8th International Conference on Advanced Power System Automation and Protection (APAP2019)

882 / 2019-04-30 12:29:46

Power System Transient Frequency Prediction Model Based on Support Vector Machine Incremental Learning

frequency situation prediction,support vector machine,incremental learning

In order to effectively manage the new sample data continuously generated in power system operation, it is necessary to classify and learn the new data in time. If we re-learn all the data, it will take a lot of time, and may even cause the learning speed lagging behind the data update speed. a power grid transient frequency prediction mode is proposed based on SVM incremental learning. A fast SVM incremental learning method is adopted in this paper.It constructs a recursive solution and adds new data to the solution.Karush-Kuhn-Tucker conditions are maintained for all the previous used training data. The effectiveness of the transient prediction model updating algorithm is verified in the IEEE39 bus test system. The algorithm model updating time-consuming and prediction accuracy advantages are obvious, and it can adapt to the growth of power system transient stability sample set.