Forecast of area‐scale behaviours of behind‐the‐metre solar power and load based on smart‐metering net demand data

Abstract

Abstract Local energy self‐sufficiency, in which the supply and demand of electricity are controlled such that the generated power from distributed energy resources (DERs) is consumed locally based on a power supply‐and‐demand forecast, mitigates the burden on the power system and contributes to the efficient use of DERs in smart cities. However, widely available smart metres cannot measure behind‐the‐metre pure demand and generation from prosumers. Pure demand and generation forecasts without additional metering contribute to advanced supply‐and‐demand control in smart cities, including demand response. This study proposes a method of forecasting spatio‐temporal behaviours of behind‐the‐metre pure demand and generation by focussing on the information of net demand distribution observable from the smart metres; the proposed method initially predicts the spatio‐temporal net demand distribution with a combined forecaster based on the persistence and non‐parametric regression models, and then separately estimates the behind‐the‐metre pure demand and generation by using demand forecast result of neighbouring pure‐consumers extracted by considering the area‐scale behaviours of the smart metering data. The simulation results demonstrate that the proposed method provides accuracy comparable to forecasts conducted by directly measuring pure demand and generation, without requiring the installation of additional metres.

Keywords

• Artificial Intelligence
• city brain and smart cities metrics
• Data Structures
• Data Analytics and Machine Learning
• distributed power generation
• power meters