The Application Of Deep Learning Model Based On Recurrent Neural Network

With the development of society and power enterprises,the electricity consumption is increasing year by year,and the requirements for the stability of the power system are also increasing year by year.Because electricity is a kind of special data secondary energy,its particularity decides that it can not be stored in large quantities.To ensure sufficient supply of electricity and to ensure energy conservation and sustainable development as much as possible,The accuracy of load forecasting requirements gradually increase.In recent years,various prediction methods emerge in an endless stream.Although some traditional load forecasting has been relatively mature,in the face of increasingly higher accuracy requirements,existing methods must be improved or new prediction methods established.In this paper,by contrasting and studying several neural networks and deep learning models at the present stage,this paper analyzes the charact eristics of load forecasting data,combines the limited Boltzmann model and recursive neural network,The hidden layer of the machine joined the recursive loop,forming a new type of network model.Then,the structure and algorithm of the model are introdu ced.Experiments and comparison methods are used to verify the network with power load data and compared with other neural networks.The future of this paper is illustrated by introducing the network in this paper and comparing with other networks.Through experimental analysis,the network constructed in this paper can overcome the shortcomings of a single network model,and can fully combine the advantages of the two,and achieved good results in power load forecasting.