基于知识图谱与特征识别的新能源微网消纳数据分析算法研究

Research on consumption data analysis algorithm of new energy microgrid based on knowledge atlas and feature recognition

针对现有电力数据分析算法缺乏结果可视性以及准确性不理想的问题,提出一种基于知识图谱与特征识别的新能源微网消纳数据分析算法。该算法根据微网架构明确新能源的出力模型,并利用结合长短时记忆网络和条件随机场的知识图谱技术,将微网中新能源消纳数据转化成图形结构。同时通过图卷积神经网络模型识别数据的图特征,并根据识别结果制定微网能量调控措施,以提高微网新能源的消纳能力。基于某微网的真实数据集对所提算法进行实验论证,结果表明所提算法的准确率、召回率和F1值分别为91.53%、89.95%、90.81%,均优于其他算法,故具有良好的工程应用价值。

In view of the lack of result visibility and poor accuracy of existing power data analysis algorithms,a new energy microgrid consumption data analysis algorithm based on knowledge atlas and feature recognition is proposed. The algorithm defines the output model of new energy according to the microgrid architecture,and uses the knowledge atlas technology combining long short-term memory network and conditional random field to convert the new energy consumption data in microgrid into graphic structure. The graph features of data are identified by graph convolution neural network model,and the microgrid energy regulation measures are formulated according to the recognition results,so as to improve the consumption capacity of microgrid new energy. The experimental results show that the accuracy,recall and F1 values of the proposed algorithm are 91.53%,89.95% and 90.81% respectively,which are better than other algorithms and have good engineering application value.