基于多维气象信息时空融合和MPA-VMD的短期电力负荷组合预测模型

为提高电力负荷预测精度，需考虑区域内不同地区多维气象信息对电力负荷影响的差异性。在空间维度上，提出多维气象信息时空融合的方法，利用Copula理论将多座气象站的风速、降雨量、温度、日照强度等气象信息与电力负荷进行非线性耦合分析并实现时空融合。在时间维度上，采用海洋捕食者算法(MPA)实现变分模态分解(VMD)核心参数的自动寻优，并采用加权排列熵构造MPA-VMD适应度函数，实现负荷序列的自适应分解。通过将时间维度各分量与空间维度各气象信息进行融合构造长短期记忆(LSTM)网络模型与海洋捕食者算法-最小二乘支持向量机(MPA-LSSVM)模型的输入集，得到各分量预测结果，根据评价指标选择各分量对应的预测模型，重构得到整体预测结果。算例分析结果表明，所提预测模型优于传统预测模型，有效提高了电力负荷预测精度。

Combination forecasting model of short-term power load based on multi-dimensional meteorological information spatio-temporal fusion and MPA-VMD

In order to improve the accuracy of power load forecasting, it is needed to consider the influence difference of multi-dimensional meteorological information on power load in different places of a region. In the spatial dimension, a spatio-temporal fusion method of multi-dimensional meteorological information is proposed, and the Copula theory is used for the nonlinear coupling analysis between meteorological information of multiple meteorological stations such as wind speed, rainfall, temperature, sunshine intensity and power load, so as to realize spatio-temporal fusion. In the time dimension, the marine predator algorithm(MPA) is adopted to realize the automatic optimization of core parameters of variational modal decomposition(VMD),and the weighted permutation entropy is adopted to construct the adaptation function of MPA-VMD, which realizes the adaptive decomposition of load sequence. The input sets of long short-term memory(LSTM) network model and marine predator algorithm-least squares support vector machine(MPA-LSSVM) model are constructed by fusing each component of time dimension and each meteorological information of spatial dimension, the forecasting results of each component are obtained, the forecasting model corresponding to each component is selected according to the evaluation index, and the overall forecasting results are reconstructed. The example analysis results show that the proposed forecasting method is better than the traditional forecasting method, and effectively improves the accuracy of power load forecasting.