Internationally, little is known about the aging characteristics of midspan compression joints on overhead lines. It is widely believed that the joints do not age appreciably, but evidence is emerging that in some cases significant deterioration is occurring. This may result in a risk of joint failure before the conductor reaches the end of its useful life. Further, as current trends to uprate older lines continue, aged joints are having to work ever harder, bringing joint aging characteristics into ever sharper focus.

Hidden Bad Joints When Transpower New Zealand Ltd., Wellington, New Zealand, suffered an unexpected failure of a transmission line midspan joint over an urban area during a sustained fault, the incident provided the initial impetus for a closer look at the condition of the mid-span joints on many of Transpower's transmission lines. Transpower has approximately 60,000 mid-span ACSR joints in service. Although the average joint age is 35 years (some are more than 60 years old) the utility knew almost nothing about the overall condition of these joints.

While Transpower has used thermal imaging equipment for many years to check lines for defective joints, it had become increasingly apparent that this technology was only finding joints that were in the later stages of failure. In other words, the joints were already serious reliability risks. Many other defective and potentially defective joints were not being found.

Evidence showed that the combined effect of cooling from the wind, the heat sink effect of the adjacent conductor, relatively low load currents and perhaps some masking from solar heating was making defective joint detection using thermal imaging difficult. Only in the later stages of the joint life, where the resistance had increased significantly, was a detectable heat signature developed at a particular load. By that stage the joint may either be close to failure or may have reached a high enough resistance that thermal runaway could occur under emergency loading conditions or abruptly under a fault.

For these reasons, Transpower decided to find a more reliable inspection method so it could determine the true condition of the joints on its lines.

Deterioration Processes A literature search, research information from Electricite de France, Paris, France, and Vattenfall AB, Vallingby, Sweden, plus in-house experience indicated that at least six separate deterioration processes may have been at work. Corrosion. Mid-span joints appear homogeneous when sectioned, but actually have micro-voids that admit water, corrosive contaminants and oxygen. The corrosion mechanisms that most affect joints seem to be: - Pitting corrosion--localized corrosion that forms holes in the surface of the unit. - Crevice corrosion--corrosion in confined crevices where acidic build up occurs. - Corrosion (rusting/pitting) of the internal steel sleeve and core wire. - Oxidation of contact surfaces.

Ice jacking. Water wicks into the small voids and expands when frozen, thus slowly jacking the compression joint apart (some joints in which joint sleeves were visibly distorted through ice jacking have recently been replaced on the Transpower network ).

Thermal cycling. The thermal heating and cooling cycles that joints are subjected to under normal operation gradually work to relieve compressive stresses which in turn increase the contact resistance.

Line faults. Fault current can induce a bulk joint temperature rise exceeding 150 degrees C which leads to annealing of the aluminium and even melting of the contact interfaces.

Joint sleeve creep which reduces the mechanical stress of the unit over time (this is of least significance).

All of the above processes increase joint resistance either over the whole joint or on half of the joint. Further, the thermal overheating and subsequent abrupt physical failure of a joint (due to aging) is preceded by a long period of gradual joint resistance increase.

Transpower decided that the only reliable method of determining the condition of a joint was to measure its resistance. Resistance measurement was the only known technology that was responsive to all the known deterioration processes. In addition, the technique was sensitive enough to provide consistently reliable information independent of all external factors early enough in the deterioration process to be able to predict future joint life.

Measurement Project Initiated In 1995 Transpower decided to launch a project to obtain joint condition information on a representative sample of its 220-kV and 110-kV network. The prime project objectives were to: - Prove and refine a live-line measurement technology. - Develop the tooling to enable measurements to be taken from helicopters in rural regions or from bucket trucks in urban locations. - Introduce and develop helicopter live working techniques in New Zealand. - Obtain sample joint data from a range of Transpower lines and conductor types. - Develop a feel for the aging characteristics of Transpower's joint population, including the effect of conductor type and atmospheric environments, so that projections of future joint life could be based on real data. - Enable joint management to be incorporated into long-term asset management strategies.

International tenders for the project were called in 1996. The project called for measuring approximately 1000 energized joints to better than 10% accuracy on a variety of lines. Both halves of the joints were to be separately measured.

The contract was awarded to a joint venture partnership consisting of Mainpower New Zealand Ltd., a local company, and Aeropower Australia, a live-line helicopter operator.

Development of Measuring Device Mainpower embarked on the development of a joint resistance measuring device that could easily be placed directly on the "live" conductor joint using bare hand techniques from a skid-mounted helicopter platform. The device transmitted data via a radio link to a PC located in the helicopter.

The research and development phases of the device included studies into the following: - Skin effects in conductors and joints. - Joint cooling by helicopter rotor wash. - Instrument calibration. - Protection of computer equipment against circulating currents. - Contact resistance of joints. - Electronic design and reliability in strong electric fields. -Effects of bonding-on currents on electronic equipment.

