necessary to either periodically change this silicone fluid or clean it continuously until an extraction rate of less than 50 ppm is reached for 90 days.

Periodic replacement is very expensive, and because both silicone and PCBs are essentially non-volatile, distillation cannot be used to separate them and other separation methods are expensive or inefficient. The Dow Corning RetroSil process uses continuous carbon filtration to purify this fluid ("The RetroSil PCB Removal System", Dow Corning Corp. Publication No. 10-205-82 (1982); Jacqueline Cox, "Silicone Transformer Fluid from Dow Reduces PCB Levels to EPA Standards, Paper Trade Journal, September 30, 1982; T. O'Neil and JJ Kelly, Silicone Retrofill of Askarel Transformers, Proc. Elec./Electron. Insul. Conf., 13, 167 ... 170 (1977), WC Page and T. Michaud, "Development of Methods to Retrofill Transformers with Silicone Transformer Liquid", Proc, Elec./Electron. Insul. Conf., 13, 159-166 (1977) ). In U.S. Patent No. 4,124,834, Westinghouse has patented 7,738,367 of a transformer and associated filtration method for removing PCBs from the coolant, while RTE discloses the use of chlorinated polymers as an adsorbent in EP Patent Publication No. 0023111. However, the filters used in these methods are very expensive and PCB removal is very inefficient due to both the lack of selectivity and the very low concentration of PCB to be filtered. Instead of filtration, methods using decantation have been proposed (U.S. Pat. No. 4,299,704), which, however, are impractical due to the limitations of solubility and are good only in the presence of high concentrations; or methods using extraction with polyglycols (F. J. Iaconianni, Ά.

J. Saggiomo and SW Osborn, "PCB Removal from Transformer Oil", EPRI PCB Seminar, Dallas, Texas, December 3, 1981) or with supercritical carbon dioxide (Richard P. deFilippi, "CO2 as a Solvent; Application to Fats, Oils and Other Materials ", Chem. and Ind., June 19, 1982, pp. 390-94) and the chemical destruction of PCBs with sodium (GB Patent 2,063,908). None of these solutions have been found to be economically or commercially viable in the case of askarel transformers. However, the filtration solution could be a relatively efficient, albeit expensive, operation, except for the fact that the extraction rate is so slow that it could take several years to reduce the residual PCB content to a level where final extraction has slowed to an acceptable value (Gilbert Addis and Bentsu Ro, “Equilibrium Study of PCB’s Between Transformer Oil and Transformer Solid Materials,” EPRI PCB Seminar, December 3, 1981).

This problem and its causes are discussed in L. A. Morgan and R. C. Ostoff, "Problems Associated with the Retrofilling of Askarel Transformers," IEEE Power Eng. Soc., Winter Meeting, N.Y., N.Y., January 30-February 4, 1977, inscription A77, page 120 ... 9. The solubility of a typical silicone oil in PCBs is virtually zero (less than 0.5%) at temperatures of at least 100 ° C, while the solubility of PCBs in silicone at 25 ° C is only 10% and at 100 ° C only 12%. Although this limited solubility does not prevent the majority of the silicone from dissolving the available free PCB, it does limit the ability of the PCB to diffuse from the pores or spacers of the cellulosic material.

For any pore filled with PCB, out-diffusion of PCB must be accompanied by silicone in-diffusion. At a certain point, an interface between the PCB and the silicone must form inside the pore, through which neither material can diffuse very quickly. Because PCB is more soluble in silicone than inverted, PCB slowly diffuses into silicone as the interface gradually progresses into the pore. This limited solubility limits the rate of diffusion, and although this mechanism allows the pore to eventually be cleaned of PCBs, it is several orders of magnitude slower than if the two liquids were miscible with each other. The high viscosity of silicone (and many other refrigerants) is also a barrier. The result is an extended extraction period, possibly several years, during which the silicone must be continuously filtered or replaced periodically to remove PCBs. Thus, the slow extraction of PCBs with silicone from a solid insulating material is a worse option than not performing this extraction at all, as emissions of PCB-containing materials are at risk for years. Experimental studies by Morgan and Osthoff showed, for example, that the effective diffusions of PCBs in a typical silicone oil were only 1/10 of the diffusibilities obtained with a hydrocarbon oil having a viscosity of 10 centimeters (mm / s), although for this reason the use of such a hydrocarbon oil However, if the susceptibility of hydrocarbons to combustion were not to be taken into account, the problem of still separating PCB from contaminated hydrocarbon oil having a high boiling point as with PCB and as with silicone oil would still be solved.

The present invention is based on the fact that there are suitable coolants which are more useful than silicone oil for limited time operations in which the extraction is completed. They are volatile enough to separate from the PCB by distillation, easily miscible with the PCB, and have a relatively low viscosity, which ensures rapid diffusion into the pores of the insulating material. The other ingredients in Askarel mixtures, i.e. trichlorobenzene and tetrachlorobenzene, have been found to be ideal liquids for this purpose. They can be used as temporary or temporary extraction and cooling fluids when there is a risk of fire, while light hydrocarbons could be used if there is no risk of fire.

