When we turn on our computers, we take for granted that the power is there. Behind the scenes, however, that power is there because utilities knew you were going to need it -- perhaps even before you did.
Getting the energy use forecasts right is a critical job for utilities. Whether the forecasts are too high or too low, an error can lead to millions of dollars wasted.
Council Associate Partner TROVE is working to help utilities better predict the future and eliminate the waste. Its efforts are attracting more attention, including investments from Avista, a utility that serves the U.S. northwest, and Council Lead Partner Itron, which helps utilities manage energy and water.
Utilities have to delicately balance supply with demand, a job that’s made more difficult that they have to make some supply decisions days or even weeks in advance. They typically use weather forecasts combined with historical data to craft a forecast that is also based on a certain amount of gut feeling.
TROVE provides utilities with predictive analytics. It combines data fusion, data enrichment and sophisticated data science to give utilities better predictions of consumer behavior, allowing them to better predict the demand for energy on any given day.
TROVE says it’s also important that utility managers can trust the data. If it forecasts a significant increase or drop in usage, it also tells the utilities why. Avista says TROVE has helped it increase the accuracy of its three-to-five-day forecasts. Southern California Edison and Oklahoma Gas and Electric Company are also using TROVE’s analytics.
Even slight improvement can save millions
Analytics have been at or near the top of utility wish lists for years. While all utilities can benefit from them, it’s perhaps easiest to visualize with hydroelectric power. If the forecasted demand is too high, utilities send more water through dams than they really need to and potential energy is lost. If the forecast is too low, most utilities have to use rapid fire generators that burn fuel, which is expensive and has a much bigger environmental impact than renewable sources.
Avista uses TROVE’s solution to produce hour-by-hour forecasts 15 days in advance and study the forecasts after the fact to determine accuracy and identify areas for improvement. It says even slight improvements in the quality of the forecasts can save it money.
Weather is one of the biggest culprits in bad forecasts. During the summer and winter, a forecast that’s off by even five degrees can result in an unexpected surge in energy use, costing the utility $8 million a year. More accurate forecasts allow it to make better use of its hydroelectric sources, which is says could easily result in $2 million in savings over a five-year period.
Forecasting job is getting harder
California is among the areas mandating that utilities make more use of renewable energy sources, like wind, which complicates the forecasts further. If there’s not enough wind, utilities have to find other sources. The less notice they have, the more those sources costs.
And on the demand side, consumers are growing more complex, making forecasting models that utilities have used successfully for decades obsolete. That’s putting even more pressure on utilities to adopt new predictive analytics tools to not only improve, but, in some cases, achieve even the same level of accuracy they’ve had in the past.
More stories …
Itron shows how IoT is changing utilities (and why it's good news for cities)
How analytics help utilities speed storm response (and even prevent outages)
New Itron survey is a wake-up call for utilities