Green and low-carbon is an essential direction for environmentally sustainable development. Under the vision of carbon neutrality, the entire society has accelerated the development of electrification, resulting in a continuous increase in electricity demand. Therefore, it is of great practical significance to analyze and forecast electricity demand reasonably and accurately. However, existing single forecasting models based on deterministic forecasting are difficult to capture the seasonal characteristics and nonlinear fluctuations in power demand data and obtain high forecasting accuracy. To fill this disadvantage, a hybrid electricity demand forecasting system based on a seasonal selection method, completely non-recursive decomposition strategy, and the sunflower optimization algorithm is proposed. Actual electricity demand data obtained in different seasons in various Australian states were used for the simulation. Empirical results indicate that the seasonality of irregular data is successfully classified by the seasonal selection method; the high-frequency noise of the original data are removed by the completely non-recursive decomposition strategy, and the forecasting accuracy and reliability of the optimized forecasting module are effectively improved. Based on the sunflower optimization algorithm, the kernel parameters and forecasting system are optimized to provide technical support for the development of electricity demand response and environmental energy protection.

绿色低碳是环境可持续发展的重要方向。在碳中和愿景下，全社会加快电气化发展，电力需求不断增加。因此，合理、准确地分析和预测电力需求具有重要的现实意义。然而，现有的基于确定性预测的单一预测模型难以捕捉电力需求数据的季节特征和非线性波动并获得较高的预测精度。为了弥补这一缺点，提出了一种基于季节选择方法、完全非递归分解策略和向日葵优化算法的混合电力需求预测系统。使用澳大利亚各州不同季节获得的实际电力需求数据进行模拟。实证结果表明，季节选择方法成功地对不规则数据的季节性进行了分类；通过完全非递归的分解策略去除了原始数据的高频噪声，有效提高了优化后的预测模块的预测精度和可靠性。基于向日葵优化算法，对内核参数和预测系统进行优化，为电力需求响应和环境能源保护的发展提供技术支撑。实证结果表明，季节选择方法成功地对不规则数据的季节性进行了分类；通过完全非递归的分解策略去除了原始数据的高频噪声，有效提高了优化后的预测模块的预测精度和可靠性。基于向日葵优化算法，对内核参数和预测系统进行优化，为电力需求响应和环境能源保护的发展提供技术支撑。实证结果表明，季节选择方法成功地对不规则数据的季节性进行了分类；通过完全非递归的分解策略去除了原始数据的高频噪声，有效提高了优化后的预测模块的预测精度和可靠性。基于向日葵优化算法，对内核参数和预测系统进行优化，为电力需求响应和环境能源保护的发展提供技术支撑。