It has been found that no method previously known in the art provides a solution for providing a substantially PCB-free transformer by removing, rinsing and eluting ascarel mixtures from transformers containing these mixtures with a temporary dielectric fluid, or the steps of filling a transformer tank with a transformer tank. with a miscible, temporarily used dielectric coolant capable of penetrating said electrical insulation and separable from the PCB, or a step in which the transformer is in its electrical operation while the PCB is eluted with the temporarily used dielectric and allowing this electrical operation to continue a period of time sufficient to elute the PCB impregnated in the solid insulator into this temporary dielectric coolant and in which the temporary coolant contained in the PCB is discharged out of the transformer; the filling cycle is repeated with pure temporary coolant, the electrical operation and the effluent are repeated sufficiently 10 7 8 3 6 7 times until the elution rate of the PCB drops to 50 ppm, based on the weight of the permanent refrigerant used, and after 90 days of electrical operation. the coolant is drained out of the transformer and separated from the PCB it contains, whereby the tank can be filled with a PCB-free, permanently usable dielectric coolant which remains substantially PCB-free during subsequent electrical operation.

The characteristic features of the invention appear from the appended claims.

The present invention is based on the use of a suitable temporary or temporary coolant to replace PCB-containing refrigerants in electrical induction devices, such as transformers, where a vessel (such as a tank) contains a refrigerant, an electrical coil and a porous solid cellulosic insulator embedded in and impregnated with the PCB. allowing this transformer to operate electrically for a period of time sufficient to elute the PCB from the solid electrical insulating material contained in the transformer. During this operation, the temporarily used dielectric coolant is changed to speed up the elution process, with the desirable goal of eluting the extractable PCB to such an extent that the transformer can then operate for 90 days without exceeding the PCB concentration of the permanent coolant for the transformer. After the amount of extractable PCB in the transformer has been reduced to this desired level, the temporary dielectric coolant is removed from the tank, after which this tank is permanently filled with a PCB-free dielectric coolant suitable for use in this transformer.

In the pictures of the accompanying drawings

Figure 1 shows a comparison between the PCB analysis values obtained in the first section of the three transformers.

Figure 2 shows the effect of temperature on PCB elution rate. Figures 3 to 5 show the PCB content of the temporary di- 11 7 8 3 6 7 electric liquid used in the transformer as a function of the elapsed days.

The following is the method of the present invention for replacing a PCB-containing liquid in a transformer with a permanently used PCB-free coolant: (1) The transformer is turned off (energy level is set to zero) and the PCB-containing liquid is drained and disposed of in an environmentally acceptable manner. measures. The transformer can be flushed with a liquid flushing agent, such as trichlorobenzene or trichlorethylene, which may be a gas or liquid, to remove "free" PCB liquid.

(2) The transformer is filled with a temporarily or temporarily used coolant such as trichlorobenzene, TCB, or a mixture of it and tetrachlorobenzene which mixes or dissolves PCB and which is able to penetrate the pores of the electrical insulating material and which can also be easily separated from the PCB. the electrical operation of the transformer is restored.

(3) The coolant temperature is monitored, and if the electrical load on the transformer does not produce a sufficiently high coolant temperature to provide the desired rate of PCB elution, thermal insulation or even external heating may be used. Likewise, recirculation of coolant through the outer loop and pump can be applied to heat the coolant or improve internal recirculation.

(4) The elution rate of PCBs in the temporary coolant can be determined by periodic sampling and analysis. The accumulated PCB is removed at certain intervals by removing the temporary coolant contained in the PCB and distilling this temporary coolant, such as trichlorobenzene (TCB), from the PCB. This can be done by shutting down the transformer and bringing its energy level to zero, draining the spent coolant out of the transformer for distillation and replacing it with a new temporary coolant such as TCB. Alternatively, the transformer may be allowed to operate with new temporary coolant, such as TCB, when adding and removing old TCB as a propeller stream or through a recirculation loop.

(5) The PCB-contaminated TCB liquid is distilled to obtain a substantially PCB-free TCB distillate and, as a base product, TCB-contaminated PCB. PCBs can be disposed of by measures approved by the U.S. government, such as incineration.

(6) When the PCB elution rate reaches the desired level, preferably less than 50 ppm PCB based on the weight of the permanent coolant to be used over a 90-day period (i.e., the elution rate is 5/9 ppm per day), refilling with the permanent coolant can be performed. The transformer is switched off (the energy level is reduced to zero), the transformer is drained and filled with silicone oil or other permanently used coolant suitable for this transformer. After that, its operation is restored.

(7) In order to comply with US regulations for PCB-free transformers, the PCB analysis should show after 90 days that the PCB content is less than 50 ppm PCB (based on the weight of the permanent coolant used), after which the transformer is reclassified PCB-free (ie "PCB free").

Although efficient emptying and rinsing techniques should be used, considering the rinsing step, these techniques do not in themselves constitute the present invention, but are part of all previously known refilling methods. They are a prerequisite for the most efficient embodiment of the invention itself, but their importance has hitherto been overestimated, as the slow extraction rate, rather than the rinsing efficiency, has been found to limit the PCB exit rate. Numerous different Liu 13 78367 traps can be used in the flushing step, including hydrocarbons such as gasoline, fuel oil, mineral oil or white spirit, toluene, turpentine or xylene, various chlorinated aliphatic or aromatic hydrocarbons, alcohols, esters and the like. However, in view of the workability of the materials as well as the separation of PCBs, it is practical to avoid using all other types of chemicals that are not necessary, so the use of a temporary extraction fluid, ie TCB or mixtures of tetrachlorobenzene, as the original rinse aid is most practical.

