Short-term Power Load Forecasting Based On Improved SFLA-GRU

With the gradual improvement of people’s living standards,electricity is assuming an increasingly important role in people’s daily lives,and the power industry is an important industry that sustains national production and life.Accurate power load forecast can provide power generation basis for power plants,arrange next day or next week’s power generation plan in advance,and coordinate among power plants in advance to meet the demand for electricity in different areas;meanwhile,although new energy generation is being gradually put into use,it cannot fully meet the existing demand for electricity in the short term due to its high volatility and instability and small volume,and domestic power generation is still dominated by thermal power and hydropower.The coal resources are limited,and from the perspective of rational use of resources,accurate load forecasting can effectively reduce the waste of resources.To address the problems of high volatility and low prediction accuracy of short-term power load,this paper establishes a bidirectional gated recurrent unit model based on attention mechanism and improved shuffled frog leaping algorithm,and verifies it through simulation experiments,and the results show that the model in this paper can effectively improve the accuracy of load prediction.The details are as follows:(1)To clarify the background and significance of the current research,and to briefly introduce the classification,influencing factors and prediction methods of load forecasting according to the current status of domestic and international research.(2)Briefly introduce the three data sets used as examples for validation in the paper,explain the characteristics of each data set,make a preliminary analysis of the relationship between factors affecting load and load prediction,and pre-process the data to fill in missing values,correct abnormal values,and normalize all data.(3)Explain the basic concepts of deep learning and neural networks,introduce long and short term memory neural networks and gated recurrent units and their advantages and disadvantages,and explain the reasons for choosing bidirectional gated recurrent unit in this paper.The shuffled frog leaping algorithm is briefly explained,and the principle of the improved shuffled frog leaping algorithm proposed in this paper is introduced in detail,and its effectiveness is proved by the Rastrigin test function.The results show that the model can effectively reduce the prediction error and improve the prediction accuracy.(4)To clarify the basic principle of the attention mechanism,a bidirectional gated recurrent unit neural network model based on the attention mechanism and the improved shuffled frog leaping algorithm is established,and its prediction effect is verified with three cases.The results show that the model established in this part can further reduce the prediction error and improve the prediction accuracy on the basis of the above model.