The electrical energy is basic necessity for daily operations. Therefore, its efficient and intelligent management is required. However, the utilities require rapid and efficient decision-making which depends on efficient load forecasting models. The existing load forecasting models stand obsolete in the dynamic environments, where rapid decisions are required; based on online data streams. The online data streams in electrical environment comprise of different generation modalities and consist of different variations (concept drift) in pattern, parameter, and modalities. Therefore, this study explores the possible concept drift scenarios in electrical load forecasting with different generation modalities and provide a framework to resolve the issue. © 2022 IEEE.