Liquid substances other than the commonly used liquid trichlorobenzene, TCB, or a mixture of it and tetrachlorobenzene may be used. The most preferred temporary coolant has the following characteristics: (a) it is suitable for use with PCBs (i.e., it is desirable that it dissolves at least 50% by weight of PCBs, and more preferably it dissolves at least 90% by weight of PCBs, and most preferably it mixes with the PCB in all proportions}; (b) it has a low molecular weight low enough to ensure good mobility of the molecules, allowing it to penetrate the pores and spacers of the solid insulating material, and promotes rapid diffusion of both components, desirably with a viscosity of 25 ° C. at a temperature of 10 centimeters or less, and preferably 3 centimeters (mm / s) or less, (c) it can be easily separated from the PCB, for example by distillation, desirably having a boiling point of 275 ° C or less, and preferably 260 ° C or less, (d) it is currently considered to be non-hazardous to the environment, and (e) it is suitable for not inside the transformers.

14 7836 7, TCB or mixtures thereof with tetrachlorobenzene are preferably used, but numerous alternatives, as mentioned above, may also be used. These could be modified and synthetic hydrocarbons as well as a wide variety of halogenated aromatic and aliphatic compounds. Numerous different mixtures of liquid isomers of trichlorobenzene are also available. The most preferred TCB liquid would be a mixture of these isomers containing or lacking isomers of tetrachlorobenzene. The advantage in this case is that the freezing point of such a mixture is lower than that of the individual isomers, thus reducing the possibility of this mixture solidifying inside the transformers when used in very cold climates. Furthermore, such mixtures are a common result of the manufacturing process, so they may cost less than the separate isomers separated and purified.

Since the aim here is preferably to cause the PCB to be extracted from the transformer at the highest practically achievable rate, in the most preferred embodiment the transformer is allowed to operate to achieve the highest possible diffusion rates defined in step (3) above. When the transformer is operated at its full rated load, it should automatically generate sufficient heat for this purpose. However, because many transformers operate below their rated load as well as below their safe rated temperature, sufficiently high temperatures (i.e., at least 50 ° C) may not be achieved without thermal insulation or external heating. Although this thermal control is a preferred embodiment of the present invention, it is nevertheless an optional and not a necessary requirement, as there are several transformers for which such isolation or heating is impractical. Extraction at lower temperatures, even at ambient temperature, can be used, but requires more time.

is 78 36 7

The liquid recirculation shown in formula (3) is optional, but is a preferred embodiment, as such recycling prevents the formation of concentration gradients that may delay diffusion. Since elution is a slow process, the recycling rate does not have to be very high. Violent recycling should, of course, be avoided to avoid damaging the transformer's internal structures. It has been found that many transformers, due to their structure or location, cannot be easily converted to use a recycling loop, and such recycling is not considered necessary, but is only one embodiment of the present invention that increases elution rates. In most transformers, the naturally occurring heat gradients alone provide adequate recycling, especially in cases where a relatively low viscosity mobile refrigerant such as TCB is used.

As the PCB content in the TCB or other temporary coolant used in the transformer increases, it may potentially reach a point where PCB is no longer extracted by diffusion from the cellulosic pores or spacers of the insulating material in the transformer tank. The decrease in elution rate determined by sample analysis is an indication that this is the case. If it is demonstrated that this point is reached, then, as shown in step (4), it may be necessary to replace this PCB-rich, temporarily used dielectric coolant with a new, PCB-free liquid. This is most easily accomplished by shutting down the transformer, draining the contaminated extraction liquid (temporary dielectric coolant) out of the transformer, and replacing it with new liquid. Practically speaking, instead of monitoring the elution rate to determine the point at which PCB is no longer efficiently extracted by diffusion from the pores and spacers of the electrical insulating material, it is much more practical to change the transformer coolant at regular intervals. If the target is a PCB-free 16,786,367 transformer, then the refrigerant changes are performed after a certain length of electrical operating cycles until the refrigerant is no longer able to elute 50 ppm of PCB during 90 operating days. The length of the electrical periods between refrigerant changes can be selected from 20 days to one year (or longer if the transformer owner can only interrupt the transformer at certain rare moments, such as during special holiday periods where the downtime may be longer than one year and the transformer may be interrupted only every two years), preferably for 30 to 120 days and most preferably for 45 to 90 days.

The contaminated extraction liquid can then be distilled and concentrated for reuse, leaving a PCB-containing product from the distillation apparatus that is incinerated or otherwise disposed of in accordance with U.S. regulations. The temporary refrigerant is preferably completely replaced, however, it is possible that the inconvenience caused by additional downtime requires a different procedure, i.e., adding new clean liquid to the transformer while removing old contaminated liquid while the transformer is in operation. This method is less efficient because the new liquid mixes with the old liquid in the transformer, and in fact, a liquid with a reduced PCB content is removed. Thus, in order to remove all the PCBs from the transformer, more extraction liquid has to be removed compared to the recommended method. This disadvantage can be reduced by ensuring that undue mixing does not occur. For example, new cooled TCB or other temporary dielectric coolant may be fed to the bottom of the transformer, thereby simultaneously removing the old, warm, PCB-rich temporary dielectric coolant from the top of the transformer. The difference in densities delays mixing. Regardless of the method used, it is necessary to repeat in this process until the desired level of PCB, i.e. less than 50 ppm, is maintained in the silicone oil coolant for at least 90 days.

