The traditional long-term load forecasting only predicts total annual load.As the load as a random variable,it is crucial to meet the demand of long-term load forecasting under flexible and changeable environment.We p...The traditional long-term load forecasting only predicts total annual load.As the load as a random variable,it is crucial to meet the demand of long-term load forecasting under flexible and changeable environment.We propose a panoramic probabilistic optimizer,which can refine the long-term load forecasting.The optimizer can describe long-term load changes over the next 8760 h.First,according to the statistical characteristics,the 8760-h load curve is decomposed into long-term trend,medium-term variation,and short-term variation.The analysis includes multiple time scale load indicators.Second,the load probability optimization model is established with the maximum load probability as the objective function,with transforming the panoramic load prediction into an optimization problem.According to the load characteristics of different time scales,constraint conditions are formed to solve the optimization model and restore the refined annual 8760-h load curve.Finally,it validates the approach by load data from Hubei and Guangxi,China.The results demonstrate the panoramic load can be accurately described in the form of refined 8760-h load curve.It can effectively improve the long-term prediction accuracy and has good generalization and interpretability.展开更多
基金supported by the National Natural Science Foundation of China(52177110)China Power Engineering Consulting Group CO.,LTD.Project Funding(DG2-X02-2022).
文摘The traditional long-term load forecasting only predicts total annual load.As the load as a random variable,it is crucial to meet the demand of long-term load forecasting under flexible and changeable environment.We propose a panoramic probabilistic optimizer,which can refine the long-term load forecasting.The optimizer can describe long-term load changes over the next 8760 h.First,according to the statistical characteristics,the 8760-h load curve is decomposed into long-term trend,medium-term variation,and short-term variation.The analysis includes multiple time scale load indicators.Second,the load probability optimization model is established with the maximum load probability as the objective function,with transforming the panoramic load prediction into an optimization problem.According to the load characteristics of different time scales,constraint conditions are formed to solve the optimization model and restore the refined annual 8760-h load curve.Finally,it validates the approach by load data from Hubei and Guangxi,China.The results demonstrate the panoramic load can be accurately described in the form of refined 8760-h load curve.It can effectively improve the long-term prediction accuracy and has good generalization and interpretability.
