In an article for Public Utilities Fortnightly, Navigant says utilities can leverage big data to increase accuracy and replace outdated forecast models
For many utilities, load forecasting errors have grown to an unprecedented level. And as increasing complexities present themselves in today’s ever-evolving energy landscape, Navigant’s Ken Seiden and Brian Eakin say it’s time to consider a new long-term forecasting model — one that’s bottom-up and combines strategic customer segmentation, customer choice modeling, and granular load analytics to deliver load forecasts that decision-makers can trust.
In an article for Public Utilities Fortnightly, Seiden and Eakin explain that forecasting models have largely been the same since the 1970s, when utilities started using econometric and mixed end-use econometric models. However, as changing customer demands and the proliferation of energy-consuming and producing alternatives take hold, taking advantage of data analytics could increase forecast accuracy significantly.
“The bottom-up load forecasting framework we have outlined takes advantage of existing grid modernization and information technology investments, along with advances in machine learning and structural analysis techniques,” the pair wrote. “…Moving to this forecasting approach that is driven by big data is not a trivial undertaking. It takes a thoughtful and coordinated effort to make use of all information available to the utility. And a preparedness to respond to insights that are far more detailed and accurate than have been previously produced.”