Short Term Power Load Forecasting Based on Image Matrix

Aiming at the problem of short-term power load forecasting, this paper proposes a load forecasting method based on image matrix. This method converts the time series of power load into a visual graph(VG), then extracts statistical features from the visual graph, captures the fine-grained features of the time series, and constructs the image matrix of the sequence. Finally, it inputs the load forecasting model based on convolutional neural network. Compared with the traditional load features based on time series, the image matrix has more fine-grained load features. The experimental results show that the short-term load forecasting accuracy of the proposed is better than the time series.