Cabling for power transmission and distribution must give reliable continuous service over many years. In addition to maintaining electrical integrity it may be required to operate at relatively elevated temperatures and to continue providing service under fire conditions. Many of the materials used in cable manufacture are polymeric, and some are cross-linked to permit service at elevated temperature and to provide required electrical and mechanical strength.

Thermal analysis offers the possibility of investigating ageing properties, cross-linking processes, service conditions and behaviour during the burning process, as well as being an invaluable tool for the analysis of materials to verify composition of particular formulations for special applications. In addition, it is possible to make use of thermograms of unknown materials for identification purposes. Although the emphasis here is on applications to electrical cabling, it is envisaged that they would be readily transferred to other industries engaged in the manufacture of polymeric products.

• Oxidation studies (OIT test)

For polymeric products required to operate at temperatures appreciably higher than ambient, degradation processes, whether purely thermal or as a result of accelerated oxidation, are important and can significantly affect life expectancy if appropriate grades, adequately protected by suitable anti-degradants, are not selected. Oven ageing trials, with periodic mechanical testing of test pieces, are good predictors of performance but are necessarily slow. The measurement of oxidation onset by differential scanning calorimetry (DSC), at higher temperatures than those used for oven ageing trials, is useful in some cases.

The basis of the measurement is that of holding a sample under isothermal conditions under an air or oxygen atmosphere, the sample being contained in a pan which is open to the purge gas, and to measure the time taken for the oxidation exotherm to appear in the DSC trace. The experiment is repeated at several temperatures so that data from different samples can be compared, either directly or by using the reciprocal of the OIT as a pseudo rate and plotting the logarithm of the pseudo rate against reciprocal absolute temperature in an Arrhenius plot.

This approach was adopted for a study of polypropylene used as the filler in the interstices between the cores of some types of three-phase power distribution cables. The cores consisted of a conductor with polymeric insulation and extruded screens which were in turn screened with copper tape. Polypropylene twine is used to pack the interstices as the cable is laid up and then the outer parts of the cable are applied later. Thus, the polypropylene is in direct contact with the copper tape.

Copper is well known as a catalyst of the autocatalytic oxidation of polyolefins, catalysing hydroperoxide breakdown and chain scission; polypropylene is particularly susceptible to this degradation route and it was therefore vital that grades were selected which contained a suitable metal deactivator as well as adequate antioxidant. OIT was measured at temperatures in the range 180–210◦C for a range of samples, either in aluminium pans alone or with 3% of metallic copper also added to the pan. Sample size was kept as close as possible to 10 mg for the sake of consistency between tests. Samples were heated to the test temperature under nitrogen and the purge changed to oxygen at the start of the isothermal part of the measurement.

In addition, some samples were reformed by pressing into sheets; small samples were then cut for ageing in air-circulated ovens at various temperatures, between 120 and 150◦C, both in and out of contact with copper. These oven ageing trials were conducted in order to validate the extrapolation to higher temperatures so that future material selection based solely on DSC measurement of OIT could be validated. The time to failure in the oven ageing trials was assessed by a simple three-point bending test, since it was found that the polypropylene lost all mechanical integrity rather rapidly once the oxidation process was under way. The time to failure was reported as OIT at the lower temperatures and these results incorporated into the Arrhenius plots.

Interestingly, the extrapolation from the DSC measurement temperature range to that of the oven ageing trials proved successful, and useful Arrhenius plots were obtained. The initial trial permitted the selection of suitable grades of polypropylene, with adequate resistance to copper-catalysed oxidation, for the designated use with confidence. Subsequently, OIT studies were used as a screening process for other candidate grades for this application, thus saving considerably time and effort which might have been spent on oven ageing trials.

Other materials such as certain types of cable sheath also have to give specified levels of performance in oven ageing trials. It was decided to experiment with some fire-retardant cable sheathing in purely oxidative studies since it was known that the oven ageing trials for this material gave very convincing Arrhenius plots. The DSC measurements were carried out in the range 190–220◦C, and very consistent replicates and good Arrhenius plots were obtained. Unfortunately, the two sets of Arrhenius data had different slopes, indicating that the reaction mechanisms for degradation at the higher temperature range differed from those at the lower oven ageing range, unlike the oxidation of polypropylene which had followed the same reaction mechanism over a very wide range of temperature.

This second trial demonstrated the necessity of making adequate measurements at oven ageing temperature ranges to test the validity of using data obtained at higher temperatures in the DSC for extrapolation purposes. It is clearly vital to ensure that the reaction mechanism remains the same over the whole range of interest.

• Thermal history studies

When power cables are energised, some heating takes place as a result of resistive and other effects. Cables are therefore designed with these effects in mind and such that they can accommodate the rated load for the circuit without overheating and consequent accelerated ageing. It is useful, therefore, to have some means of investigating the thermal history of cable insulation.

Much cable is insulated with cross-linked polyethylene and in this case the crystalline melting envelope gives an opportunity to investigate past heating effects. When polyethylene is heated at a temperature within the crystalline melting envelope, reorganization occurs for that part of the structure which can melt at the temperature achieved. On cooling, this new structure is retained. As a result, the shape of the crystalline melting envelope, obtained in a DSC experiment, has a pronounced irregularity at the final highest temperature of heating, when compared to a sample cooled from a temperature above the crystalline melting point. This work was based on a report by ERA Technology.

It thus becomes possible to study cross-linked polyethylene insulations with a view to investigating possible past overheating events in service and also, in the case of very large cables with thick insulation, to investigate temperature gradients across the thickness of the insulation by making measurements at intervals across the thickness. Data are required from the cable in the as-manufactured condition and in the fully recrystallised state. Figure 1 shows an example.

Figure 1 Crystalline melting endotherms of cross-linked polyethylene insulation as manufactured (1) and after recrystallisation at 140◦C (2).

It will be noted that there is a small event between 50 and 60◦C. This is a result of a thermal cycle in manufacturing to remove by-products of the cross-linking process. Curves for insulation, which have been oven aged at temperatures that might be expected in normal light service and under conditions of overheating, are shown in Figure 2.

Figure 2 Crystalline melting endotherms for cross-linked polyethylene which had been oven aged at 78◦C (1) and 102◦C (2).

These measurements were carried out under nitrogen purging, at a heating rate of 10 K/min. Samples were approximately 7 mg in weight and held in crimped (unsealed) pans with unperforated lids. A previously measured baseline, obtained with empty pans in both reference and sample positions, was subtracted from the raw data.

It is notable how closely the onsets of the crystalline melting event match the measured oven ageing temperatures, and it is also worth noticing that the previous thermal history of the manufacturing cycle has, in all cases where a subsequent heating had taken place, been recrystallised out. Thus, if two thermal histories were observed, it would be possible to conclude that the lower temperature event occurred after the higher. If no thermal history was observed at all, it would be reasonable to conclude that the insulation had fully recrystallized because the thermal history of the manufacturing process would have disappeared (see Figure 1). Since the normal service rating for cross-linked polyethylene-insulated cables is designed to give a maximum temperature of 85–90◦C at the conductor, depending on design, full recrystallisation would be indicative of a period of being run under appreciable overload conditions, although the actual temperature achieved cannot be measured. In this case it has sometimes been possible to examine the outer cover of the cable, where these are manufactured from medium- or high-density polyethene which both have higher peak temperatures for their crystalline melting points than does the low-density polyethene usually employed for insulation purposes.