Accordingly, Parliament passed the Electricity (Supply) Act of 1926, which created a Central Electricity Board with the duty of constructing and operating a new transmission system, or grid, connecting the power stations in England, Wales and central Scotland, and adopting a standard frequency of 50 Hz. The grid was to operate at 132,000 volts (132 kV), twice the voltage previously used in the United Kingdom, so it was not surprising that, despite high standards of engineering and design, some operational problems occurred initially.
The most serious problem was the flash-over of the 132 kV overhead line insulators in fog. This flashover—the sparking that occurs around an insulator when subjected to a voltage high enough to overcome its resistance—would cause the line to be switched out, possibly leading to an interruption of the electricity supplied to consumers. In the early 1930s, the fogs in winter were very dense in industrial areas because of severe air pollution caused by emissions from factory chimneys and smoke from coal-burning domestic fires. Visibility often was only a few yards, and in the dark all road traffic came to a standstill.
These fogs deposited a moist conducting layer on the insulator surfaces which soon led to the flashover of the insulators, even though the insulator length seemed to be adequate in comparison with practice in other countries. We had to study the question in some detail to elucidate the mechanism of flashover and discover a remedy.
We selected a site adjacent to the 132 kV substation at Croydon, where fogs were then prevalent and flashovers of substation insulators had been experienced. There we constructed a full-scale testing station using a 132 kV transmission tower on which insulators of various designs were suspended and their performance observed under the full grid voltage supplied from a test transformer. We inserted an additional insulator unit between each insulator under test and the steel tower, and connected each insulator to the laboratory so its performance could be measured.
The classical method of measuring the quality of an electrical insulator is to use a Schering bridge, an electric circuit that indicates the "loss factor" of an insulator when an alternating voltage is applied. This bridge must be balanced manually—but we were unable to balance the bridge when conditions were critical and the leakage current over the insulator was varying rapidly over wide limits. So we resorted to a simpler method: measurement of the leakage current over the surface of the insulator.
This current could not be measured by the usual pointer instrument, because in fog it varied so erratically. It happened, however, that in the substation there was a chart recorder scaled 0-100 mA for recording neutral currents. This instrument had a long paper chart driven by clockwork at a rate of one inch per hour, the current being recorded by an ink-filled pointer that rested on the chart. It never seemed to record anything in the substation, so we borrowed it and connected it to record the leakage current over an insulator.
Normally the current was too small to measure, but one day the fog came down and then what a sight met our eyes! The chart was recording a surge of current every few minutes with amplitudes of up to 100 mA.
Now we understood the mechanism of insulator flashover in fog and realized that a flashover developed only when a particularly large leakage current surge occurred. We were able to determine a standard for insulator performance in fog, based on the number and amplitude of the surges of leakage current.
Recording the leakage current had one disadvantage, however; most of the time the recorder was registering zero and producing yards and yards of useless chart. The solution also started in the substation, where some telephone contractors were using telephone-call counters; we borrowed one and connected it in an insulator leakage-current circuit. When a fog occurred, the counter ticked up all the surges above a certain value; with surge counters connected in each of the insulator test circuits, the number of surges recorded could be used as a measure of the performance of the individual insulators. This method was subsequently used in many countries to measure insulator performance in polluted atmospheres.
By using this method of measuring the performance of insulators in fog it was possible to select the best of the available insulators and test new designs of "anti-fog" insulators. Many of the original disc insulators were changed to designs with longer leakage distances; their performance improved, and a satisfactory standard of service on the grid system was achieved.
In the 1950s and 1960s, higher-voltage grid systems operating at 275 kV and 400 kV had to be built to carry the increasing loads. Fog was no longer such a serious problem, because we could select a satisfactory type of insulator at a test station in fog or in a test chamber in which fog was simulated. Furthermore, the use of gas and electricity for heating and the virtual disappearance of the domestic coal fire have greatly reduced air pollution, and the "pea-soup" fogs have gone forever.
But in the early days of the first grid system, a difficult problem was solved by making full-scale tests in the actual weather conditions and assessing the results using simple instrumentation that happened to be available by chance.