Document EMwj7mEExmO4nNMyp6EzyZgg

A HEU ESTHETIC IMPREGNATE FOR EIC-E TENSION HOLLOW CORE CABLES ... ... . .. ' U. Pelagatti . . . .-s.-'. . . Centro Pirelli Milano, Italy The "paper/thin mineral oil" composite dielectric used as insulation in high voltage cables dates back to 1924 when the first holler;; core cable (a 130 kv pioneer) was experimentally laid at Brugherio (Milan, Italy). Since then, thanks to continuous investigations and growing experience, each single characteristic of this impregnant has been brought more and more into focus, eventually reaching the at present strictly defined features . Tee general properties which characterize the presently "used thin mineral oils can be regarded as the best obtainable with petroleum derivatives; as a matter of face mineral oils are made up of complex mixtures of many different hydrocarbons some of which behave satisfactorily whilst others, unfortunately very difficult to re move, show 'undesirable properties. An additional disadvantage which is comecued with mineral oils and is related to their natural origin, is the difficulty in obtaining them with absolutely consistent characteristics. To overcome this situation and reach the desirable target of better impreg- nants, the ideal solution could lie in synthetic products made up of those "tailored molecules" which have already been suggested for solid-insulating materials. Such a solution, however, -would require someone in the chemical industry -willing to investi gate and perfect new products without a reasonable certainty of very large future consumptions; a very serious limitation to start with. ' For this and other reasons our attention was thus oriented towards the great bulk of synthetics, trying to see if there were existing "tailored" materials able to fit the new situation. _2 an the course of a search which started some six years ago , an interesting material was selected: dodecylbenzene. Such a material, produced by the big chemic:. companies as an intermediate to detergents of world-wide use, gave promise that ti de sired impregnant was near to being identified. - Tnis assumption was based on the following theoretical considerations: technical dodecylbenzene, even if not consisting only of one single chemical bu f - rather a large mixture of monoalkylbenzenes, is nevertheless made up of homolc; s molecules which should show almost the same behavior; thus the drawback of the _.e- sirable components present in natural products would be avoided. With regain. : ..a 'esirable constant composition this end also should be attained, as dcdecylb .z; j is DSW 201956 Paper presented at the 1963 Conference on.Electrical Insulation STLCOPCB4060023 obtained through well-defined and easily controlled processes starting from pure mate rials (benzene and propylene). With regard to the behavior connected with its molecular structure, the expected results were thought to be the following: the benzene ring should ensure: l) exceptional stability to both thermal and electrical stressj and 2) an extremely good gas absorption property^. In its turn, the saturated aliphatic chain was likely to provide, due to the'sufficient number (twelve) of carbon atoms, physical properties (such as viscosity, etc.) very close to those of the best thin mineral oils. ' Dodecylbenzenes of technical grade were then tested and after exhaustive experiments the chemists had the satisfaction of realizing that all was behaving to their expectations. The general properties of a dodecylbenzene of electrical grade are given in Table I compared with those of widely used mineral oils (normal and low viscosity grades) which can be considered among the best available at present, Table I - Comparison of Some Properties of Dodecylbenzene (D.B.D.) and Thin Mineral Gils, Normal Viscosity (N.V.) and Low Viscosity (L.V.) G^TcbG.0 r]t---Qrr, PROCEDURE UNIT D.D.B. N.V. L.V. Specific gravity at 15C Kinematic viscosity at 20cC Pour point ASTM D 1298-55 gr/ml astm D 145-61 cSt _ ASTM D 97-57 c 0.8769 0.8920 ' 12.5 ; 26.6 -63 ' -47 0.8856 14 -63 Plash Point (Cleveland O.C.) ASTM D 92-57 Autogenous ignition temperature ASTM D 286-58T C C ' 135 . 440 155 ' 135 260 250 Power Factor at 100C Pirelli R.T.127 original '. #' 40.3 4 0.3 40.3 after ageing* dP 40.3 ^0.6 40.6 Resistivity at 100C original after ageing* . Pirelli R.T.127- MD cm >25*107 ^5*107 >5*107 II . .. .. >15 .io7 >i.5.io7:>1.5.107 Gas absorption after 300 mins. Pirelli L.R.356i ml at 80C under a 16 kv stress 2.3 1.4 0.5 *Ageing: 24 hours at 100C in the presence of both air and metallic copper. Prom the examination of these data, conclusions are that dodecylbenzene not only has characteristics at least equal to those of mineral oils but that it is definitely very superior in two important' instances, showing: 1) the expected higher gas absorption under electric stress and 2) a higher autogenous ignition temperature which leads to higher resistance to both thermal and electrical cracking. Moreover its low viscosity, even lower than that of the so-called low viscosity oils, is an additional point in its favor. DSW 201957 STLCOPCB4060024 k2 GAS A6SCK?TCN-E'KatfCN CHWSTEfcSTlCS iMACrvoui** a**t tf *mxo Miiut ^LucnM) GAS HYDROGEN CAS ASSOWO^CVOLUTiOl^ CHAftACTERSTCS ikAi**cavm b*sci m **aq awc GAS ASSCRPTION EVOLUTiCM CK.ARACTEKIST.CS TCSTIkS TEMPSRATUftettO(T-APPUED VOLTAGE: W kV fWCLU GAS: ARGDM As regards the higher gas ab sorption, some results are given in Figures 1 to 3 which are self-explanatory. In such tests comparison is made between the dodecylbenzene and the thin mineral oil (normal viscosity grade) the charac teristics of which are reported in Table I. These tests were carried out following the Pirelli method" with the only exception that the testing period was extended to 