COLUMN-Power price, frequency link suggests trade opportunity: Wynn
(The author is a Reuters columnist. The opinions expressed are his own)
LONDON, Oct 8 (Reuters) - A link between British grid frequency - set by the speed at which power turbines spin - and electricity prices suggests a trading opportunity in European spot markets, where liquidity is rising to balance a more variable power supply.
Traditionally, grid operators have focused on balancing electricity generation with variation in demand, which peaks in Britain at around 0900-1000 and 1830-1930 local time.
However, wind and solar power have added variability on the power supply side as well, for example as a result of weather forecasting inaccuracies.
That has led to a greater need for grid balancing and a rise in shorter dated power supply contracts, in increasingly liquid intra-day markets.
Grid operators must maintain system frequency within a narrow range to maintain grid stability: a fall in frequency implies greater electricity demand than supply, and a need to ramp up power output, and vice versa.
As system frequency rises and falls, grid operators have a range of options to call upon, including requiring an automated, remote change in output at operating power plants, as well as through trades in spot wholesale markets.
Frequency can therefore be an indicator of power demand.
British data show a statistically significant, negative correlation between system frequency and spot power prices, suggesting frequency can indeed be a price discovery tool.
At fossil fuel power plants, steam turbines drive large spinning generators and so set a speed of rotation, which in Britain should be 50 times a second.
That in turn sets the frequency of alternating current at 50 Hertz (Hz).
Generators rotate synchronously, coupled together through the power grid.
When there is not enough power available to meet demand, generators slow down together and the frequency falls.
Ultimately, if power frequency falls below a certain threshold then power supply is insufficient and operators will shut off some supply, causing blackouts, as an emergency procedure to keep generators spinning.
That scenario should not arise given a well-established response system over a time period of seconds to hours - from remote frequency response, to the start-up of gas combustion engines and ramping up of power plants already operating at part-load.
Over the past five years, onshore wind power generation has more than doubled in Britain, to 12,121 gigawatt hours (GWh) in 2012 from 5,788 GWh in 2008, government statistics show.
Data for daily wind power generation over the last two years illustrates both the growth and variability of supply.
Given that grid operators have to maintain frequency in a narrow band, there is no evidence from the available frequency data of rising volatility.
However, there is evidence of growing liquidity in power markets for spot power, indicative of an increasing need to balance a more variable supply at shorter notice.
For example, intra-day trade in the Netherlands in September more than doubled year-on-year, and was up 65 percent on the Belgian Belpex power exchange, the power market exchange APX says.
APX publishes prices for UK half-hour contracts, based on a volume-weighted reference price for trades during each half-hour settlement period each day.
These prices can be compared with the average, British grid system frequency over the corresponding half-hour period.
One would expect there to be a negative relationship: the higher the frequency the more the system is in over-supply, implying power supply and prices should fall.
Historical frequency data are available in five-minute intervals from the Balancing Mechanism used by the National Grid to balance power flows, summarised and published online by Gridwatch (http://www.gridwatch.templar.co.uk/index.php).
A Reuters analysis compared two sets of corresponding data for 54 selected half-hour time periods between June 3 and June 7, 2013.
They show a correlation of -0.26, on a scale of zero to one.
That value and sample size show with a 95 percent certainty that there is a relationship between the two sets of data, rather than the sets varying by chance, statistical tables show.
This is a quick study of the relationship between grid system frequency and intra-day power prices.
More study and a larger dataset may show different results.
But the findings suggest that trading against frequency data may indeed be a useful strategy in increasingly liquid intra-day power markets.
(Reporting by Gerard Wynn; Editing by Pravin Char)