17,783 6 7

Although distillation is the preferred method used to separate TCB or other temporary dielectric coolant from PCB, other methods may be possible, especially when a liquid other than TCB is selected as this temporary liquid.

There has been some concern that TCB itself, or other chlorinated temporary dielectric coolants such as TTCB and halogenated solvents, may sometimes be considered hazardous to health and that the transformer, although free of PCBs, may be contaminated with TCB or other possibly with a questionable temporary fluid. A further advantage of the process according to the present invention is that such contamination can be easily removed if necessary, since the temporary TCB or other liquid is more volatile than silicone oil or heavy hydrocarbon liquid, or other coolant of relatively high viscosity permanently used in the transformer, and can be distill from this coolant. Thus, the chlorinated portion of the refrigerant can be replaced with pure refrigerant and the old batch can be sent to a distillery for effortless purification. Two or three such refrigerant exchanges over several months will result in a substantially halogen-free system, if desired.

Other preferred refrigerants of a permanent nature that may be used in place of the silicone oil used for final filling include, for example, dioctyl phthalate, modified hydrocarbon oils such as RTEmp from RTE Corporation, polyalphaolefins such as Uniroyal PAO-13-C, and other synthetic esters. corresponding liquids suitable for permanent use. It is also preferred that the permanently used dielectric liquid is characterized by a high boiling point compared to said temporary dielectric solvent, whereby this temporary dielectric solvent can be separated from the permanently used coolant, if necessary, and also by the evaporation of the permanently used liquid. in the event that the transformer tank is damaged.

Although the following substances have been proposed as permanent dielectric fluids, and although they have been used in some cases, they are not as recommended as those permanently used dielectric fluids with high viscosity and high boiling point:

Tetrachlorodiarylmethane with or without isomers of tetrachlorotoluene, freon, halogenated hydrocarbons, tetrachlorethylene, isomers of trichlorobenzene and isomers of tetrachlorobenzene. The isomers of trichlorobenzene, the isomers of tetrachlorobenzene and these mixtures have a high flash point and are similar in physical properties to askarel mixtures, so they are preferably used from these less preferred, permanently used liquids.

The following examples are presented. The following abbreviations are used in these examples: TCB trichlorobenzene TTCB tetrachlorobenzene TCB-mix 30 to 35% by weight of tetrachlorobenzene (TTCB) in trichlorobenzene (TCB) (containing an effective amount of chlorine scavenging, epoxide-based inhibitor PCB or TCB mix in refrigerant, by weight Askarel mixture Askarel type A, 60% by weight Aroclor 1260, 40% by weight TCB

Aroclor 1260 Polychlorinated biphenyl L-305 containing 60% by weight of chlorine is a silicone oil of the formula A above, having a viscosity of 50 centistokes (mm2 / s) at 25 ° C.

"Period" means the period between refrigerant changes. "Part of a cycle" means a portion of a cycle in which the rate of extraction into the refrigerant is clearly different from that occurring in an earlier or later portion of the cycle.

Examples 1, 2, A, B and C

Liquids were drained from each of the four transformers shown in Table I * The transformers were flushed and filled with the refrigerant indicated in section 1 of Table I in each case.

In each of Examples A and 1, the refrigerant in the transformer prior to draining and rinsing was mineral oil (Exxon Inhibited Univolt Oil, Transformer Class) containing PCB at the initial concentration shown in Table I. The transformers 459 and 461 of Examples A and 1 had sometimes been ascarel-filled transformers which had subsequently started to use mineral oil and which contained 9150 and 7800 ppm PCBs, respectively. The transformer 459 of Example A was emptied, spray rinsed with mineral oil, then filled with pure mineral oil. The transformer 461 of Example 1 was emptied, spray rinsed with TCB-Mix, and then filled with pure TCB-Mix. The transformers were allowed to operate at a temperature of about 80 ° C and a length for the periods shown in Table I. In Example 1, the transformer was emptied of TCB-Mix and refilled with TCB-Mix twice more to provide two complete cycles and one progressive cycle.

20,783 6 7

In Examples 2, B, and C, each corresponding transformer was a member of a series of three similar Westinghouse transformers, each with a nominal capacity of 333 KVA. Each of these transformers contained approximately 190 gallons (approximately 719 L) of askarel type A, a refrigerant with a weight ratio of 60/40 between Aroclor 1260 (PCB) and trichlorobenzene. These three transformers were connected to the same three-phase controller and all operated at the same capacity and otherwise under similar conditions, with the exceptions set out below. The transformers of Examples 2, B and C had loads well below their nominal value and had annual average operating temperatures of about 40 ° C, with the exceptions below. Transformer 669 (Example 2) and Transformer 667 (Example C) were emptied of the askarel mixture, spray rinsed twice with TCB-Mix, and then refilled with TCB-Mix. Transformer 668 (Example B) was emptied, spray rinsed twice with silicone oil (L-305), then refilled with silicone oil L-305. The energy level of all transformers was restored and liquid samples were taken at regular intervals to check the PCB content.

All five transformers were allowed to operate and PCB concentrations were determined at the end of the various time intervals shown in Table I, the daily increase in PCB in coolant, expressed in ppm, was calculated and converted to silicone oil refrigerant (if not already based). as shown in Table I.

