1.School of Electrical Engineering, Northeast Electric Power University, Jilin 132012, China;2.School of Computer Science and Technology, Beihua University, Jilin 132021, China

Accurate spatial load situation awareness (SLSA) can provide a scientific guidance for the optimal planning of urban power grid. This paper proposes an SLSA method based on isolation forest, variational mode decomposition, multilayer perceptron and gated recurrent unit (iForest-VMD-MLP-GRU). In the stage of situation perception, the isolation forest algorithm is used to identify the outliers of the measured data of class Ⅰ cell load under the given spatial resolution in the power geographic information system, and the Lagrangian interpolation method is used to modify it, so as to determine the reasonable class Ⅰ cell load data. In the stage of situation comprehension, the perceived class Ⅰ cell load data is decomposed by the variational mode decomposition method to obtain the components of different central frequencies. The trend component and low-frequency component are determined according to their energy values. In the stage of situation forecasting, the multilayer perceptron and the gated recurrent unit are used to forecast the trend component and the low-frequency component respectively, and the forecasting results of the two components are inversed and reconstructed to obtain the situation awareness results of class Ⅰ cell loads in the target year. The grid technology is then used to transform the results of class Ⅰ cells into the results based on class Ⅱ cells. The results of example analysis prove that the proposed method is correct and effective.