Evaluation Of Smart Energy System Based On Machine Learning

In the context of the gradual shortage of fossil energy which is the main energy and the increasingly serious environmental pollution,the voice of energy transformation is increasing year by year.With the development of the Internet,along with the emergence of new technologies such as artificial intelligence,big data,and the Internet of Things,smart energy,as the comprehensive energy system that integrates multiple fields to achieve interconnection and good interaction among various systems to solve the current energy problem,is proposed as a new concept to solve the sustainable problem of energy development,and is paid more and more attention by all parties.It is the key research direction of energy industry upgrading in the world,and has gradually become a part of national strategies.However,the current development of smart energy is limited by the lack of objective and reasonable evaluation.Objective and reasonable evaluation can guide and correct the development direction of smart energy,which is of great significance to the research of smart energy system.But now,only a small number of evaluation methods have been proposed and they are all based on traditional evaluation methods,with more subjectivity and limitations.In view of this,this paper innovatively studies the evaluation system of smart energy system based on machine learning method for the first time to provide new ideas for research in this field.The main work and results of this paper are as follows:1.The definition and connotation of smart energy are studied in depth,and the previous definitions are integrated to give the explanation of smart energy’s definition and connotation with stronger universality and greater tolerance.Based on this,the current evaluation index system of smart energy system is obtained and optimized,and the weight feature vector of the index system is obtained by using the analytic hierarchy process.2.Constructing evaluation models based on three mature and widely used machine learning methods.For the three classical models including support vector machine,decision tree and BP neural network,the algorithm design of three machine learning methods is given,and the corresponding evaluation model is constructed.Then two optimization strategies are proposed.In order to solve the problem that the number of input data samples is discrete and scarce due to the difficulty in obtaining large amounts of relevant data in smart energy systems,Seq GAN is selected to reasonably expand the sample data.Aiming at the optimization setting of weights and thresholds of BP neural network,genetic algorithm is selected to obtain the optimized weights and thresholds,and the process of optimization algorithm is given and the corresponding model network is constructed.3.The evaluation models built before and after optimization are simulated and compared.The example results show that the machine learning method plays a good role in the study of smart energy system evaluation.The optimization method has high robustness and fitting ability to discrete nonlinearity.It is suitable for the evaluation of smart energy systems with high feature dimensions.In addition,compared with the traditional methods,the machine learning evaluation model is more objective and efficient.Therefore,this paper proves the feasibility of evaluation based on machine learning,and reveals that there is a broader space for optimization research in this field.4.Based on the above research results,the specific needs are analyzed,the system uses are clarified,and a complete set of smart energy system evaluation prototype system is designed and implemented.This all provides an objective basis and reference for the specific evaluation of cities implemented by smart energy in the future and the formulation of their subsequent improvement strategies to realize the convenience and intelligence of smart energy evaluation.