In Example 2, transformer 669 was allowed to operate for 96 days, after which it was emptied, spray rinsed with TCB-Mix, and refilled with TCB-Mix and continued to operate in section 2, and the same procedures were repeated for transformer in sections 3 and 4. In example 2 in the case of transformer 669, this transformer was emptied at the end of section 4 and spray-purged with TCB-Mix, after which it was filled with TCB-Mix and allowed to continue its electrical operation. Table I shows the daily intervals used in each section, the PCB concentrations at the end of these intervals, the total elution rate and the elution rates converted to ppm / day.

In Example B, transformer 668 was first emptied of the askarel mixture, spray rinsed twice with L-305, and filled with pure L-305. After 390 days, the transformer was re-emptied, spray-flushed with L-305, then filled with pure L-305 and continued in Section 2. Table I lists the intervals measured in days, PCB concentrations at the end of the intervals, and total elution rates in ppm, respectively. / day.

Transformer 667 (Example C) was first emptied of the askarel mixture, spray rinsed twice with TCB-Mix and filled with TCB-Mix. At the end of 96 days, it was emptied, rinsed with spray and filled with pure TCB-Mix. The following sections are in Table I.

22 78367

W

AZ | ^ Q c - ho \ ^ onm co CN vo ro vo c li cmo es vo mo · τ ov vo • r coroto (0 1) φ (Q ^ ^ * k ^ ^ vv ^ k, ^ «v • H ·· 3 Ji VO VO O (N m ro ro O c

§M .u ρ (N (N ro H H

O 0)> «N

on 3 \ I Ui g se & s P (0 H 0) di tn o o o o o o o o o o o in ro I 3 <l in (N rj <oo o o oo ro σν o vchS.

m d E vo oo H ro o vo oo ro uo rr * r c □ O B. d ro

ft H fl H H

S δ • hq, in oo »t ro o vo H m v co tn ro o tn O rt vo vo h in ov vo n ov ro h h m

4J 3H H N H (Ί (N f i — t (N <N

(0 (0 • ro flj O • h p 3 h o oi o m oo m o o vo h n vf o tn ro o; fl 5 h cn vo vo m uv vo n σν hh n

>> frf iH i — I (N (N ι-l (V | N

ä Π): ¾ -h in in mooom in in m in in m iJ-p oo oo oo oo rr (N m oo oo coooooooco> 0> 1> i> 1> o • ro · η · η · η · Η -— 'rH r — lr — tr — tf — t Φ -Ό: 0: θ: 0: 0 C »H · Η« H Ή

> 1 * »H rH rH rH rH rH

J-> C <0 <3 X X X X »H X X X X X X ro <g (d <Q

«P -H · <0 · Η · Η · Η · Η <D · Η · Η · Η · Η · Η · Η <0 <0 <0 <0 1 | o o ρ E E E Ε ρ E E E E E Ε ρ ρ ρ ρ m 4J Ο: Π3 Ο Τ L Τ I ill Τ I Τ Τ I φ φ m φο, ^ ι3: {β Cgggg “ΡΡΡΡΡΡ C | C C Γ0% tn 3 3 • wo o o d · η o o in

3 10. O O H

± £ m 4J E p- o σν H ö -h B o <& dd vo to to to to to to to wwww www 0 0 OO 0 0 0 i — l (Ti (Tl VO · * (D · · lO · · · * 3 * rH (N kO rH <N rH (N ro (0 rd ^ 3 lO ^ * 3 (0 »i ^« 3 3 · ΠΗ rH CN CO 3 · πΗ hn ro ^ id α -ΠΗ HHN 3 > 5 QQQ oi J3qOQQOQ4J SOOOO 2 5 cwwwwm cwwwwwwtt) cwww tn m M U3 3 Ji. *. *. * -Ro 3. *. *. *. *. *. * - ro 3 JC J *. *. * · (- <3 <d id ro id 3 fO ro <0 ro <0 ro 3 id id <0 id O 2 * 3 »3 Ό» 3 2 * 3 * 3> 3 * 3> 3> 3 2 * 3 * 3> 3 »3 s .S.

S ω 8 H CM <23 7 8 3 6 7 I c ^ o c ψ c go oo go vo co ** m go m

η Φ ΦΑΙ OS m r-i * 00 in 1 ^ Π IN

ή ** 3 AND «« «h« «^ ^« I

SID-UU GO GO * 3 * ^ liUOOIVflH

OW> Ο <H NΙΛ H

is 3 \ GO

JA I U E

It & B

• *

4J (0 • H oS

Qt W _ ooo ooooooo 13¾. mom oooocooovo

ffl ag ID ID OlOMICmH

S31 »a aaH

S3 • hq co cm oo oisrimieooo no cm O 'co m o' m cm go O 'go D drl GO ^ 1 · r— (CM GO GO' S *

Co (0 • ΓΟ -X

(0 O • H l-i 3 rH O -X OOOCN oovo — imvoo: <0 3 <H CM O 'm O' Ό CM GO O '

>> (0 GO rH CM GO GO

m y L co σ

Oi CO · »O

Eh y o h m oommmom n *: ¾ -h o ** (0 oo «f ti · m oo oo ^ j · oo · M 4- *>.

t

O

CM

^ 8 3 I

> 1 · Η ψ r- »·· H c m γη ι-ιχκχχκ δ: <o m

4J · Η »α) OI 'H · Η -H · Η · Η 5 H O

V 4J £ m m m m ygEgEE-mm ΉΗ co OM <0000 3 1 1 I I TOO co: cd i

> i 3: <o .x co eo co jc α ο ώ n m n n ι; ΐο J

Sat: S Λ Ai 8BSBBSAA St s • ή jc a 4J g * E * 4-1 (0: SJ M n 3 H H 3 3 rH u • n o o co oi ra

