The Power Short-term Load Forecasting Based On Large Data Analysis Of The Research And Development

In this paper, the application of Hadoop and Spark in power load forecasting is studied. Spark Mlib and Mahout clustering algorithm is used to find the clustering algorithm based on the traditional clustering algorithm.This paper uses R to cluster the open source cluster, which can be used in different industries. The model is based on neural network and support vector machine. In short term load forecasting, support vector machine algorithm is not very good, but in short term load forecasting, the two algorithms are introduced. The results are good.Based on the research of this paper, the function modules of power short term load forecasting system are divided into the practical situation and functional requirements. In the traditional algorithm, the data is processed by a large data, which can’t run the same day. So the paper uses the related algorithms of Hadoop and Spark in the big data environment. The power load forecasting system is developed, and the total load forecasting and spatial load forecasting are carried out by the system. The results show that the prediction results are in good agreement with the actual situation. Prove that the software has good practicability.The innovation of this paper mainly has three points:first, the user in accordance with the big industry, small industry classification, then the user’s load of the cluster analysis, the better the user type; two, the use of Spark, Hadoop to carry out load forecasting; three, proposed an overall architecture scheme.