To comply with Transpower specifications, the test device had to be capable of measuring the voltage drop over each half of a joint, the coincident current through the joint (from which the a/c resistance was derived to a + or - 5% tolerance) and the joint surface temperature (tolerance + or - 2 degrees C). Wind speed at the time of the measurement was also recorded.

The key components of the device included: - An auto-ranging voltmeter capable of operating in the electric and magnetic field adjacent to the in-service joint. - Ammeter. - Thermocouple. - Analog to digital converter. - Microprocessors and software. - A two-way radio link. - Motorized clamps and controls. - Power supplies. - Voltage, current and temperature probes.

The device was extensively tested on various joint samples of known resistance to verify its accuracy at varying loads.

Live techniques for bare handing, hot sticking and lowering the device on rope evolved for accessing the various conductor configurations on the Transpower grid. To ensure continued data integrity, the measuring device was re-calibrated at the end of each work day.

The contractors commenced work in March 1997 and completed the project in June 1997.

Project Results The resistance of the original joint population, when new, would have a normal distribution; the spread of which would depend on the quality of the jointing techniques employed at the time of installation. As the population ages, the average resistance tends to increase. The key questions are, "How much will the resistance increase and is the increase linear with time or does it accelerate with age?"

Figure 4 shows the spread of resistance data obtained from the 940 "Goat" joints measured on the project, along with the envisioned spread when the joints were new. The spread for the new joints was based on sample measurements from newly made joints and from calculated theoretical maximum and minimum resistance values.

The joint resistance values have moved upward over time from an estimated approximate mean value of 35 mW/m to 70 mW/m (a doubling of resistance). However, the exact extent of the increase and the rates of change over time is uncertain. As additional information is gathered in subsequent years, the rate of aging versus time will begin to become more clear.

A typical resistance value for ACSR Goat conductor is 89.1 mW/m. The Transpower maximum acceptable value for a newly made joint is 50 mW/m. Fifty joints fall into the "replacement" category (above 89.1 mW /m) where the joint resistance is greater than the equivalent unit length of conductor. Other joint types tested indicate a similar joint resistance pattern.

As suspected, some joints displayed a large discrepancy between the two half-joint resistance values. This is indicative of an off-center outer sleeve and confirms the lack of proper jointing care during construction which has contributed to joint failures in the past.

Further Work Transpower is in the process of obtaining additional information on the joint aging process to assist with the formulation of a joint management process and joint life prediction model. The utility's processes will be adapted for the diverse local climatic conditions.

In particular, the study has prompted the following actions: - Replacement of inferior joints. - The potential for joint life extension through "live" re-compression of rogue joints using differently shaped and/or slightly undersized dies. - Laboratory testing of removed joints to investigate out-of-balance half resistance values, corrosion and oxidation levels, degree of aluminium annealing and stress reduction, installation geometry, joint bulk temperatures generated by operational and fault current and the efficiency exhibited by jointing compound in the old joints. - Confirmation of joint dimension adequacy for operating current density. - Measurement of urban joints (where helicopter use is not appropriate) using hot sticks to apply the measuring device from cranes or bucket trucks. Further development using this technology is being investigated for the measurement of: - Conductor corrosion using a resistance index (i.e. confirming that significant corrosion of inner layers of aluminium can be detected by resistance increase). - Conductor dead-end assembly, including jumper flag resistance (testing is under way). - Addition of a magnetic detector to indicate the relative location of the steel sleeve.

Summary This project has shown that it is technically feasible to routinely measure the resistance of in-service line compression joints, and that bad joints can now be found reliably years before they fail. It has also confirmed that aging processes slowly increase the resistance of all joints and that this factor needs to be considered when uprating older lines. Further, the new measurement techniques offer encouragement that this technology may potentially be used to find the location of internal corrosion on ACSR conductor.

It is also apparent that we need to learn much more about the deterioration processes affecting older joints. For example, we need to learn about the effects of jointing compound, age/load profiles, so that the speed of deterioration can be quantified with greater accuracy. We also need to decide what the safe upper joint resistance limits are for various potential fault duties.

Wal Marshall is an engineering support manager, transmission lines with the Operations Group of Transpower New Zealand Ltd. Marshall has worked in the electric utility industry for 27 years. The last 17 years have been spent on engineering and policy development related to transmission line maintenance work. He manages an internal lines consultancy group, which provides overhead line engineering, data, maintenance management and procurement services within Transpower and externally to consultants and contractors. Within Transpower, Marshall has been intimately involved in the introduction of live-line bare hand maintenance work, live-line tower painting systems, live-line midspan joint resistance measurement techniques, live-line tower replacement techniques and the development of a computerized predictive maintenance modeling system.

Trevor Jacobs holds a bachelor of science degree in civil engineering and has worked as a technical specialist for the Operations Group of Transpower New Zealand, Ltd. since July 1996. He previously spent 13 years in transmission engineering and distribution project management with ESKOM in South Africa. Within Transpower, Jacobs is involved with the live-line joint resistance project, live replacement of HVDC towers, line engineering and standards.