Oi -H O O 3> O,

3 10 0 O HH

· * P) -K E o o <y cq co HUH H o o H4J n <ft a a m m js en • M * Qi 53 5

rH 4J

n m co n; co · ή 8 ra ra m oj ra> O O o m. * n 3 S · · m · · h δΓ i3

Ji VO rH CM VO r — I CM rH

a: 0 <0 3

CO «CO ** 3 CO» · ». 3 C H 4J

Ή 3 · ΠΗ iH CM 3 TlH rH CM GO TG m VO 3 CC 4-1 3 w <0 cq .x en .x: <S -d Φ 3

> 4JQQQ4J4JQQQOQQQ4J: cg g C 4J

3 e ra ra ra co cmrarararararaco ra -he 4-> m UJ 34C4C ^ -n: (0 Q 3 3 _ 5 (0 (0 (0 3 (0 (0 (0 (0 (3 (0 (0 a- ö 5 3 o sr> t -) (- 3 2 (313 (313 (3 (3 (3 ra ^ E &

^ Il II II

2 -S · h co (0 m o «o S.

E- · cqcffl O> coti 24 78367

In Example A, for transformer 459, the extraction rate during the second part of the first period had decreased to 0.36 ppm / day as the average PCB extraction rate. This is less than 0.55 ppm / day, ie less than 50 ppm in 90 days, which must be reached in order to reclassify the transformer as PCB-free. Thus, on day 220, the mineral oil was drained from the transformer and replaced with a permanently used refrigerant in L-305 silicone oil. As can be seen from Table I, the total amount of PCBs removed during Period I corresponded to 475 ppm of the volume of refrigerant contained in this transformer. This is less than the 1220 ppm removed at the same time with TCB-Mix from transformer 461 of the same size and type, indicating that the mineral oil was not as effective as the extractant as TCB-mix.

For comparison, transformer 461 was extracted with TCB-Mix. On day 68, the transformer was emptied of TCB solvent, which was fed back to this same transformer. On day 165, the TCB mix was drained from the transformer and replaced with pure TCB mix. The overall rate for cycle 3 was 1.64 ppm per day, but this rate decreased during the cycle and after 245 days the PCB elution rate had decreased to 0.05 ppm / day, well below the target value of 0.55 ppm / day. Thus, the coolant can now be replaced with a permanently used coolant in silicone oil.

Transformer 669 (Example 2) was initially filled with an askarel mixture. Thus, it contained much more extractable PCB in its insulating material than transformer 461 (Example 1). As a result, extraction to an acceptable level required significantly more cycles and a significantly longer time. When the extraction rate falls below the target rate of 50 ppm in 90 days, the TCB mix can be replaced with silicone. The velocity values shown in the last column of Table I show a steady decrease over time, and the target speed should be reached after approximately 600 days. It should be noted that transformers 667, 668, and 669 were assumed to be the most difficult to extract. They are helically wound transformers where the thickness of the paper insulation and hence the length of the diffusion path may be several inches. In contrast, many transformers are of the "pan-cake" type in construction, where the length of the diffusion path is less than an inch.

The transformers 668 and 667 of Examples B and C are comparable examples because in Example B silicone oil was originally used as the coolant and in Example C the transition to silicone oil occurred before the elution rate was reduced to 0.55 ppm PCB per day.

Transformers 667 and 669 were originally filled with ascarel mixture, which was replaced with TCB-Mix, while transformer 668 was originally filled with askarel mixture, which was replaced with L-305 silicone oil. Figure 1 compares the PCB analysis values obtained in the first section of these three transformers. These results have been converted to the actual total amount of PCB removed in grams. Approximately 60,000 to 70,000 grams of PCBs were rapidly removed (during the first 28 days), but PCB removal was then significantly slower and these rates are shown in Figure 1 by straight lines passing through the measuring points. It is assumed that most of the PCB retained by the relatively loose insulating material can be easily extracted regardless of the solvent, but it is also assumed that the process speed is limited by the PCB retained by the tightly wrapped paper and the insulation board, in which case the efficiencies of the eluents vary. This difference is shown in Figure 1. Although the measurement points are somewhat scattered due to the difficulties associated with accurate PCB analysis, it is nevertheless obvious that the amount of PCB that can be removed with TCB-Mix in 60 days requires 400 days to remove with silicone. Comparing the actions of the straight angles of Figure 1, it can be seen that the TCB mix is about 8.5 to 9.0 times more effective extractant than L-305 silicone oil. The key point of the present invention is that this efficiency ratio is so high. Thus, a process which, when performed with silicone alone, could take 5 to 10 years, can be performed in a much shorter time using, for example, TCB mix as a temporary coolant.

Figure 2 shows the effect of transformer temperature on PCB elution rate. The temperature of both transformer 667 and transformer 669 was about 40 ° C during the first period. They were changed to pure TCB mix after 96 days. As winter approached and the transformers were not operated at the capacity needed to keep their temperature high, it was assumed that low temperatures would interfere with PCB extraction. As a result, transformer 667 was artificially heated by conducting heat to the cooling fins. For comparison, transformer 669 was not heated. The average temperature of transformer 667 was about 55 ° C, and that of transformer 669 averaged about 23 ° C. The scatter of the measuring points is again quite large, but it is clear that PCB eluted from the warm transformer 1.6 times faster than from the cold transformer. This 1.6-fold velocity may not be linear, and thus the increase in velocity may not be as noticeable at higher temperatures. Subsequent extraction was performed at 85 ° C, if possible, and thus most of the extraction shown in Table I took place at this higher temperature.

