Research on Power Load Forecasting Model Based on Adaptive Feedback Loop Network

With the development of the power industry and the increasing use of distributed energy, short-term power load forecasting has become very important as a fundamental task of power system operation and planning. However, power loads are influenced by many factors and have complex nonlinear, and time-varying characteristics. It is challenging to capture dynamic changes using traditional forecasting methods accurately. Based on the above, this paper studies a power load forecasting model based on an adaptive feedback loop network (AFLN). First, we use the grey correlation analysis to select similar days as the training data set and construct the meteorological feature vector as the input variable. Second, the adaptive feedback loop network is used to learn and fit the training data set. The network can adaptively adjust the feedback weight and time stepping to improve the modeling ability of time series data. Finally, experiments are conducted on the actual load data and compared with the traditional feedback loop network (FLN) and long-short-term memory network (LSTM). The results show that the proposed AFLN method has high prediction accuracy and stability and has research significance in short-term power load forecasting.

I. Introduction

With the gradual development of deep learning technology, neural networks have been applied to various fields, such as image processing, natural language processing, and time series prediction. However, neural networks have many problems, including too many parameters, slow training, and overfitting. Therefore, people need to improve it.