Transient transmission system faults are of extremely short duration, milliseconds, but their effect and cost to a utility and its customers can be considerable. A 2001 study by the Electric Power Research Institute concluded that poor power quality costs the American economy between US$15 billion to $25 billion per annum. Similar South African studies show the cost is highly variable, with the estimated customer cost varying between ZAR 5000 and ZAR 150,000 (US$700 to $21,000) per event.

The performance of Eskom's transmission lines in terms of transient line faults has been the subject of supply quality studies since the inception of the utility's transmission division in 1993. Fires beneath and near transmission lines account for some 20% of line faults. Eskom has implemented an Advanced Fire Information System (AFIS) to monitor fires under the transmission lines in real time and convey this information to field staff. The system also improves the management of fires by providing the opportunity to collate historical data on burn scar areas.

Faults Attributable to Fires

Fires in South Africa are part of the landscape and most plants have adapted to fire, with some even dependent on it for germination of their seeds. These fires should not be compared with the high-intensity fires that regularly occur in Australia and California, rather these are low-intensity grass fires that produce sufficient ionized air to reduce the insulation properties of air-causing line faults. The effect of these fires on the duty cycles and life of transmission equipment remains largely unquantified and difficult to determine, but approximately 11,500 line faults of all types have been recorded on Eskom's transmission system since 1993.

To illustrate the effect of these fires, in 2002, two grass fires occurring on the same day some 295 km (183 miles) apart gave rise to the tripping and lockout of two 765-kV lines in the Standerton area and the Province of the Free State, resulting in the loss of supplies to Cape Town. This single event led Eskom to revisit its policy with respect to vegetation management against fires. The main findings of the subsequent investigation was the role of atmospheric conditions during fires, its influence on fire intensity and the ability of the fire to cause a line fault.

Prior this event, Eskom adopted a traditional approach with regard to fires by reducing the vegetation (or fuel load) under the lines; typically, this was in the form of grass cutting. However, this form of fuel reduction had limited success due to the fact that, in most cases, the cut grass remained in situ and, as a result, the overall fuel load remained the same. In situations where the lines crossed wetlands or rocky areas, access was limited and cutting was normally not undertaken, leaving those areas vulnerable to fires.

Weather Forecasting

As a result of the major system outage in 2002, Eskom arranged a daily weather forecast that predicted the expected ambient temperature, relative humidity and wind speed for the whole country. These atmospheric parameters are combined into what is known as the fire danger index, a term used by the forestry industry across the world.

It became clear to Eskom that dangerous fire conditions are often associated with pre-frontal weather patterns, which result in hot, dry and windy conditions during the last part of the Austral winter. The weather forecast became a useful tool in many aspects for the operation of the transmission system. Field staff knew when to be on standby, and the national control centre could use the information in the permitting outages for maintenance work. In addition to being able to predict fire weather, Eskom realized that a system to detect fires in real time also was necessary in the fight against fires.

Satellite Applications

Today, satellites are used for a variety of purposes, such as navigation, weather forecasting, real-time tracking, radio Internet communications and even TV broadcasts. Satellite images have numerous uses, namely in agriculture, geology, forestry, regional planning, education, conservation and national intelligence. The launching of the Aqua and Terra satellites with the moderate-resolution imaging spectroradiometer (MODIS) by NASA in 1999 and 2002, respectively, provided the world with a tool to be used for fire tracking. The University of Maryland demonstrated the successful use of this data in the mapping of fires across the globe.

Near-real-time fire detection was not possible with the Maryland system. However, when the South African Department of Agriculture purchased the satellite data and a MODIS antenna in 2003, fast detection became a reality for Eskom.

In 2004, Eskom and the Satellite Application Centre (SAC) of the Council of Scientific and Industrial Research (CSIR) launched a joint research project to demonstrate the ability to track active fires by using polar orbiting satellites. MODIS, a medium-resolution scanner, provided four updates daily with a 1-sq km (0.4-sq mile) resolution, making the detection of grass fires as small as 0.25 hectares (0.62 acres) possible.

Eskom required information on fires every 15 minutes, and while the MODIS data was sufficiently high in spatial resolution, the temporal resolution, with updates on 6-hour intervals, was less than satisfactory. Consequently, CSIR proposed the use of the spinning-enhanced visible and infrared imager (SEVIRI) sensor onboard the Meteosat Second Generation (MSG) satellite. This satellite is in geostationary orbit above the equator, and the SEVIRI sensor transmits data every 15 minutes with a spatial resolution of 5 sq km (2 sq miles).

Alerting Field Personnel

Once the processing of the hot spots is completed, the information is published on a website or e-mail alerts are sent. Since Eskom field personnel are normally on field patrols, text messages warning of fires are sent to their mobile phones. In the case of overhead lines, only fires within 2.5 km (1.6 miles) of a circuit are reported. Where possible, the national control centre can temporarily isolate the circuit under threat, and field staff can activate fire-suppression teams where available.

This system was the first of its kind in the world where an electrical utility applied remote sensing together with cell phone technology in the monitoring of fires under overhead lines. Eskom does not have in-house firefighting capability and relies on the many Fire Protection Associations (FPAs) and fire brigades across the country to extinguish fires. Consequently, the web and text message service is extended to these entities free of charge.

The system developed through this research project was named Advanced Fire Information System (AFIS) and was initially developed on Esri's proprietary ArcIMS software. Most users adapted to the system without much trouble, although the system sometimes proves to be slow in areas of low bandwidth. In addition to the real-time functionality of the system, the historical record of fires accumulated on the system proves to be most valuable. The system permits queries on fires observed for any stated period, and the data is used in the planning of vegetation management, investigations of line faults and insurance claims.

Although the system has proved to be valuable in the fight against fires, it has some shortcomings as not all fires are detected. The data shows that up to 82% of fires associated with flashovers are detected before the flashover, and in about 60% of cases, fires are observed at the time of the flashover. Because grass fires in South Africa normally last a short time, the higher temporal resolution of the MSG shows a distinct advantage in spite of its lower spatial resolution.

Toward the end of 2007, it was decided to add a further MODIS antenna to improve the redundancy of AFIS. Coupled with this was the decision by Eskom to use “open” systems software in South Africa, Eskom considered it time to upgrade AFIS.