As further examples, the following illustrative cases of Examples 3 to 5 are presented. Although they do not show results from actual transformers, they are based on the expected performance of the process of this invention under the conditions set forth below for each example, applied to transformers from which PCB elution by the method of this invention is relatively easier than in Examples 1, 2, A, B and C of the transformers used.

Il 2? 7 8 3 6 7

In each of Examples 3 to 5, a transformer with a capacity of 200 gallons (about 757 L) of liquid volume is used, the cellulosic material inside, i.e., paper insulating the windings, retains up to 6 gallons (about 22.7 L) of liquid, and which contains 200 gallons (about 757 l), more or less of a 50% (500,000 ppm) ascarel mixture of PCBs, with the exception of Example 4, where the transformer contains about 200 gallons (about 757 l) of mineral oil with PCB concentration is 10,000 ppm.

Figures 3 to 5 show the concentration of PCB in the transient di-electric liquid (TCB) used in the transformer in ppm, plotted on a vertical logarithmic scale as a function of days (extraction time), and graphically show the expected results of the present invention.

Example 3

In Example 3, the energy level of the transformer is first set to zero. It is then emptied of its askarel mixture, which is eventually disposed of in an approved manner. The transformer is flushed with a small amount (e.g., 25 gallons, about 94.6 L) of trichlorobenzene to reduce the amount of ascarel remaining in the free liquid system to 0.5 percent of its original value. The system is then inspected for logically leaking bushings or other physical problems that may now require repair.

The transformer is then filled with 200 gallons (approximately 757 L) of trichlorobenzene (TCB) (or alternatively a mixture of trichlorobenzene and tetrachlorobenzene), sealed, and restored to its energy level after proper inspection. Since the rinsing is not complete, the initial level of PCB in the new transformer liquid is assumed to be 2500 ppm, i.e. 0.5% of the original PCB levels. It is assumed that the PCB retained by the cellulosic materials is recovered per day at a rate ranging from 0.001 to 0.01%. Although these values may seem arbitrary, they can potentially be achieved in easily extractable transformers, and higher or lower rates affect only the total time required to perform the extraction, and not the basic method itself. In the representation of Figure 3, the curve drawn at the top shows the PCB concentration (on a logarithmic scale) that can be expected to be found in the transformer fluid as a function of time. In actual commercial applications of this method, it would not be necessary to determine all of these concentrations. However, it may be necessary to recover the old liquid to be replaced and determine its PCB content. Figure 3 shows this in open circles. Although the exact length of the extraction cycles is arbitrary, experience with a particular type of transformer shows the most practical lengths of these cycles in terms of the total process time as well as the total number of fluid changes. In this example, extraction periods of 60 days are used.

After 60 days, the energy level of the transformer is brought to zero again, the liquid is drained out of the transformer and the liquid is sampled for analysis. The system can be re-flushed with approximately 25 gallons (approximately 94.6 L) of TCB, and the flushing fluid and other spent fluid are taken to a facility where TCB can be recovered by distillation (and residual PCBs are disposed of by appropriately approved EPA methods. ).

The transformer is refilled with TCB, and this time the expected initial concentration of PCB is about 83 ppm (due to the residual liquid). The expected PCB concentration again follows the upper curve of Figure 3 for the next 60 (... 120) days, after which the TCB in the transformer is replaced as before, with one exception. As the PCB content of the effluent TCB is less than the initial value found in the first filling, the effluent does not need to be sent to a distillery for separation but can instead be used as the first filling liquid in the second PCB transformer to convert the PCB. to a space that does not contain. This saves valuable distillation time and energy, as well as transportation or handling costs.

This refill process is repeated one more time. Table II lists the assumed analytical results plotted in circles in the graphical representation of Figure three. It is clear from Table II and Figure 3 that the PCB content in the fourth charge does not rise above 50 ppm, which is the limit value for PCB-free transformers in accordance with U.S. government regulations. Thus, after 180 days, the transformer is filled with a permanently used liquid which is silicone oil, for example L-305. The PCB value that was expected to be reached after the next 60 days (240 days) is only 16 ppm, and this concentration is still expected to be only 18 ppm after the 90 days (270 days total) required by US regulations. Therefore, this transformer can be classified as a PCB-free transformer.

Table II

Elapsed days PCB concentration, ppm_

Empty Original liquid concentration when refilling 0 500000 2500 60 16600 83 120 896 4 180 Refilling 101 with silicone <1 240 (16) not emptied 270 (18) not emptied »30 7836 7

Example 4

In Example 4, extraction cycles of 60 days are used, but flushing of residual liquid out of the transformer was omitted. It is assumed that 98% of the liquid can be suitably drained out of the transformer, leaving 2% of the liquid. In this case, the initial concentrations are 2% of these freshly drained liquids instead of the 0.5 found in Example 3. The procedures of Example 3 are repeated in this example.

The expected results in Example 4 are shown in Table III, as well as in the graphical representation of Figure 4. It should be noted that the target is still being reached and that the system can be refilled with silicone or other permanently used oil on the 180th day. The absence of a very efficient out-flushing of the residual liquid is expected to lead to a slightly increased PCB content in the final liquid, but this fact does not substantially change the fact that the target PCB-free transformer has been reached.