20 hours. A perculiar and unexpected prop erty of dodecylbenzene is that not only does it absorb gas at temperatures and/or for testing periods at which mineral oils evolve, but that its capability to absorb takes place, after an evolving transient, even with water vapor (see Figure 4), a phenomenon never observed with mineral oils of current use. 3 Xt has to be mentioned that even if the reported values could be quantitatively incorrect, because of some difficulties encountered' in performing this last test, they are nevertheless qualitatively exact, as confirmed by other tests described farther on. ' In addition to the laboratory investigations, technological experiments which took some five years to complete, fully confirmed the previous findings and our expectations. As a direct consequence dodecylbenzene has, since the beginning of this OSVM 201958 STLCOPCB4060025 i ^0 GAS AEStlFTtCJ IVCiiT::;: C-LT-^'JSTsCS TESTiKS 7SK?3!OT52:ttOd.APPUSO VOUACSstS fcV (iV-itU ASTWXi} 0;&v/;iixv;rc,7: 31s. 5 year, completely replaced thin mineral cil in the 0.?. cable manufacture at the plant of the firm with which the author is associated. One of the more .interesting ex periments carried out on actual cables will be briefly described hereunder. Two lengths of a hollow core cable de- . signed for a nominal voltage of 50 kv, one impregnated with conventional thin ' mineral oil and the other with dodecylben- sene, were subjected to a ho Zcv voltage (applied between the conductor and the outer screen and corresponding to a maxi mum gradient of 10 kv/mm) whilst kept at 120C by means of external heating. The cable impregnated with mineral oil shoved after about 500, days an increase of power factor which con tinuously rose' (see Figure p) -until too high to be measured. On the 640th day the cable failed. In the meantime a great gas evolution -was observed. Such gases collected and analyzed by means of gas chromatography technique were found to be: c% (84.6$) Eg (7.8$); CO (5-5$); 0^ (l.4$)j CgHj, (0.7$). 'When the cable was dismantled for inspection, the insu lation was found strongly altered, resambling a unique compact material in which the paper tapes were- unvindable and extremely brittle. The impregnant, extracted by solvent, was a very dark liquid with an extremely high viscosity (55,000 cSt at 40c)j the gaps were com pletely filled -with "X" Waxes, an un- ' meltable and insoluble material origi nating from mineral oil when its altera tion takes place under intense electrical. DSW 201959 STLCOPCB4060026 stress. In addition a great amount of water, given off by the cellulose on account of its thermal degradation, was found in the oil feeder. ' After a same period the cable impregnated with dodecylbenzene was perfectly efficient; the power factor was still quite good (see Figure 5), and no sympton of gas evolution was observed. It is worth stressing the meaning of this finding, i.e., dodecylbenzene not only did not evolve gas but did absorb those gases (CO - C0o - E.O) . which originate from the thermal deterioration of the paper. Tne test was thus protracted up to 9^1 days before dismantling the cable for inspection. Tne insulation was practically unaltered and the paper tapes (which of course underwent a certain embrittlement) could be easily unwound; no traces of "X Waxes" were found in the gaps. As regards the impregnant, dodecylbenzene shewed un altered physical properties; a slight increase in the power factor was noticed and a small; chemical alteration (ascertained by I. R. spectrography) was found. This latter alteration is certainly to be related to seme modifications which took place on ac count of the absorption of the products evolved by the cellulose during its degrada tion, In addition, no water was noticed in the oil reservoir, a confirmation of its capability to absorb water-vapor as already ascertained by the gassing test previously mentioned. With regard to the dodecylbenzene used in the tests and now introduced in actual production, it is composed, to an extent of 99-5$ of a mixture of moncalkylben- zenes whose aliphatic chain (which is saturated and shows a prevailing branched structure) is in the range between C^q and inclusive, Tne main portion, over 65$, consists of whilst C-, and are 12$ and 18$ respectively. The mean molecular " weight corresponds to that of dodecylbenzene; for this reason this term is the one broadly used in defining such a product. As regards the whole structure, the aromatic ring can, of course, be linked to the main aliphatic chain at different positions: thus either the head or other linking positions are likely to occur. Dodecylbenzene, obviously is not the only alkylbenzene which can be advan tageously used as cable impregnant. Products whose bulk is characterized by aliphatic chains higher, than C^, are generally not interesting as they show too high an increase in viscosity without adding any other advantage. On the contrary, for special applications (as for example, lengths of O.P, cables which for particular reasons could require very widely spaced oil feeders) those products whose bulk presents chains in the range from to C^q can be very interesting as they possess an extremely low viscosity (about 5 cSt at 20C for C-. Q). Products with a chain smaller than C^q are however, not advisable as their flash point decreases very rapidly (ll8C open cup for C^q) and danger of fire can arise in jointing operations. .References 1, "Insulating oils for electric cables" by C. Palan&ri and U. Pelagatti - Presented ^ at the 195k Convention of the Associazione Elettrotecnica - Eellagio, Oct. 3"10,19":^. 2. U. Pelagatti - "Prodotto per L'impregnazione dei cavi ad alta tensione e cavi ccsi ottenuti" - Italian Patent Ko. 59^-75, ktg. 2, 1953 . DSW 201960 STLCOPCB4060027