Load forecasting tech to tide over power woes
Burdened with escalating power purchase costs, the two electricity distributing companies owned by BSES have imported a new technology that makes them predict Delhi’s appetite for power more accurately and, hopefully, save on both overbuying and last-minute scouting for power.delhi Updated: Jul 14, 2011 00:44 IST
Burdened with escalating power purchase costs, the two electricity distributing companies owned by BSES have imported a new technology that makes them predict Delhi’s appetite for power more accurately and, hopefully, save on both overbuying and last-minute scouting for power. Imported from Sweden, the technology cost the discoms Rs 2.5 crore.
It takes into account lifestyle pattern and income of consumers, population trends and the retail energy sales pattern — along with variations in weather — to draw up a more accurate picture of how much power Delhiites would need.
Known as SAS Smart Load Forecaster, it’s a level up from the prevalent practice of “intelligent guesswork” — that is the result of tracking historical load patterns coupled with the current weather forecast and past weather trends.
Copenhagen Energy and the Northern Virginia Electric Cooperative are among the leading utilities currently using this new tech.
“More accurate load forecasts will help us plan power purchases or sales more prudently and with greater accuracy, leading to a reduction in power purchase costs,” said Gopal Saxena, CEO, BSES Rajdhani.
The company claims that the savings accrued from the more precise planning will ultimately benefit the consumers; as reduced power costs will have a direct bearing on the customer tariffs. With the new technology in place, the discoms will also be able to plan their surplus power sales better, leading to higher returns from such sales to other utilities.
In manual load forecasting, the discoms divide the day into 96 slots, of 15 minutes each, and prepare estimates using historical data of consumption and weather.
The new tool uses statistical models to account for long-term trends and assess the effect of each parameter on demand at any given time. Hence forecasting with this level of around 1% accuracy — which was not possible manually — wherein, at best, a 3% error margin can be achieved.
Discoms say that in case of inaccurate load forecasting, utilities end up buying more power than necessary, then find it difficult to sell the extra power at a premium. The new system allows flexibility to quickly work in a cooler summer or colder winter and countless variations in between.