Table III

Days elapsed PCB content, ppm_ emptied Initial liquid concentration during refilling 0 500000 10000 60 23900 480 120 1440 30 180 Refilling 145 with silicone 3 240 (21) not emptied 270 (21) not emptied

II

3i 78 3 67

Example 5

Based on the shape of the concentration curves shown in Figures 3 and 4, it might be assumed that fluid changes should be performed more frequently, e.g., every 30 days, instead of 60-day periods. Example 5 is identical to Example 4, except that the extraction periods used are 30 days. The expected analytical results are shown in Table IV and the corresponding curves are shown in Figure 5. The trend is evident from the graphs in Figure 5, with the first refill achieving almost as good a concentration reduction as in Example 4, but then these reductions begin to level off. The sixth refilling can be performed with a permanently usable fluid, and time has thus been saved slightly at the expense of two additional refills with TCB. This example illustrates the relationship between the time required and the number of refills, and the choice in favor of either factor depends on which of these factors is considered more valuable in each case under consideration.

Table IV

Days elapsed PCB content, ppm_ emptied Initial liquid concentration during refilling 0 500000 10 000 30 15800 316 60 1260 25 90 310 6 120 120 3 150 Refilling 50 with silicone 1 180 (21) not emptied 240 (32) not emptied 32 783 6 The use of this invention is not limited to transformers, but can be used in any electrical induction device using a dielectric coolant, such as electromagnets, liquid-cooled electric motors, and capacitors, for example, in directional devices used in fluorescent lights.

Claims (19)

A method of replacing PCB-containing refrigerant in an electrical induction device with a tank containing the refrigerant and an electrical winding and porous solid cellulose-based electrical insulation immersed in the PCB-containing refrigerant, with substantially PCB-free dielectric permanent refrigerant having a high boiling point to convert said electrical device into a seed in which the elution of the PCB to the coolant Mr below the maximum permissible elution rate to the coolant of an electrical device classified as non-PCB-containing, wherein the porous solid the electrical insulation is impregnated with the PCB-containing refrigerant, characterized in that the process comprises the following steps: (a) the PCB-containing refrigerant is drained from the tank to remove most of the PCB-containing refrigerant therein; (b) the tank is filled with an interim dielectric coolant which mixes with the PCB and has a sufficiently low viscosity to circulate inside the tank and penetrate the gaps in the porous solid electrical insulation and which can be easily separated from the PCB; (c) the electrical induction device was allowed to function electrically and the electrical function was continued for a period sufficient to elute PCBs contained in the PCB-containing refrigerant impregnated in the porous solid insulation from the insulation to the interim dielectric coolant; (d) after which the interim dielectric coolant containing the eluted PCB is drained from the tank; (e) the cycle formed by steps (b), (c) and (d) is repeated if the rate at which the PCB elutes to the interim dielectric coolant exceeds 0.55 ppm PCB per day, based on the weight of the permanent dielectric refrigerant; and (f) the tank is filled with a substantially PCB-free permeable refrigerant selected from the group comprising silicone oils, liquid synthetic esters, poly-alpha-olefin oils and hydrocarbon oils having a high boiling point and high viscosity, whereby the electrical device may be reclassified as a non-PCB containing device. Process according to Claim 1, characterized in that the interim dielectric coolant is trichlorobenzene, tetrachlorobenzene or a mixture thereof and that the permanent coolant is a dielectric coolant of silicone oil. Method according to claim 1, characterized in that each step (c) is continued between 20 days and one year. Method according to claim 1, characterized in that each step (c) is continued between 30 and 120 days. Process according to claim 1, characterized in that each step (c) is continued between 45 and 90 days. Method according to claim 1, characterized in that while step (d) of the preceding cycle and step (b) of the following cycle is performed, the interim cooling coolant is drained from the top of the tank while fresh cooled interim cooling liquid is measured from the bottom of the tank. the electrical function of the device continues. Process according to claim 1, characterized in that steps (d) and (f) are carried out by feeding the PCB-free permanent coolant into the bottom of the tank while removing the interim dielectric coolant in the tank from the top of the tank, and the electrical function of the appliance continues. Method according to claim 1, characterized in that the tank is provided with heat insulation to raise the operating temperature of the interim dielectric coolant contained therein during each step (c) while allowing the electrical induction device to operate electrically. Method according to Claim 1, characterized in that the interim dielectric coolant in the tank is heated during step (c) while allowing the electrical induction device to operate electrically. 10. A process according to claim 1 characterized in that the interim dielectric coolant is removed from the tank during step (c), heated and fed back to the tank while maintaining a sufficient amount of interim dielectric coolant in the tank and allowing the electrical induction device to operate electrically. 11. A method according to claim 1, characterized in that the interim dielectric coolant is more volatile than the PCB and is separated from the contained PCB by distilling away the interim dielectric coolant. 12. A method according to claim 1, characterized in that the interim dielectric coolant containing PCB eluted from the solid insulation is drained from the tank as a slip current while allowing the electric induction device to function electrically and to supply fresh electrolyte; imistic dielectric coolant in an amount substantially equal to the amount of PCB-containing interim dielectric fluid drained away as a propeller stream. Process according to claim 1, characterized in that the tank is flushed with a solvent for PCB after step (a) and before step (b). 14. A method according to claim 13, characterized in